Merge branch 'develop' into update_ctr

.gitignore

-.DS_Store
+paddle/operators/check_t.save
+paddle/operators/check_tensor.ls
+paddle/operators/tensor.save
+python/paddle/v2/fluid/tests/book/image_classification_resnet.inference.model/
+python/paddle/v2/fluid/tests/book/image_classification_vgg.inference.model/
+python/paddle/v2/fluid/tests/book/label_semantic_roles.inference.model/
+*.DS_Store
+*.vs
+build/
+build_doc/
+*.user
+
 .vscode
 .idea
 .project
 .cproject
 .pydevproject
 .settings/
+
 *.pyc
-.*~
-fluid/neural_machine_translation/transformer/deps
-fluid/neural_machine_translation/transformer/train.data
-fluid/neural_machine_translation/transformer/train.pkl
-fluid/neural_machine_translation/transformer/train.sh
-fluid/neural_machine_translation/transformer/train.tok.clean.bpe.32000.en-de
-fluid/neural_machine_translation/transformer/vocab.bpe.32000.refined
+CMakeSettings.json
+Makefile
+.test_env/
+third_party/
+
+*~
+bazel-*
+third_party/
+
+build_*
+# clion workspace.
+cmake-build-*
+model_test
\ No newline at end of file

fluid/DeepASR/.gitignore

+.idea

fluid/DeepASR/data_utils/util.py

@@ -25,16 +25,6 @@ def to_lodtensor(data, place):
     return res
 
 
-def lodtensor_to_ndarray(lod_tensor):
-    """conver lodtensor to ndarray
-    """
-    dims = lod_tensor._get_dims()
-    ret = np.zeros(shape=dims).astype('float32')
-    for i in xrange(np.product(dims)):
-        ret.ravel()[i] = lod_tensor.get_float_element(i)
-    return ret, lod_tensor.lod()
-
-
 def split_infer_result(infer_seq, lod):
     infer_batch = []
     for i in xrange(0, len(lod[0]) - 1):

fluid/DeepASR/decoder/.gitignore

+ThreadPool
+build
+post_latgen_faster_mapped.so
+pybind11

fluid/DeepASR/examples/aishell/.gitignore

+aux.tar.gz
+aux
+data
+checkpoints

fluid/DeepASR/examples/aishell/train.sh

 export CUDA_VISIBLE_DEVICES=4,5,6,7
-python -u ../../train.py --train_feature_lst data/train_feature.lst \
+python -u  ../../train.py --train_feature_lst data/train_feature.lst \
                    --train_label_lst data/train_label.lst \
                    --val_feature_lst data/val_feature.lst \
                    --val_label_lst data/val_label.lst \
@@ -7,7 +7,8 @@ python -u ../../train.py --train_feature_lst data/train_feature.lst \
                    --checkpoints checkpoints \
                    --frame_dim 80  \
                    --class_num 3040  \
+                   --print_per_batches 100 \
                    --infer_models '' \
-                   --batch_size 64 \
+                   --batch_size 16 \
                    --learning_rate 6.4e-5 \
                    --parallel

fluid/DeepASR/model_utils/model.py

@@ -5,19 +5,21 @@ from __future__ import print_function
 import paddle.fluid as fluid
 
 
-def stacked_lstmp_model(frame_dim,
+def stacked_lstmp_model(feature,
+                        label,
                         hidden_dim,
                         proj_dim,
                         stacked_num,
                         class_num,
                         parallel=False,
                         is_train=True):
-    """ The model for DeepASR. The main structure is composed of stacked 
-        identical LSTMP (LSTM with recurrent projection) layers.
+    """ 
+    The model for DeepASR. The main structure is composed of stacked 
+    identical LSTMP (LSTM with recurrent projection) layers.
 
-        When running in training and validation phase, the feeding dictionary
-        is {'feature', 'label'}, fed by the LodTensor for feature data and 
-        label data respectively. And in inference, only `feature` is needed.
+    When running in training and validation phase, the feeding dictionary
+    is {'feature', 'label'}, fed by the LodTensor for feature data and 
+    label data respectively. And in inference, only `feature` is needed.
 
     Args:
         frame_dim(int): The frame dimension of feature data.
@@ -28,80 +30,45 @@ def stacked_lstmp_model(frame_dim,
         is_train(bool): Run in training phase or not, default `True`.
         class_dim(int): The number of output classes.
     """
+    conv1 = fluid.layers.conv2d(
+        input=feature,
+        num_filters=32,
+        filter_size=3,
+        stride=1,
+        padding=1,
+        bias_attr=True,
+        act="relu")
 
-    # network configuration
-    def _net_conf(feature, label):
-        conv1 = fluid.layers.conv2d(
-            input=feature,
-            num_filters=32,
-            filter_size=3,
-            stride=1,
-            padding=1,
-            bias_attr=True,
-            act="relu")
-
-        pool1 = fluid.layers.pool2d(
-            conv1, pool_size=3, pool_type="max", pool_stride=2, pool_padding=0)
-
-        stack_input = pool1
-        for i in range(stacked_num):
-            fc = fluid.layers.fc(input=stack_input,
-                                 size=hidden_dim * 4,
-                                 bias_attr=None)
-            proj, cell = fluid.layers.dynamic_lstmp(
-                input=fc,
-                size=hidden_dim * 4,
-                proj_size=proj_dim,
-                bias_attr=True,
-                use_peepholes=True,
-                is_reverse=False,
-                cell_activation="tanh",
-                proj_activation="tanh")
-            bn = fluid.layers.batch_norm(
-                input=proj,
-                is_test=not is_train,
-                momentum=0.9,
-                epsilon=1e-05,
-                data_layout='NCHW')
-            stack_input = bn
-
-        prediction = fluid.layers.fc(input=stack_input,
-                                     size=class_num,
-                                     act='softmax')
+    pool1 = fluid.layers.pool2d(
+        conv1, pool_size=3, pool_type="max", pool_stride=2, pool_padding=0)
 
-        cost = fluid.layers.cross_entropy(input=prediction, label=label)
-        avg_cost = fluid.layers.mean(x=cost)
-        acc = fluid.layers.accuracy(input=prediction, label=label)
-        return prediction, avg_cost, acc
-
-    # data feeder
-    feature = fluid.layers.data(
-        name="feature",
-        shape=[-1, 3, 11, frame_dim],
-        dtype="float32",
-        lod_level=1)
-    label = fluid.layers.data(
-        name="label", shape=[-1, 1], dtype="int64", lod_level=1)
-
-    if parallel:
-        # When the execution place is specified to CUDAPlace, the program will
-        # run on all $CUDA_VISIBLE_DEVICES GPUs. Otherwise the program will 
-        # run on all CPU devices.
-        places = fluid.layers.device.get_places()
-        pd = fluid.layers.ParallelDo(places)
-        with pd.do():
-            feat_ = pd.read_input(feature)
-            label_ = pd.read_input(label)
-            prediction, avg_cost, acc = _net_conf(feat_, label_)
-            for out in [prediction, avg_cost, acc]:
-                pd.write_output(out)
+    stack_input = pool1
+    for i in range(stacked_num):
+        fc = fluid.layers.fc(input=stack_input,
+                             size=hidden_dim * 4,
+                             bias_attr=None)
+        proj, cell = fluid.layers.dynamic_lstmp(
+            input=fc,
+            size=hidden_dim * 4,
+            proj_size=proj_dim,
+            bias_attr=True,
+            use_peepholes=True,
+            is_reverse=False,
+            cell_activation="tanh",
+            proj_activation="tanh")
+        bn = fluid.layers.batch_norm(
+            input=proj,
+            is_test=not is_train,
+            momentum=0.9,
+            epsilon=1e-05,
+            data_layout='NCHW')
+        stack_input = bn
 
-        # get mean loss and acc through every devices.
-        prediction, avg_cost, acc = pd()
-        prediction.stop_gradient = True
-        avg_cost = fluid.layers.mean(x=avg_cost)
-        acc = fluid.layers.mean(x=acc)
-    else:
-        prediction, avg_cost, acc = _net_conf(feature, label)
+    prediction = fluid.layers.fc(input=stack_input,
+                                 size=class_num,
+                                 act='softmax')
 
+    cost = fluid.layers.cross_entropy(input=prediction, label=label)
+    avg_cost = fluid.layers.mean(x=cost)
+    acc = fluid.layers.accuracy(input=prediction, label=label)
     return prediction, avg_cost, acc

fluid/DeepASR/train.py

@@ -14,7 +14,6 @@ import data_utils.augmentor.trans_add_delta as trans_add_delta
 import data_utils.augmentor.trans_splice as trans_splice
 import data_utils.augmentor.trans_delay as trans_delay
 import data_utils.async_data_reader as reader
-from data_utils.util import lodtensor_to_ndarray
 from model_utils.model import stacked_lstmp_model
 
 
@@ -24,7 +23,8 @@ def parse_args():
         '--batch_size',
         type=int,
         default=32,
-        help='The sequence number of a batch data. (default: %(default)d)')
+        help='The sequence number of a batch data. Batch size per GPU. (default: %(default)d)'
+    )
     parser.add_argument(
         '--minimum_batch_size',
         type=int,
@@ -147,29 +147,72 @@ def train(args):
     if args.infer_models != '' and not os.path.exists(args.infer_models):
         os.mkdir(args.infer_models)
 
-    prediction, avg_cost, accuracy = stacked_lstmp_model(
-        frame_dim=args.frame_dim,
-        hidden_dim=args.hidden_dim,
-        proj_dim=args.proj_dim,
-        stacked_num=args.stacked_num,
-        class_num=args.class_num,
-        parallel=args.parallel)
-
-    # program for test
-    test_program = fluid.default_main_program().clone()
-
-    #optimizer = fluid.optimizer.Momentum(learning_rate=args.learning_rate, momentum=0.9)
-    optimizer = fluid.optimizer.Adam(
-        learning_rate=fluid.layers.exponential_decay(
-            learning_rate=args.learning_rate,
-            decay_steps=1879,
-            decay_rate=1 / 1.2,
-            staircase=True))
-    optimizer.minimize(avg_cost)
+    train_program = fluid.Program()
+    train_startup = fluid.Program()
 
+    with fluid.program_guard(train_program, train_startup):
+        with fluid.unique_name.guard():
+            py_train_reader = fluid.layers.py_reader(
+                capacity=10,
+                shapes=([-1, 3, 11, args.frame_dim], [-1, 1]),
+                dtypes=['float32', 'int64'],
+                lod_levels=[1, 1],
+                name='train_reader')
+            feature, label = fluid.layers.read_file(py_train_reader)
+            prediction, avg_cost, accuracy = stacked_lstmp_model(
+                feature=feature,
+                label=label,
+                hidden_dim=args.hidden_dim,
+                proj_dim=args.proj_dim,
+                stacked_num=args.stacked_num,
+                class_num=args.class_num)
+            # optimizer = fluid.optimizer.Momentum(learning_rate=args.learning_rate, momentum=0.9)
+            optimizer = fluid.optimizer.Adam(
+                learning_rate=fluid.layers.exponential_decay(
+                    learning_rate=args.learning_rate,
+                    decay_steps=1879,
+                    decay_rate=1 / 1.2,
+                    staircase=True))
+            optimizer.minimize(avg_cost)
+            fluid.memory_optimize(train_program)
+
+    test_program = fluid.Program()
+    test_startup = fluid.Program()
+    with fluid.program_guard(test_program, test_startup):
+        with fluid.unique_name.guard():
+            py_test_reader = fluid.layers.py_reader(
+                capacity=10,
+                shapes=([-1, 3, 11, args.frame_dim], [-1, 1]),
+                dtypes=['float32', 'int64'],
+                lod_levels=[1, 1],
+                name='test_reader')
+            feature, label = fluid.layers.read_file(py_test_reader)
+            prediction, avg_cost, accuracy = stacked_lstmp_model(
+                feature=feature,
+                label=label,
+                hidden_dim=args.hidden_dim,
+                proj_dim=args.proj_dim,
+                stacked_num=args.stacked_num,
+                class_num=args.class_num)
+    test_program = test_program.clone(for_test=True)
     place = fluid.CPUPlace() if args.device == 'CPU' else fluid.CUDAPlace(0)
     exe = fluid.Executor(place)
-    exe.run(fluid.default_startup_program())
+    exe.run(train_startup)
+    exe.run(test_startup)
+
+    if args.parallel:
+        exec_strategy = fluid.ExecutionStrategy()
+        exec_strategy.num_iteration_per_drop_scope = 10
+        train_exe = fluid.ParallelExecutor(
+            use_cuda=(args.device == 'GPU'),
+            loss_name=avg_cost.name,
+            exec_strategy=exec_strategy,
+            main_program=train_program)
+        test_exe = fluid.ParallelExecutor(
+            use_cuda=(args.device == 'GPU'),
+            main_program=test_program,
+            exec_strategy=exec_strategy,
+            share_vars_from=train_exe)
 
     # resume training if initial model provided.
     if args.init_model_path is not None:
@@ -181,15 +224,24 @@ def train(args):
         trans_splice.TransSplice(5, 5), trans_delay.TransDelay(5)
     ]
 
-    feature_t = fluid.LoDTensor()
-    label_t = fluid.LoDTensor()
+    # bind train_reader
+    train_data_reader = reader.AsyncDataReader(
+        args.train_feature_lst,
+        args.train_label_lst,
+        -1,
+        split_sentence_threshold=1024)
 
-    # validation
-    def test(exe):
-        # If test data not found, return invalid cost and accuracy
-        if not (os.path.exists(args.val_feature_lst) and
-                os.path.exists(args.val_label_lst)):
-            return -1.0, -1.0
+    train_data_reader.set_transformers(ltrans)
+
+    def train_data_provider():
+        for data in train_data_reader.batch_iterator(args.batch_size,
+                                                     args.minimum_batch_size):
+            yield batch_data_to_lod_tensors(args, data, fluid.CPUPlace())
+
+    py_train_reader.decorate_tensor_provider(train_data_provider)
+
+    if (os.path.exists(args.val_feature_lst) and
+            os.path.exists(args.val_label_lst)):
         # test data reader
         test_data_reader = reader.AsyncDataReader(
             args.val_feature_lst,
@@ -197,86 +249,101 @@ def train(args):
             -1,
             split_sentence_threshold=1024)
         test_data_reader.set_transformers(ltrans)
-        test_costs, test_accs = [], []
-        for batch_id, batch_data in enumerate(
-                test_data_reader.batch_iterator(args.batch_size,
-                                                args.minimum_batch_size)):
-            # load_data
-            (features, labels, lod, _) = batch_data
-            features = np.reshape(features, (-1, 11, 3, args.frame_dim))
-            features = np.transpose(features, (0, 2, 1, 3))
-            feature_t.set(features, place)
-            feature_t.set_lod([lod])
-            label_t.set(labels, place)
-            label_t.set_lod([lod])
-
-            cost, acc = exe.run(test_program,
-                                feed={"feature": feature_t,
-                                      "label": label_t},
-                                fetch_list=[avg_cost, accuracy],
-                                return_numpy=False)
-            test_costs.append(lodtensor_to_ndarray(cost)[0])
-            test_accs.append(lodtensor_to_ndarray(acc)[0])
-        return np.mean(test_costs), np.mean(test_accs)
 
-    # train data reader
-    train_data_reader = reader.AsyncDataReader(
-        args.train_feature_lst,
-        args.train_label_lst,
-        -1,
-        split_sentence_threshold=1024)
+        def test_data_provider():
+            for data in test_data_reader.batch_iterator(
+                    args.batch_size, args.minimum_batch_size):
+                yield batch_data_to_lod_tensors(args, data, fluid.CPUPlace())
+
+        py_test_reader.decorate_tensor_provider(test_data_provider)
+
+    # validation
+    def test(exe):
+        # If test data not found, return invalid cost and accuracy
+        if not (os.path.exists(args.val_feature_lst) and
+                os.path.exists(args.val_label_lst)):
+            return -1.0, -1.0
+        batch_id = 0
+        test_costs = []
+        test_accs = []
+        while True:
+            if batch_id == 0:
+                py_test_reader.start()
+            try:
+                if args.parallel:
+                    cost, acc = exe.run(
+                        fetch_list=[avg_cost.name, accuracy.name],
+                        return_numpy=False)
+                else:
+                    cost, acc = exe.run(program=test_program,
+                                        fetch_list=[avg_cost, accuracy],
+                                        return_numpy=False)
+                sys.stdout.write('.')
+                sys.stdout.flush()
+                test_costs.append(np.array(cost)[0])
+                test_accs.append(np.array(acc)[0])
+                batch_id += 1
+            except fluid.core.EOFException:
+                py_test_reader.reset()
+                break
+        return np.mean(test_costs), np.mean(test_accs)
 
-    train_data_reader.set_transformers(ltrans)
     # train
     for pass_id in xrange(args.pass_num):
         pass_start_time = time.time()
-        for batch_id, batch_data in enumerate(
-                train_data_reader.batch_iterator(args.batch_size,
-                                                 args.minimum_batch_size)):
-            # load_data
-            (features, labels, lod, name_lst) = batch_data
-            features = np.reshape(features, (-1, 11, 3, args.frame_dim))
-            features = np.transpose(features, (0, 2, 1, 3))
-            feature_t.set(features, place)
-            feature_t.set_lod([lod])
-            label_t.set(labels, place)
-            label_t.set_lod([lod])
-
+        batch_id = 0
+        while True:
+            if batch_id == 0:
+                py_train_reader.start()
             to_print = batch_id > 0 and (batch_id % args.print_per_batches == 0)
-            outs = exe.run(fluid.default_main_program(),
-                           feed={"feature": feature_t,
-                                 "label": label_t},
-                           fetch_list=[avg_cost, accuracy] if to_print else [],
-                           return_numpy=False)
+            try:
+                if args.parallel:
+                    outs = train_exe.run(
+                        fetch_list=[avg_cost.name, accuracy.name]
+                        if to_print else [],
+                        return_numpy=False)
+                else:
+                    outs = exe.run(program=train_program,
+                                   fetch_list=[avg_cost, accuracy]
+                                   if to_print else [],
+                                   return_numpy=False)
+            except fluid.core.EOFException:
+                py_train_reader.reset()
+                break
 
             if to_print:
-                print("\nBatch %d, train cost: %f, train acc: %f" %
-                      (batch_id, lodtensor_to_ndarray(outs[0])[0],
-                       lodtensor_to_ndarray(outs[1])[0]))
+                if args.parallel:
+                    print("\nBatch %d, train cost: %f, train acc: %f" %
+                          (batch_id, np.mean(outs[0]), np.mean(outs[1])))
+                else:
+                    print("\nBatch %d, train cost: %f, train acc: %f" % (
+                        batch_id, np.array(outs[0])[0], np.array(outs[1])[0]))
                 # save the latest checkpoint
                 if args.checkpoints != '':
                     model_path = os.path.join(args.checkpoints,
                                               "deep_asr.latest.checkpoint")
-                    fluid.io.save_persistables(exe, model_path)
+                    fluid.io.save_persistables(exe, model_path, train_program)
             else:
                 sys.stdout.write('.')
                 sys.stdout.flush()
+
+            batch_id += 1
         # run test
-        val_cost, val_acc = test(exe)
+        val_cost, val_acc = test(test_exe if args.parallel else exe)
 
         # save checkpoint per pass
         if args.checkpoints != '':
             model_path = os.path.join(
                 args.checkpoints,
                 "deep_asr.pass_" + str(pass_id) + ".checkpoint")
-            fluid.io.save_persistables(exe, model_path)
+            fluid.io.save_persistables(exe, model_path, train_program)
         # save inference model
         if args.infer_models != '':
             model_path = os.path.join(
                 args.infer_models,
                 "deep_asr.pass_" + str(pass_id) + ".infer.model")
             fluid.io.save_inference_model(model_path, ["feature"],
-                                          [prediction], exe)
+                                          [prediction], exe, train_program)
         # cal pass time
         pass_end_time = time.time()
         time_consumed = pass_end_time - pass_start_time
@@ -285,6 +352,19 @@ def train(args):
               (pass_id, time_consumed, val_cost, val_acc))
 
 
+def batch_data_to_lod_tensors(args, batch_data, place):
+    features, labels, lod, name_lst = batch_data
+    features = np.reshape(features, (-1, 11, 3, args.frame_dim))
+    features = np.transpose(features, (0, 2, 1, 3))
+    feature_t = fluid.LoDTensor()
+    label_t = fluid.LoDTensor()
+    feature_t.set(features, place)
+    feature_t.set_lod([lod])
+    label_t.set(labels, place)
+    label_t.set_lod([lod])
+    return feature_t, label_t
+
+
 if __name__ == '__main__':
     args = parse_args()
     print_arguments(args)

fluid/PaddleCV/faster_rcnn/README.md

@@ -111,10 +111,10 @@ Evalutaion result is shown as below:
 
 | Model              | RoI function    | Batch size     | Max iteration    | mAP  |
 | :--------------- | :--------: | :------------:    | :------------------:    |------: |
-| [Fluid RoIPool minibatch padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_pool_minibatch_padding.tar.gz) | RoIPool | 8   |    180000        | 0.314 |
-| [Fluid RoIPool no padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_pool_no_padding.tar.gz)  | RoIPool | 8   |    180000        | 0.316 |
-| [Fluid RoIAlign no padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_align_no_padding.tar.gz)  | RoIAlign | 8   |    180000        | 0.345 |
-| [Fluid RoIAlign no padding 2x](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_align_no_padding_2x.tar.gz)  | RoIAlign | 8   |    360000        | 0.364 |
+| [Fluid RoIPool minibatch padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_pool_minibatch_padding.tar.gz) | RoIPool | 8   |    180000        | 0.316 |
+| [Fluid RoIPool no padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_pool_no_padding.tar.gz)  | RoIPool | 8   |    180000        | 0.318 |
+| [Fluid RoIAlign no padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_align_no_padding.tar.gz)  | RoIAlign | 8   |    180000        | 0.348 |
+| [Fluid RoIAlign no padding 2x](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_align_no_padding_2x.tar.gz)  | RoIAlign | 8   |    360000        | 0.367 |
 
 * Fluid RoIPool minibatch padding: Use RoIPool. Images in one batch padding to the same size. This method is same as detectron.
 * Fluid RoIPool no padding: Images without padding.

fluid/PaddleCV/faster_rcnn/README_cn.md

@@ -105,10 +105,10 @@ Faster RCNN 目标检测模型
 
 | 模型                   |   RoI处理方式  | 批量大小   | 迭代次数   | mAP  |
 | :--------------- | :--------: | :------------:    | :------------------:    |------: |
-| [Fluid RoIPool minibatch padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_pool_minibatch_padding.tar.gz) | RoIPool | 8   |    180000        | 0.314 |
-| [Fluid RoIPool no padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_pool_no_padding.tar.gz)  | RoIPool | 8   |    180000        | 0.316 |
-| [Fluid RoIAlign no padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_align_no_padding.tar.gz)  | RoIAlign | 8   |    180000        | 0.345 |
-| [Fluid RoIAlign no padding 2x](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_align_no_padding_2x.tar.gz)  | RoIAlign | 8   |    360000        | 0.364 |
+| [Fluid RoIPool minibatch padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_pool_minibatch_padding.tar.gz) | RoIPool | 8   |    180000        | 0.316 |
+| [Fluid RoIPool no padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_pool_no_padding.tar.gz)  | RoIPool | 8   |    180000        | 0.318 |
+| [Fluid RoIAlign no padding](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_align_no_padding.tar.gz)  | RoIAlign | 8   |    180000        | 0.348 |
+| [Fluid RoIAlign no padding 2x](http://paddlemodels.bj.bcebos.com/faster_rcnn/model_align_no_padding_2x.tar.gz)  | RoIAlign | 8   |    360000        | 0.367 |
 
 
 

fluid/PaddleCV/faster_rcnn/eval_helper.py

@@ -23,29 +23,43 @@ from PIL import ImageFont
 from config import cfg
 
 
-def box_decoder(target_box, prior_box, prior_box_var):
-    proposals = np.zeros_like(target_box, dtype=np.float32)
-    prior_box_loc = np.zeros_like(prior_box, dtype=np.float32)
-    prior_box_loc[:, 0] = prior_box[:, 2] - prior_box[:, 0] + 1.
-    prior_box_loc[:, 1] = prior_box[:, 3] - prior_box[:, 1] + 1.
-    prior_box_loc[:, 2] = (prior_box[:, 2] + prior_box[:, 0]) / 2
-    prior_box_loc[:, 3] = (prior_box[:, 3] + prior_box[:, 1]) / 2
-    pred_bbox = np.zeros_like(target_box, dtype=np.float32)
-    for i in range(prior_box.shape[0]):
-        dw = np.minimum(prior_box_var[2] * target_box[i, 2::4], cfg.bbox_clip)
-        dh = np.minimum(prior_box_var[3] * target_box[i, 3::4], cfg.bbox_clip)
-        pred_bbox[i, 0::4] = prior_box_var[0] * target_box[
-            i, 0::4] * prior_box_loc[i, 0] + prior_box_loc[i, 2]
-        pred_bbox[i, 1::4] = prior_box_var[1] * target_box[
-            i, 1::4] * prior_box_loc[i, 1] + prior_box_loc[i, 3]
-        pred_bbox[i, 2::4] = np.exp(dw) * prior_box_loc[i, 0]
-        pred_bbox[i, 3::4] = np.exp(dh) * prior_box_loc[i, 1]
-    proposals[:, 0::4] = pred_bbox[:, 0::4] - pred_bbox[:, 2::4] / 2
-    proposals[:, 1::4] = pred_bbox[:, 1::4] - pred_bbox[:, 3::4] / 2
-    proposals[:, 2::4] = pred_bbox[:, 0::4] + pred_bbox[:, 2::4] / 2 - 1
-    proposals[:, 3::4] = pred_bbox[:, 1::4] + pred_bbox[:, 3::4] / 2 - 1
-
-    return proposals
+def box_decoder(deltas, boxes, weights):
+    if boxes.shape[0] == 0:
+        return np.zeros((0, deltas.shape[1]), dtype=deltas.dtype)
+
+    boxes = boxes.astype(deltas.dtype, copy=False)
+
+    widths = boxes[:, 2] - boxes[:, 0] + 1.0
+    heights = boxes[:, 3] - boxes[:, 1] + 1.0
+    ctr_x = boxes[:, 0] + 0.5 * widths
+    ctr_y = boxes[:, 1] + 0.5 * heights
+
+    wx, wy, ww, wh = weights
+    dx = deltas[:, 0::4] * wx
+    dy = deltas[:, 1::4] * wy
+    dw = deltas[:, 2::4] * ww
+    dh = deltas[:, 3::4] * wh
+
+    # Prevent sending too large values into np.exp()
+    dw = np.minimum(dw, cfg.bbox_clip)
+    dh = np.minimum(dh, cfg.bbox_clip)
+
+    pred_ctr_x = dx * widths[:, np.newaxis] + ctr_x[:, np.newaxis]
+    pred_ctr_y = dy * heights[:, np.newaxis] + ctr_y[:, np.newaxis]
+    pred_w = np.exp(dw) * widths[:, np.newaxis]
+    pred_h = np.exp(dh) * heights[:, np.newaxis]
+
+    pred_boxes = np.zeros(deltas.shape, dtype=deltas.dtype)
+    # x1
+    pred_boxes[:, 0::4] = pred_ctr_x - 0.5 * pred_w
+    # y1
+    pred_boxes[:, 1::4] = pred_ctr_y - 0.5 * pred_h
+    # x2 (note: "- 1" is correct; don't be fooled by the asymmetry)
+    pred_boxes[:, 2::4] = pred_ctr_x + 0.5 * pred_w - 1
+    # y2 (note: "- 1" is correct; don't be fooled by the asymmetry)
+    pred_boxes[:, 3::4] = pred_ctr_y + 0.5 * pred_h - 1
+
+    return pred_boxes
 
 
 def clip_tiled_boxes(boxes, im_shape):
@@ -73,7 +87,6 @@ def get_nmsed_box(rpn_rois, confs, locs, class_nums, im_info,
     variance_v = np.array(cfg.bbox_reg_weights)
     confs_v = np.array(confs)
     locs_v = np.array(locs)
-    rois = box_decoder(locs_v, rpn_rois_v, variance_v)
     im_results = [[] for _ in range(len(lod) - 1)]
     new_lod = [0]
     for i in range(len(lod) - 1):
@@ -81,9 +94,11 @@ def get_nmsed_box(rpn_rois, confs, locs, class_nums, im_info,
         end = lod[i + 1]
         if start == end:
             continue
-        rois_n = rois[start:end, :]
+        locs_n = locs_v[start:end, :]
+        rois_n = rpn_rois_v[start:end, :]
         rois_n = rois_n / im_info[i][2]
-        rois_n = clip_tiled_boxes(rois_n, im_info[i][:2])
+        rois_n = box_decoder(locs_n, rois_n, variance_v)
+        rois_n = clip_tiled_boxes(rois_n, im_info[i][:2] / im_info[i][2])
 
         cls_boxes = [[] for _ in range(class_nums)]
         scores_n = confs_v[start:end, :]

fluid/PaddleCV/image_classification/README.md

@@ -59,7 +59,7 @@ python train.py \
        --model=SE_ResNeXt50_32x4d \
        --batch_size=32 \
        --total_images=1281167 \
-       --class_dim=1000
+       --class_dim=1000 \
        --image_shape=3,224,224 \
        --model_save_dir=output/ \
        --with_mem_opt=False \
@@ -80,6 +80,9 @@ python train.py \
 * **lr**: initialized learning rate. Default: 0.1.
 * **pretrained_model**: model path for pretraining. Default: None.
 * **checkpoint**: the checkpoint path to resume. Default: None.
+* **model_category**: the category of models, ("models"|"models_name"). Default: "models".
+
+Or can start the training step by running the ```run.sh```.
 
 **data reader introduction:** Data reader is defined in ```reader.py```. In [training stage](#training-a-model), random crop and flipping are used, while center crop is used in [evaluation](#inference) and [inference](#inference) stages. Supported data augmentation includes:
 * rotation
@@ -183,26 +186,23 @@ Test-12-score: [15.040644], class [386]
 
 ## Supported models and performances
 
+Models consists of two categories: Models with specified parameters names in model definition and Models without specified parameters, Generate named model by indicating ```model_category = models_name```.
+
 Models are trained by starting with learning rate ```0.1``` and decaying it by ```0.1``` after each pre-defined epoches, if not special introduced. Available top-1/top-5 validation accuracy on ImageNet 2012 are listed in table. Pretrained models can be downloaded by clicking related model names.
 
+- Released models: specify parameter names
+
+|model | top-1/top-5 accuracy
+|- | -:
+|[AlexNet](http://paddle-imagenet-models-name.bj.bcebos.com/AlexNet_pretrained.zip) | 56.34%/79.02%
+|[VGG11](http://paddle-imagenet-models-name.bj.bcebos.com/VGG11_pretained.zip) | 68.86%/88.64%
+|[MobileNetV1](http://paddle-imagenet-models-name.bj.bcebos.com/MobileNetV1_pretrained.zip) | 70.7%/89.41%
+|[ResNet50](http://paddle-imagenet-models-name.bj.bcebos.com/ResNet50_pretrained.zip) | 76.46%/93.04%
+|[ResNet101](http://paddle-imagenet-models-name.bj.bcebos.com/ResNet101_pretrained.zip) | 77.65%/93.71%
+
+- Released models: not specify parameter names
+
 |model | top-1/top-5 accuracy
 |- | -:
-|[AlexNet](http://paddle-imagenet-models.bj.bcebos.com/alexnet_model.tar) | 57.21%/79.72%
-|VGG11 | -
-|VGG13 | -
-|VGG16 | -
-|VGG19 | -
-|GoogleNet | -
-|InceptionV4 | -
-|MobileNet | -
-|[ResNet50](http://paddle-imagenet-models.bj.bcebos.com/resnet_50_model.tar) | 76.63%/93.10%
-|ResNet101 | -
-|ResNet152 | -
+|[ResNet152](http://paddle-imagenet-models.bj.bcebos.com/ResNet152_pretrained.zip) | 78.29%/94.11%
 |[SE_ResNeXt50_32x4d](http://paddle-imagenet-models.bj.bcebos.com/se_resnext_50_model.tar) | 78.33%/93.96%
-|SE_ResNeXt101_32x4d | -
-|SE_ResNeXt152_32x4d | -
-|DPN68 | -
-|DPN92 | -
-|DPN98 | -
-|DPN107 | -
-|DPN131 | -

fluid/PaddleCV/image_classification/models/__init__.py

 from .alexnet import AlexNet
 from .mobilenet import MobileNet
+from .mobilenet_v2 import MobileNetV2
 from .googlenet import GoogleNet
 from .vgg import VGG11, VGG13, VGG16, VGG19
 from .resnet import ResNet50, ResNet101, ResNet152
@@ -7,4 +8,4 @@ from .resnet_dist import DistResNet
 from .inception_v4 import InceptionV4
 from .se_resnext import SE_ResNeXt50_32x4d, SE_ResNeXt101_32x4d, SE_ResNeXt152_32x4d
 from .dpn import DPN68, DPN92, DPN98, DPN107, DPN131
-import learning_rate
+from .shufflenet_v2 import ShuffleNetV2_x0_5, ShuffleNetV2_x1_0, ShuffleNetV2_x1_5, ShuffleNetV2_x2_0

fluid/PaddleCV/image_classification/models/dpn.py

@@ -5,8 +5,8 @@ import os
 import numpy as np
 import time
 import sys
-import math
 import paddle.fluid as fluid
+import math
 
 __all__ = ["DPN", "DPN68", "DPN92", "DPN98", "DPN107", "DPN131"]
 
@@ -62,7 +62,6 @@ class DPN(object):
             pool_padding=1,
             pool_type='max')
 
-        #conv2 - conv5
         for gc in range(4):
             bw = bws[gc]
             inc = inc_sec[gc]

fluid/PaddleCV/image_classification/models/googlenet.py

@@ -13,7 +13,7 @@ train_parameters = {
     "learning_strategy": {
         "name": "piecewise_decay",
         "batch_size": 256,
-        "epochs": [30, 60, 90],
+        "epochs": [30, 70, 100],
         "steps": [0.1, 0.01, 0.001, 0.0001]
     }
 }

fluid/PaddleCV/image_classification/models/mobilenet_v2.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle.fluid as fluid
+from paddle.fluid.initializer import MSRA
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = ['MobileNetV2']
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 60, 90],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class MobileNetV2():
+    def __init__(self):
+        self.params = train_parameters
+
+    def net(self, input, class_dim=1000, scale=1.0):
+
+        bottleneck_params_list = [
+            (1, 16, 1, 1),
+            (6, 24, 2, 2),
+            (6, 32, 3, 2),
+            (6, 64, 4, 2),
+            (6, 96, 3, 1),
+            (6, 160, 3, 2),
+            (6, 320, 1, 1),
+        ]
+
+        input = self.conv_bn_layer(
+            input,
+            num_filters=int(32 * scale),
+            filter_size=3,
+            stride=2,
+            padding=1,
+            if_act=True)
+
+        in_c = int(32 * scale)
+        for layer_setting in bottleneck_params_list:
+            t, c, n, s = layer_setting
+            input = self.invresi_blocks(
+                input=input,
+                in_c=in_c,
+                t=t,
+                c=int(c * scale),
+                n=n,
+                s=s, )
+            in_c = int(c * scale)
+
+        input = self.conv_bn_layer(
+            input=input,
+            num_filters=int(1280 * scale) if scale > 1.0 else 1280,
+            filter_size=1,
+            stride=1,
+            padding=0,
+            if_act=True)
+
+        input = fluid.layers.pool2d(
+            input=input,
+            pool_size=7,
+            pool_stride=1,
+            pool_type='avg',
+            global_pooling=True)
+
+        output = fluid.layers.fc(input=input,
+                                 size=class_dim,
+                                 act='softmax',
+                                 param_attr=ParamAttr(initializer=MSRA()))
+        return output
+
+    def conv_bn_layer(self,
+                      input,
+                      filter_size,
+                      num_filters,
+                      stride,
+                      padding,
+                      channels=None,
+                      num_groups=1,
+                      use_cudnn=True,
+                      if_act=True):
+        conv = fluid.layers.conv2d(
+            input=input,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=padding,
+            groups=num_groups,
+            act=None,
+            use_cudnn=use_cudnn,
+            param_attr=ParamAttr(initializer=MSRA()),
+            bias_attr=False)
+        bn = fluid.layers.batch_norm(input=conv)
+        if if_act:
+            return fluid.layers.relu6(bn)
+        else:
+            return bn
+
+    def shortcut(self, input, data_residual):
+        return fluid.layers.elementwise_add(input, data_residual)
+
+    def inverted_residual_unit(self, input, num_in_filter, num_filters,
+                               ifshortcut, stride, filter_size, padding,
+                               expansion_factor):
+        num_expfilter = int(round(num_in_filter * expansion_factor))
+        channel_expand = self.conv_bn_layer(
+            input=input,
+            num_filters=num_expfilter,
+            filter_size=1,
+            stride=1,
+            padding=0,
+            num_groups=1,
+            if_act=True)
+        bottleneck_conv = self.conv_bn_layer(
+            input=channel_expand,
+            num_filters=num_expfilter,
+            filter_size=filter_size,
+            stride=stride,
+            padding=padding,
+            num_groups=num_expfilter,
+            if_act=True,
+            use_cudnn=False)
+        linear_out = self.conv_bn_layer(
+            input=bottleneck_conv,
+            num_filters=num_filters,
+            filter_size=1,
+            stride=1,
+            padding=0,
+            num_groups=1,
+            if_act=False)
+        if ifshortcut:
+            out = self.shortcut(input=input, data_residual=linear_out)
+            return out
+        else:
+            return linear_out
+
+    def invresi_blocks(self, input, in_c, t, c, n, s):
+        first_block = self.inverted_residual_unit(
+            input=input,
+            num_in_filter=in_c,
+            num_filters=c,
+            ifshortcut=False,
+            stride=s,
+            filter_size=3,
+            padding=1,
+            expansion_factor=t)
+
+        last_residual_block = first_block
+        last_c = c
+
+        for i in range(1, n):
+            last_residual_block = self.inverted_residual_unit(
+                input=last_residual_block,
+                num_in_filter=last_c,
+                num_filters=c,
+                ifshortcut=True,
+                stride=1,
+                filter_size=3,
+                padding=1,
+                expansion_factor=t)
+        return last_residual_block

fluid/PaddleCV/image_classification/models/shufflenet_v2.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle.fluid as fluid
+from paddle.fluid.initializer import MSRA
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = [
+    'ShuffleNetV2', 'ShuffleNetV2_x0_5', 'ShuffleNetV2_x1_0',
+    'ShuffleNetV2_x1_5', 'ShuffleNetV2_x2_0'
+]
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 60, 90],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class ShuffleNetV2():
+    def __init__(self, scale=1.0):
+        self.params = train_parameters
+        self.scale = scale
+
+    def net(self, input, class_dim=1000):
+        scale = self.scale
+        stage_repeats = [4, 8, 4]
+
+        if scale == 0.5:
+            stage_out_channels = [-1, 24, 48, 96, 192, 1024]
+        elif scale == 1.0:
+            stage_out_channels = [-1, 24, 116, 232, 464, 1024]
+        elif scale == 1.5:
+            stage_out_channels = [-1, 24, 176, 352, 704, 1024]
+        elif scale == 2.0:
+            stage_out_channels = [-1, 24, 224, 488, 976, 2048]
+        else:
+            raise ValueError("""{} groups is not supported for
+                       1x1 Grouped Convolutions""".format(num_groups))
+
+        #conv1
+
+        input_channel = stage_out_channels[1]
+        conv1 = self.conv_bn_layer(
+            input=input,
+            filter_size=3,
+            num_filters=input_channel,
+            padding=1,
+            stride=2)
+        pool1 = fluid.layers.pool2d(
+            input=conv1,
+            pool_size=3,
+            pool_stride=2,
+            pool_padding=1,
+            pool_type='max')
+        conv = pool1
+        # bottleneck sequences
+        for idxstage in range(len(stage_repeats)):
+            numrepeat = stage_repeats[idxstage]
+            output_channel = stage_out_channels[idxstage + 2]
+            for i in range(numrepeat):
+                if i == 0:
+                    conv = self.inverted_residual_unit(
+                        input=conv,
+                        num_filters=output_channel,
+                        stride=2,
+                        benchmodel=2)
+                else:
+                    conv = self.inverted_residual_unit(
+                        input=conv,
+                        num_filters=output_channel,
+                        stride=1,
+                        benchmodel=1)
+
+        conv_last = self.conv_bn_layer(
+            input=conv,
+            filter_size=1,
+            num_filters=stage_out_channels[-1],
+            padding=0,
+            stride=1)
+        pool_last = fluid.layers.pool2d(
+            input=conv_last,
+            pool_size=7,
+            pool_stride=7,
+            pool_padding=0,
+            pool_type='avg')
+
+        output = fluid.layers.fc(input=pool_last,
+                                 size=class_dim,
+                                 act='softmax',
+                                 param_attr=ParamAttr(initializer=MSRA()))
+        return output
+
+    def conv_bn_layer(self,
+                      input,
+                      filter_size,
+                      num_filters,
+                      stride,
+                      padding,
+                      num_groups=1,
+                      use_cudnn=True,
+                      if_act=True):
+        conv = fluid.layers.conv2d(
+            input=input,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=padding,
+            groups=num_groups,
+            act=None,
+            use_cudnn=use_cudnn,
+            param_attr=ParamAttr(initializer=MSRA()),
+            bias_attr=False)
+        if if_act:
+            return fluid.layers.batch_norm(input=conv, act='relu')
+        else:
+            return fluid.layers.batch_norm(input=conv)
+
+    def channel_shuffle(self, x, groups):
+        batchsize, num_channels, height, width = x.shape[0], x.shape[
+            1], x.shape[2], x.shape[3]
+        channels_per_group = num_channels // groups
+
+        # reshape
+        x = fluid.layers.reshape(
+            x=x, shape=[batchsize, groups, channels_per_group, height, width])
+
+        x = fluid.layers.transpose(x=x, perm=[0, 2, 1, 3, 4])
+
+        # flatten
+        x = fluid.layers.reshape(
+            x=x, shape=[batchsize, num_channels, height, width])
+
+        return x
+
+    def inverted_residual_unit(self, input, num_filters, stride, benchmodel):
+        assert stride in [1, 2], \
+            "supported stride are {} but your stride is {}".format([1,2], stride)
+
+        oup_inc = num_filters // 2
+        inp = input.shape[1]
+
+        if benchmodel == 1:
+            x1, x2 = fluid.layers.split(
+                input,
+                num_or_sections=[input.shape[1] // 2, input.shape[1] // 2],
+                dim=1)
+
+            conv_pw = self.conv_bn_layer(
+                input=x2,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True)
+
+            conv_dw = self.conv_bn_layer(
+                input=conv_pw,
+                num_filters=oup_inc,
+                filter_size=3,
+                stride=stride,
+                padding=1,
+                num_groups=oup_inc,
+                if_act=False)
+
+            conv_linear = self.conv_bn_layer(
+                input=conv_dw,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True)
+
+            out = fluid.layers.concat([x1, conv_linear], axis=1)
+
+        else:
+            #branch1
+            conv_dw = self.conv_bn_layer(
+                input=input,
+                num_filters=inp,
+                filter_size=3,
+                stride=stride,
+                padding=1,
+                num_groups=inp,
+                if_act=False)
+
+            conv_linear_1 = self.conv_bn_layer(
+                input=conv_dw,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True)
+
+            #branch2
+            conv_pw = self.conv_bn_layer(
+                input=input,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True)
+
+            conv_dw = self.conv_bn_layer(
+                input=conv_pw,
+                num_filters=oup_inc,
+                filter_size=3,
+                stride=stride,
+                padding=1,
+                num_groups=oup_inc,
+                if_act=False)
+
+            conv_linear_2 = self.conv_bn_layer(
+                input=conv_dw,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True)
+            out = fluid.layers.concat([conv_linear_1, conv_linear_2], axis=1)
+
+        return self.channel_shuffle(out, 2)
+
+
+def ShuffleNetV2_x0_5():
+    model = ShuffleNetV2(scale=0.5)
+    return model
+
+
+def ShuffleNetV2_x1_0():
+    model = ShuffleNetV2(scale=1.0)
+    return model
+
+
+def ShuffleNetV2_x1_5():
+    model = ShuffleNetV2(scale=1.5)
+    return model
+
+
+def ShuffleNetV2_x2_0():
+    model = ShuffleNetV2(scale=2.0)
+    return model

fluid/PaddleCV/image_classification/models_name/__init__.py

+from .alexnet import AlexNet
+from .mobilenet import MobileNet
+from .mobilenet_v2 import MobileNetV2
+from .googlenet import GoogleNet
+from .vgg import VGG11, VGG13, VGG16, VGG19
+from .resnet import ResNet50, ResNet101, ResNet152
+from .inception_v4 import InceptionV4
+from .se_resnext import SE_ResNeXt50_32x4d, SE_ResNeXt101_32x4d, SE_ResNeXt152_32x4d
+from .dpn import DPN68, DPN92, DPN98, DPN107, DPN131
+from .shufflenet_v2 import ShuffleNetV2_x0_5, ShuffleNetV2_x1_0, ShuffleNetV2_x1_5, ShuffleNetV2_x2_0

fluid/PaddleCV/image_classification/models_name/alexnet.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle
+import paddle.fluid as fluid
+import math
+
+__all__ = ['AlexNet']
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [40, 70, 100],
+        "steps": [0.01, 0.001, 0.0001, 0.00001]
+    }
+}
+
+
+class AlexNet():
+    def __init__(self):
+        self.params = train_parameters
+
+    def net(self, input, class_dim=1000):
+        stdv = 1.0 / math.sqrt(input.shape[1] * 11 * 11)
+        layer_name = [
+            "conv1", "conv2", "conv3", "conv4", "conv5", "fc6", "fc7", "fc8"
+        ]
+        conv1 = fluid.layers.conv2d(
+            input=input,
+            num_filters=64,
+            filter_size=11,
+            stride=4,
+            padding=2,
+            groups=1,
+            act='relu',
+            bias_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[0] + "_offset"),
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[0] + "_weights"))
+        pool1 = fluid.layers.pool2d(
+            input=conv1,
+            pool_size=3,
+            pool_stride=2,
+            pool_padding=0,
+            pool_type='max')
+
+        stdv = 1.0 / math.sqrt(pool1.shape[1] * 5 * 5)
+        conv2 = fluid.layers.conv2d(
+            input=pool1,
+            num_filters=192,
+            filter_size=5,
+            stride=1,
+            padding=2,
+            groups=1,
+            act='relu',
+            bias_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[1] + "_offset"),
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[1] + "_weights"))
+        pool2 = fluid.layers.pool2d(
+            input=conv2,
+            pool_size=3,
+            pool_stride=2,
+            pool_padding=0,
+            pool_type='max')
+
+        stdv = 1.0 / math.sqrt(pool2.shape[1] * 3 * 3)
+        conv3 = fluid.layers.conv2d(
+            input=pool2,
+            num_filters=384,
+            filter_size=3,
+            stride=1,
+            padding=1,
+            groups=1,
+            act='relu',
+            bias_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[2] + "_offset"),
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[2] + "_weights"))
+
+        stdv = 1.0 / math.sqrt(conv3.shape[1] * 3 * 3)
+        conv4 = fluid.layers.conv2d(
+            input=conv3,
+            num_filters=256,
+            filter_size=3,
+            stride=1,
+            padding=1,
+            groups=1,
+            act='relu',
+            bias_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[3] + "_offset"),
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[3] + "_weights"))
+
+        stdv = 1.0 / math.sqrt(conv4.shape[1] * 3 * 3)
+        conv5 = fluid.layers.conv2d(
+            input=conv4,
+            num_filters=256,
+            filter_size=3,
+            stride=1,
+            padding=1,
+            groups=1,
+            act='relu',
+            bias_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[4] + "_offset"),
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[4] + "_weights"))
+        pool5 = fluid.layers.pool2d(
+            input=conv5,
+            pool_size=3,
+            pool_stride=2,
+            pool_padding=0,
+            pool_type='max')
+
+        drop6 = fluid.layers.dropout(x=pool5, dropout_prob=0.5)
+        stdv = 1.0 / math.sqrt(drop6.shape[1] * drop6.shape[2] *
+                               drop6.shape[3] * 1.0)
+
+        fc6 = fluid.layers.fc(
+            input=drop6,
+            size=4096,
+            act='relu',
+            bias_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[5] + "_offset"),
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[5] + "_weights"))
+
+        drop7 = fluid.layers.dropout(x=fc6, dropout_prob=0.5)
+        stdv = 1.0 / math.sqrt(drop7.shape[1] * 1.0)
+
+        fc7 = fluid.layers.fc(
+            input=drop7,
+            size=4096,
+            act='relu',
+            bias_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[6] + "_offset"),
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[6] + "_weights"))
+
+        stdv = 1.0 / math.sqrt(fc7.shape[1] * 1.0)
+        out = fluid.layers.fc(
+            input=fc7,
+            size=class_dim,
+            act='softmax',
+            bias_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[7] + "_offset"),
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=layer_name[7] + "_weights"))
+        return out

fluid/PaddleCV/image_classification/models_name/dpn.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import os
+import numpy as np
+import time
+import sys
+import paddle.fluid as fluid
+import math
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = ["DPN", "DPN68", "DPN92", "DPN98", "DPN107", "DPN131"]
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 60, 90],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class DPN(object):
+    def __init__(self, layers=68):
+        self.params = train_parameters
+        self.layers = layers
+
+    def net(self, input, class_dim=1000):
+        # get network args
+        args = self.get_net_args(self.layers)
+        bws = args['bw']
+        inc_sec = args['inc_sec']
+        rs = args['bw']
+        k_r = args['k_r']
+        k_sec = args['k_sec']
+        G = args['G']
+        init_num_filter = args['init_num_filter']
+        init_filter_size = args['init_filter_size']
+        init_padding = args['init_padding']
+
+        ## define Dual Path Network
+
+        # conv1
+        conv1_x_1 = fluid.layers.conv2d(
+            input=input,
+            num_filters=init_num_filter,
+            filter_size=init_filter_size,
+            stride=2,
+            padding=init_padding,
+            groups=1,
+            act=None,
+            bias_attr=False,
+            name="conv1",
+            param_attr=ParamAttr(name="conv1_weights"), )
+
+        conv1_x_1 = fluid.layers.batch_norm(
+            input=conv1_x_1,
+            act='relu',
+            is_test=False,
+            name="conv1_bn",
+            param_attr=ParamAttr(name='conv1_bn_scale'),
+            bias_attr=ParamAttr('conv1_bn_offset'),
+            moving_mean_name='conv1_bn_mean',
+            moving_variance_name='conv1_bn_variance', )
+
+        convX_x_x = fluid.layers.pool2d(
+            input=conv1_x_1,
+            pool_size=3,
+            pool_stride=2,
+            pool_padding=1,
+            pool_type='max',
+            name="pool1")
+
+        #conv2 - conv5
+        match_list, num = [], 0
+        for gc in range(4):
+            bw = bws[gc]
+            inc = inc_sec[gc]
+            R = (k_r * bw) // rs[gc]
+            if gc == 0:
+                _type1 = 'proj'
+                _type2 = 'normal'
+                match = 1
+            else:
+                _type1 = 'down'
+                _type2 = 'normal'
+                match = match + k_sec[gc - 1]
+            match_list.append(match)
+
+            convX_x_x = self.dual_path_factory(
+                convX_x_x, R, R, bw, inc, G, _type1, name="dpn" + str(match))
+            for i_ly in range(2, k_sec[gc] + 1):
+                num += 1
+                if num in match_list:
+                    num += 1
+                convX_x_x = self.dual_path_factory(
+                    convX_x_x, R, R, bw, inc, G, _type2, name="dpn" + str(num))
+
+        conv5_x_x = fluid.layers.concat(convX_x_x, axis=1)
+        conv5_x_x = fluid.layers.batch_norm(
+            input=conv5_x_x,
+            act='relu',
+            is_test=False,
+            name="final_concat_bn",
+            param_attr=ParamAttr(name='final_concat_bn_scale'),
+            bias_attr=ParamAttr('final_concat_bn_offset'),
+            moving_mean_name='final_concat_bn_mean',
+            moving_variance_name='final_concat_bn_variance', )
+        pool5 = fluid.layers.pool2d(
+            input=conv5_x_x,
+            pool_size=7,
+            pool_stride=1,
+            pool_padding=0,
+            pool_type='avg', )
+
+        stdv = 0.01
+        param_attr = fluid.param_attr.ParamAttr(
+            initializer=fluid.initializer.Uniform(-stdv, stdv))
+        fc6 = fluid.layers.fc(input=pool5,
+                              size=class_dim,
+                              act='softmax',
+                              param_attr=param_attr,
+                              name="fc6")
+
+        return fc6
+
+    def get_net_args(self, layers):
+        if layers == 68:
+            k_r = 128
+            G = 32
+            k_sec = [3, 4, 12, 3]
+            inc_sec = [16, 32, 32, 64]
+            bw = [64, 128, 256, 512]
+            r = [64, 64, 64, 64]
+            init_num_filter = 10
+            init_filter_size = 3
+            init_padding = 1
+        elif layers == 92:
+            k_r = 96
+            G = 32
+            k_sec = [3, 4, 20, 3]
+            inc_sec = [16, 32, 24, 128]
+            bw = [256, 512, 1024, 2048]
+            r = [256, 256, 256, 256]
+            init_num_filter = 64
+            init_filter_size = 7
+            init_padding = 3
+        elif layers == 98:
+            k_r = 160
+            G = 40
+            k_sec = [3, 6, 20, 3]
+            inc_sec = [16, 32, 32, 128]
+            bw = [256, 512, 1024, 2048]
+            r = [256, 256, 256, 256]
+            init_num_filter = 96
+            init_filter_size = 7
+            init_padding = 3
+        elif layers == 107:
+            k_r = 200
+            G = 50
+            k_sec = [4, 8, 20, 3]
+            inc_sec = [20, 64, 64, 128]
+            bw = [256, 512, 1024, 2048]
+            r = [256, 256, 256, 256]
+            init_num_filter = 128
+            init_filter_size = 7
+            init_padding = 3
+        elif layers == 131:
+            k_r = 160
+            G = 40
+            k_sec = [4, 8, 28, 3]
+            inc_sec = [16, 32, 32, 128]
+            bw = [256, 512, 1024, 2048]
+            r = [256, 256, 256, 256]
+            init_num_filter = 128
+            init_filter_size = 7
+            init_padding = 3
+        else:
+            raise NotImplementedError
+        net_arg = {
+            'k_r': k_r,
+            'G': G,
+            'k_sec': k_sec,
+            'inc_sec': inc_sec,
+            'bw': bw,
+            'r': r
+        }
+        net_arg['init_num_filter'] = init_num_filter
+        net_arg['init_filter_size'] = init_filter_size
+        net_arg['init_padding'] = init_padding
+
+        return net_arg
+
+    def dual_path_factory(self,
+                          data,
+                          num_1x1_a,
+                          num_3x3_b,
+                          num_1x1_c,
+                          inc,
+                          G,
+                          _type='normal',
+                          name=None):
+        kw = 3
+        kh = 3
+        pw = (kw - 1) // 2
+        ph = (kh - 1) // 2
+
+        # type
+        if _type is 'proj':
+            key_stride = 1
+            has_proj = True
+        if _type is 'down':
+            key_stride = 2
+            has_proj = True
+        if _type is 'normal':
+            key_stride = 1
+            has_proj = False
+
+        # PROJ
+        if type(data) is list:
+            data_in = fluid.layers.concat([data[0], data[1]], axis=1)
+        else:
+            data_in = data
+
+        if has_proj:
+            c1x1_w = self.bn_ac_conv(
+                data=data_in,
+                num_filter=(num_1x1_c + 2 * inc),
+                kernel=(1, 1),
+                pad=(0, 0),
+                stride=(key_stride, key_stride),
+                name=name + "_match")
+            data_o1, data_o2 = fluid.layers.split(
+                c1x1_w,
+                num_or_sections=[num_1x1_c, 2 * inc],
+                dim=1,
+                name=name + "_match_conv_Slice")
+        else:
+            data_o1 = data[0]
+            data_o2 = data[1]
+
+        # MAIN
+        c1x1_a = self.bn_ac_conv(
+            data=data_in,
+            num_filter=num_1x1_a,
+            kernel=(1, 1),
+            pad=(0, 0),
+            name=name + "_conv1")
+        c3x3_b = self.bn_ac_conv(
+            data=c1x1_a,
+            num_filter=num_3x3_b,
+            kernel=(kw, kh),
+            pad=(pw, ph),
+            stride=(key_stride, key_stride),
+            num_group=G,
+            name=name + "_conv2")
+        c1x1_c = self.bn_ac_conv(
+            data=c3x3_b,
+            num_filter=(num_1x1_c + inc),
+            kernel=(1, 1),
+            pad=(0, 0),
+            name=name + "_conv3")
+
+        c1x1_c1, c1x1_c2 = fluid.layers.split(
+            c1x1_c,
+            num_or_sections=[num_1x1_c, inc],
+            dim=1,
+            name=name + "_conv3_Slice")
+
+        # OUTPUTS
+        summ = fluid.layers.elementwise_add(
+            x=data_o1, y=c1x1_c1, name=name + "_elewise")
+        dense = fluid.layers.concat(
+            [data_o2, c1x1_c2], axis=1, name=name + "_concat")
+
+        return [summ, dense]
+
+    def bn_ac_conv(self,
+                   data,
+                   num_filter,
+                   kernel,
+                   pad,
+                   stride=(1, 1),
+                   num_group=1,
+                   name=None):
+        bn_ac = fluid.layers.batch_norm(
+            input=data,
+            act='relu',
+            is_test=False,
+            name=name + '.output.1',
+            param_attr=ParamAttr(name=name + '_bn_scale'),
+            bias_attr=ParamAttr(name + '_bn_offset'),
+            moving_mean_name=name + '_bn_mean',
+            moving_variance_name=name + '_bn_variance', )
+        bn_ac_conv = fluid.layers.conv2d(
+            input=bn_ac,
+            num_filters=num_filter,
+            filter_size=kernel,
+            stride=stride,
+            padding=pad,
+            groups=num_group,
+            act=None,
+            bias_attr=False,
+            param_attr=ParamAttr(name=name + "_weights"))
+        return bn_ac_conv
+
+
+def DPN68():
+    model = DPN(layers=68)
+    return model
+
+
+def DPN92():
+    onvodel = DPN(layers=92)
+    return model
+
+
+def DPN98():
+    model = DPN(layers=98)
+    return model
+
+
+def DPN107():
+    model = DPN(layers=107)
+    return model
+
+
+def DPN131():
+    model = DPN(layers=131)
+    return model

fluid/PaddleCV/image_classification/models_name/googlenet.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle
+import paddle.fluid as fluid
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = ['GoogleNet']
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 70, 100],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class GoogleNet():
+    def __init__(self):
+        self.params = train_parameters
+
+    def conv_layer(self,
+                   input,
+                   num_filters,
+                   filter_size,
+                   stride=1,
+                   groups=1,
+                   act=None,
+                   name=None):
+        channels = input.shape[1]
+        stdv = (3.0 / (filter_size**2 * channels))**0.5
+        param_attr = ParamAttr(
+            initializer=fluid.initializer.Uniform(-stdv, stdv),
+            name=name + "_weights")
+        conv = fluid.layers.conv2d(
+            input=input,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=(filter_size - 1) // 2,
+            groups=groups,
+            act=act,
+            param_attr=param_attr,
+            bias_attr=False,
+            name=name)
+        return conv
+
+    def xavier(self, channels, filter_size, name):
+        stdv = (3.0 / (filter_size**2 * channels))**0.5
+        param_attr = ParamAttr(
+            initializer=fluid.initializer.Uniform(-stdv, stdv),
+            name=name + "_weights")
+
+        return param_attr
+
+    def inception(self,
+                  input,
+                  channels,
+                  filter1,
+                  filter3R,
+                  filter3,
+                  filter5R,
+                  filter5,
+                  proj,
+                  name=None):
+        conv1 = self.conv_layer(
+            input=input,
+            num_filters=filter1,
+            filter_size=1,
+            stride=1,
+            act=None,
+            name="inception_" + name + "_1x1")
+        conv3r = self.conv_layer(
+            input=input,
+            num_filters=filter3R,
+            filter_size=1,
+            stride=1,
+            act=None,
+            name="inception_" + name + "_3x3_reduce")
+        conv3 = self.conv_layer(
+            input=conv3r,
+            num_filters=filter3,
+            filter_size=3,
+            stride=1,
+            act=None,
+            name="inception_" + name + "_3x3")
+        conv5r = self.conv_layer(
+            input=input,
+            num_filters=filter5R,
+            filter_size=1,
+            stride=1,
+            act=None,
+            name="inception_" + name + "_5x5_reduce")
+        conv5 = self.conv_layer(
+            input=conv5r,
+            num_filters=filter5,
+            filter_size=5,
+            stride=1,
+            act=None,
+            name="inception_" + name + "_5x5")
+        pool = fluid.layers.pool2d(
+            input=input,
+            pool_size=3,
+            pool_stride=1,
+            pool_padding=1,
+            pool_type='max')
+        convprj = fluid.layers.conv2d(
+            input=pool,
+            filter_size=1,
+            num_filters=proj,
+            stride=1,
+            padding=0,
+            name="inception_" + name + "_3x3_proj",
+            param_attr=ParamAttr(
+                name="inception_" + name + "_3x3_proj_weights"),
+            bias_attr=False)
+        cat = fluid.layers.concat(input=[conv1, conv3, conv5, convprj], axis=1)
+        cat = fluid.layers.relu(cat)
+        return cat
+
+    def net(self, input, class_dim=1000):
+        conv = self.conv_layer(
+            input=input,
+            num_filters=64,
+            filter_size=7,
+            stride=2,
+            act=None,
+            name="conv1")
+        pool = fluid.layers.pool2d(
+            input=conv, pool_size=3, pool_type='max', pool_stride=2)
+
+        conv = self.conv_layer(
+            input=pool,
+            num_filters=64,
+            filter_size=1,
+            stride=1,
+            act=None,
+            name="conv2_1x1")
+        conv = self.conv_layer(
+            input=conv,
+            num_filters=192,
+            filter_size=3,
+            stride=1,
+            act=None,
+            name="conv2_3x3")
+        pool = fluid.layers.pool2d(
+            input=conv, pool_size=3, pool_type='max', pool_stride=2)
+
+        ince3a = self.inception(pool, 192, 64, 96, 128, 16, 32, 32, "ince3a")
+        ince3b = self.inception(ince3a, 256, 128, 128, 192, 32, 96, 64,
+                                "ince3b")
+        pool3 = fluid.layers.pool2d(
+            input=ince3b, pool_size=3, pool_type='max', pool_stride=2)
+
+        ince4a = self.inception(pool3, 480, 192, 96, 208, 16, 48, 64, "ince4a")
+        ince4b = self.inception(ince4a, 512, 160, 112, 224, 24, 64, 64,
+                                "ince4b")
+        ince4c = self.inception(ince4b, 512, 128, 128, 256, 24, 64, 64,
+                                "ince4c")
+        ince4d = self.inception(ince4c, 512, 112, 144, 288, 32, 64, 64,
+                                "ince4d")
+        ince4e = self.inception(ince4d, 528, 256, 160, 320, 32, 128, 128,
+                                "ince4e")
+        pool4 = fluid.layers.pool2d(
+            input=ince4e, pool_size=3, pool_type='max', pool_stride=2)
+
+        ince5a = self.inception(pool4, 832, 256, 160, 320, 32, 128, 128,
+                                "ince5a")
+        ince5b = self.inception(ince5a, 832, 384, 192, 384, 48, 128, 128,
+                                "ince5b")
+        pool5 = fluid.layers.pool2d(
+            input=ince5b, pool_size=7, pool_type='avg', pool_stride=7)
+        dropout = fluid.layers.dropout(x=pool5, dropout_prob=0.4)
+        out = fluid.layers.fc(input=dropout,
+                              size=class_dim,
+                              act='softmax',
+                              param_attr=self.xavier(1024, 1, "out"),
+                              name="out",
+                              bias_attr=ParamAttr(name="out_offset"))
+
+        pool_o1 = fluid.layers.pool2d(
+            input=ince4a, pool_size=5, pool_type='avg', pool_stride=3)
+        conv_o1 = self.conv_layer(
+            input=pool_o1,
+            num_filters=128,
+            filter_size=1,
+            stride=1,
+            act=None,
+            name="conv_o1")
+        fc_o1 = fluid.layers.fc(input=conv_o1,
+                                size=1024,
+                                act='relu',
+                                param_attr=self.xavier(2048, 1, "fc_o1"),
+                                name="fc_o1",
+                                bias_attr=ParamAttr(name="fc_o1_offset"))
+        dropout_o1 = fluid.layers.dropout(x=fc_o1, dropout_prob=0.7)
+        out1 = fluid.layers.fc(input=dropout_o1,
+                               size=class_dim,
+                               act='softmax',
+                               param_attr=self.xavier(1024, 1, "out1"),
+                               name="out1",
+                               bias_attr=ParamAttr(name="out1_offset"))
+
+        pool_o2 = fluid.layers.pool2d(
+            input=ince4d, pool_size=5, pool_type='avg', pool_stride=3)
+        conv_o2 = self.conv_layer(
+            input=pool_o2,
+            num_filters=128,
+            filter_size=1,
+            stride=1,
+            act=None,
+            name="conv_o2")
+        fc_o2 = fluid.layers.fc(input=conv_o2,
+                                size=1024,
+                                act='relu',
+                                param_attr=self.xavier(2048, 1, "fc_o2"),
+                                name="fc_o2",
+                                bias_attr=ParamAttr(name="fc_o2_offset"))
+        dropout_o2 = fluid.layers.dropout(x=fc_o2, dropout_prob=0.7)
+        out2 = fluid.layers.fc(input=dropout_o2,
+                               size=class_dim,
+                               act='softmax',
+                               param_attr=self.xavier(1024, 1, "out2"),
+                               name="out2",
+                               bias_attr=ParamAttr(name="out2_offset"))
+
+        # last fc layer is "out"
+        return out, out1, out2

fluid/PaddleCV/image_classification/models_name/inception_v4.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle
+import paddle.fluid as fluid
+import math
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = ['InceptionV4']
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 60, 90],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class InceptionV4():
+    def __init__(self):
+        self.params = train_parameters
+
+    def net(self, input, class_dim=1000):
+        x = self.inception_stem(input)
+
+        for i in range(4):
+            x = self.inceptionA(x, name=str(i + 1))
+        x = self.reductionA(x)
+
+        for i in range(7):
+            x = self.inceptionB(x, name=str(i + 1))
+        x = self.reductionB(x)
+
+        for i in range(3):
+            x = self.inceptionC(x, name=str(i + 1))
+
+        pool = fluid.layers.pool2d(
+            input=x, pool_size=8, pool_type='avg', global_pooling=True)
+
+        drop = fluid.layers.dropout(x=pool, dropout_prob=0.2)
+
+        stdv = 1.0 / math.sqrt(drop.shape[1] * 1.0)
+        out = fluid.layers.fc(
+            input=drop,
+            size=class_dim,
+            act='softmax',
+            param_attr=ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name="final_fc_weights"),
+            bias_attr=ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name="final_fc_offset"))
+        return out
+
+    def conv_bn_layer(self,
+                      data,
+                      num_filters,
+                      filter_size,
+                      stride=1,
+                      padding=0,
+                      groups=1,
+                      act='relu',
+                      name=None):
+        conv = fluid.layers.conv2d(
+            input=data,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=padding,
+            groups=groups,
+            act=None,
+            param_attr=ParamAttr(name=name + "_weights"),
+            bias_attr=False,
+            name=name)
+        bn_name = name + "_bn"
+        return fluid.layers.batch_norm(
+            input=conv,
+            act=act,
+            name=bn_name,
+            param_attr=ParamAttr(name=bn_name + "_scale"),
+            bias_attr=ParamAttr(name=bn_name + "_offset"),
+            moving_mean_name=bn_name + '_mean',
+            moving_variance_name=bn_name + '_variance')
+
+    def inception_stem(self, data, name=None):
+        conv = self.conv_bn_layer(
+            data, 32, 3, stride=2, act='relu', name="conv1_3x3_s2")
+        conv = self.conv_bn_layer(conv, 32, 3, act='relu', name="conv2_3x3_s1")
+        conv = self.conv_bn_layer(
+            conv, 64, 3, padding=1, act='relu', name="conv3_3x3_s1")
+
+        pool1 = fluid.layers.pool2d(
+            input=conv, pool_size=3, pool_stride=2, pool_type='max')
+        conv2 = self.conv_bn_layer(
+            conv, 96, 3, stride=2, act='relu', name="inception_stem1_3x3_s2")
+        concat = fluid.layers.concat([pool1, conv2], axis=1)
+
+        conv1 = self.conv_bn_layer(
+            concat, 64, 1, act='relu', name="inception_stem2_3x3_reduce")
+        conv1 = self.conv_bn_layer(
+            conv1, 96, 3, act='relu', name="inception_stem2_3x3")
+
+        conv2 = self.conv_bn_layer(
+            concat, 64, 1, act='relu', name="inception_stem2_1x7_reduce")
+        conv2 = self.conv_bn_layer(
+            conv2,
+            64, (7, 1),
+            padding=(3, 0),
+            act='relu',
+            name="inception_stem2_1x7")
+        conv2 = self.conv_bn_layer(
+            conv2,
+            64, (1, 7),
+            padding=(0, 3),
+            act='relu',
+            name="inception_stem2_7x1")
+        conv2 = self.conv_bn_layer(
+            conv2, 96, 3, act='relu', name="inception_stem2_3x3_2")
+
+        concat = fluid.layers.concat([conv1, conv2], axis=1)
+
+        conv1 = self.conv_bn_layer(
+            concat, 192, 3, stride=2, act='relu', name="inception_stem3_3x3_s2")
+        pool1 = fluid.layers.pool2d(
+            input=concat, pool_size=3, pool_stride=2, pool_type='max')
+
+        concat = fluid.layers.concat([conv1, pool1], axis=1)
+
+        return concat
+
+    def inceptionA(self, data, name=None):
+        pool1 = fluid.layers.pool2d(
+            input=data, pool_size=3, pool_padding=1, pool_type='avg')
+        conv1 = self.conv_bn_layer(
+            pool1, 96, 1, act='relu', name="inception_a" + name + "_1x1")
+
+        conv2 = self.conv_bn_layer(
+            data, 96, 1, act='relu', name="inception_a" + name + "_1x1_2")
+
+        conv3 = self.conv_bn_layer(
+            data, 64, 1, act='relu', name="inception_a" + name + "_3x3_reduce")
+        conv3 = self.conv_bn_layer(
+            conv3,
+            96,
+            3,
+            padding=1,
+            act='relu',
+            name="inception_a" + name + "_3x3")
+
+        conv4 = self.conv_bn_layer(
+            data,
+            64,
+            1,
+            act='relu',
+            name="inception_a" + name + "_3x3_2_reduce")
+        conv4 = self.conv_bn_layer(
+            conv4,
+            96,
+            3,
+            padding=1,
+            act='relu',
+            name="inception_a" + name + "_3x3_2")
+        conv4 = self.conv_bn_layer(
+            conv4,
+            96,
+            3,
+            padding=1,
+            act='relu',
+            name="inception_a" + name + "_3x3_3")
+
+        concat = fluid.layers.concat([conv1, conv2, conv3, conv4], axis=1)
+
+        return concat
+
+    def reductionA(self, data, name=None):
+        pool1 = fluid.layers.pool2d(
+            input=data, pool_size=3, pool_stride=2, pool_type='max')
+
+        conv2 = self.conv_bn_layer(
+            data, 384, 3, stride=2, act='relu', name="reduction_a_3x3")
+
+        conv3 = self.conv_bn_layer(
+            data, 192, 1, act='relu', name="reduction_a_3x3_2_reduce")
+        conv3 = self.conv_bn_layer(
+            conv3, 224, 3, padding=1, act='relu', name="reduction_a_3x3_2")
+        conv3 = self.conv_bn_layer(
+            conv3, 256, 3, stride=2, act='relu', name="reduction_a_3x3_3")
+
+        concat = fluid.layers.concat([pool1, conv2, conv3], axis=1)
+
+        return concat
+
+    def inceptionB(self, data, name=None):
+        pool1 = fluid.layers.pool2d(
+            input=data, pool_size=3, pool_padding=1, pool_type='avg')
+        conv1 = self.conv_bn_layer(
+            pool1, 128, 1, act='relu', name="inception_b" + name + "_1x1")
+
+        conv2 = self.conv_bn_layer(
+            data, 384, 1, act='relu', name="inception_b" + name + "_1x1_2")
+
+        conv3 = self.conv_bn_layer(
+            data, 192, 1, act='relu', name="inception_b" + name + "_1x7_reduce")
+        conv3 = self.conv_bn_layer(
+            conv3,
+            224, (1, 7),
+            padding=(0, 3),
+            act='relu',
+            name="inception_b" + name + "_1x7")
+        conv3 = self.conv_bn_layer(
+            conv3,
+            256, (7, 1),
+            padding=(3, 0),
+            act='relu',
+            name="inception_b" + name + "_7x1")
+
+        conv4 = self.conv_bn_layer(
+            data,
+            192,
+            1,
+            act='relu',
+            name="inception_b" + name + "_7x1_2_reduce")
+        conv4 = self.conv_bn_layer(
+            conv4,
+            192, (1, 7),
+            padding=(0, 3),
+            act='relu',
+            name="inception_b" + name + "_1x7_2")
+        conv4 = self.conv_bn_layer(
+            conv4,
+            224, (7, 1),
+            padding=(3, 0),
+            act='relu',
+            name="inception_b" + name + "_7x1_2")
+        conv4 = self.conv_bn_layer(
+            conv4,
+            224, (1, 7),
+            padding=(0, 3),
+            act='relu',
+            name="inception_b" + name + "_1x7_3")
+        conv4 = self.conv_bn_layer(
+            conv4,
+            256, (7, 1),
+            padding=(3, 0),
+            act='relu',
+            name="inception_b" + name + "_7x1_3")
+
+        concat = fluid.layers.concat([conv1, conv2, conv3, conv4], axis=1)
+
+        return concat
+
+    def reductionB(self, data, name=None):
+        pool1 = fluid.layers.pool2d(
+            input=data, pool_size=3, pool_stride=2, pool_type='max')
+
+        conv2 = self.conv_bn_layer(
+            data, 192, 1, act='relu', name="reduction_b_3x3_reduce")
+        conv2 = self.conv_bn_layer(
+            conv2, 192, 3, stride=2, act='relu', name="reduction_b_3x3")
+
+        conv3 = self.conv_bn_layer(
+            data, 256, 1, act='relu', name="reduction_b_1x7_reduce")
+        conv3 = self.conv_bn_layer(
+            conv3,
+            256, (1, 7),
+            padding=(0, 3),
+            act='relu',
+            name="reduction_b_1x7")
+        conv3 = self.conv_bn_layer(
+            conv3,
+            320, (7, 1),
+            padding=(3, 0),
+            act='relu',
+            name="reduction_b_7x1")
+        conv3 = self.conv_bn_layer(
+            conv3, 320, 3, stride=2, act='relu', name="reduction_b_3x3_2")
+
+        concat = fluid.layers.concat([pool1, conv2, conv3], axis=1)
+
+        return concat
+
+    def inceptionC(self, data, name=None):
+        pool1 = fluid.layers.pool2d(
+            input=data, pool_size=3, pool_padding=1, pool_type='avg')
+        conv1 = self.conv_bn_layer(
+            pool1, 256, 1, act='relu', name="inception_c" + name + "_1x1")
+
+        conv2 = self.conv_bn_layer(
+            data, 256, 1, act='relu', name="inception_c" + name + "_1x1_2")
+
+        conv3 = self.conv_bn_layer(
+            data, 384, 1, act='relu', name="inception_c" + name + "_1x1_3")
+        conv3_1 = self.conv_bn_layer(
+            conv3,
+            256, (1, 3),
+            padding=(0, 1),
+            act='relu',
+            name="inception_c" + name + "_1x3")
+        conv3_2 = self.conv_bn_layer(
+            conv3,
+            256, (3, 1),
+            padding=(1, 0),
+            act='relu',
+            name="inception_c" + name + "_3x1")
+
+        conv4 = self.conv_bn_layer(
+            data, 384, 1, act='relu', name="inception_c" + name + "_1x1_4")
+        conv4 = self.conv_bn_layer(
+            conv4,
+            448, (1, 3),
+            padding=(0, 1),
+            act='relu',
+            name="inception_c" + name + "_1x3_2")
+        conv4 = self.conv_bn_layer(
+            conv4,
+            512, (3, 1),
+            padding=(1, 0),
+            act='relu',
+            name="inception_c" + name + "_3x1_2")
+        conv4_1 = self.conv_bn_layer(
+            conv4,
+            256, (1, 3),
+            padding=(0, 1),
+            act='relu',
+            name="inception_c" + name + "_1x3_3")
+        conv4_2 = self.conv_bn_layer(
+            conv4,
+            256, (3, 1),
+            padding=(1, 0),
+            act='relu',
+            name="inception_c" + name + "_3x1_3")
+
+        concat = fluid.layers.concat(
+            [conv1, conv2, conv3_1, conv3_2, conv4_1, conv4_2], axis=1)
+
+        return concat

fluid/PaddleCV/image_classification/models_name/mobilenet.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle.fluid as fluid
+from paddle.fluid.initializer import MSRA
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = ['MobileNet']
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 60, 90],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class MobileNet():
+    def __init__(self):
+        self.params = train_parameters
+
+    def net(self, input, class_dim=1000, scale=1.0):
+        # conv1: 112x112
+        input = self.conv_bn_layer(
+            input,
+            filter_size=3,
+            channels=3,
+            num_filters=int(32 * scale),
+            stride=2,
+            padding=1,
+            name="conv1")
+
+        # 56x56
+        input = self.depthwise_separable(
+            input,
+            num_filters1=32,
+            num_filters2=64,
+            num_groups=32,
+            stride=1,
+            scale=scale,
+            name="conv2_1")
+
+        input = self.depthwise_separable(
+            input,
+            num_filters1=64,
+            num_filters2=128,
+            num_groups=64,
+            stride=2,
+            scale=scale,
+            name="conv2_2")
+
+        # 28x28
+        input = self.depthwise_separable(
+            input,
+            num_filters1=128,
+            num_filters2=128,
+            num_groups=128,
+            stride=1,
+            scale=scale,
+            name="conv3_1")
+
+        input = self.depthwise_separable(
+            input,
+            num_filters1=128,
+            num_filters2=256,
+            num_groups=128,
+            stride=2,
+            scale=scale,
+            name="conv3_2")
+
+        # 14x14
+        input = self.depthwise_separable(
+            input,
+            num_filters1=256,
+            num_filters2=256,
+            num_groups=256,
+            stride=1,
+            scale=scale,
+            name="conv4_1")
+
+        input = self.depthwise_separable(
+            input,
+            num_filters1=256,
+            num_filters2=512,
+            num_groups=256,
+            stride=2,
+            scale=scale,
+            name="conv4_2")
+
+        # 14x14
+        for i in range(5):
+            input = self.depthwise_separable(
+                input,
+                num_filters1=512,
+                num_filters2=512,
+                num_groups=512,
+                stride=1,
+                scale=scale,
+                name="conv5" + "_" + str(i + 1))
+        # 7x7
+        input = self.depthwise_separable(
+            input,
+            num_filters1=512,
+            num_filters2=1024,
+            num_groups=512,
+            stride=2,
+            scale=scale,
+            name="conv5_6")
+
+        input = self.depthwise_separable(
+            input,
+            num_filters1=1024,
+            num_filters2=1024,
+            num_groups=1024,
+            stride=1,
+            scale=scale,
+            name="conv6")
+
+        input = fluid.layers.pool2d(
+            input=input,
+            pool_size=0,
+            pool_stride=1,
+            pool_type='avg',
+            global_pooling=True)
+
+        output = fluid.layers.fc(input=input,
+                                 size=class_dim,
+                                 act='softmax',
+                                 param_attr=ParamAttr(
+                                     initializer=MSRA(), name="fc7_weights"),
+                                 bias_attr=ParamAttr(name="fc7_offset"))
+        return output
+
+    def conv_bn_layer(self,
+                      input,
+                      filter_size,
+                      num_filters,
+                      stride,
+                      padding,
+                      channels=None,
+                      num_groups=1,
+                      act='relu',
+                      use_cudnn=True,
+                      name=None):
+        conv = fluid.layers.conv2d(
+            input=input,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=padding,
+            groups=num_groups,
+            act=None,
+            use_cudnn=use_cudnn,
+            param_attr=ParamAttr(
+                initializer=MSRA(), name=name + "_weights"),
+            bias_attr=False)
+        bn_name = name + "_bn"
+        return fluid.layers.batch_norm(
+            input=conv,
+            act=act,
+            param_attr=ParamAttr(name=bn_name + "_scale"),
+            bias_attr=ParamAttr(name=bn_name + "_offset"),
+            moving_mean_name=bn_name + '_mean',
+            moving_variance_name=bn_name + '_variance')
+
+    def depthwise_separable(self,
+                            input,
+                            num_filters1,
+                            num_filters2,
+                            num_groups,
+                            stride,
+                            scale,
+                            name=None):
+        depthwise_conv = self.conv_bn_layer(
+            input=input,
+            filter_size=3,
+            num_filters=int(num_filters1 * scale),
+            stride=stride,
+            padding=1,
+            num_groups=int(num_groups * scale),
+            use_cudnn=False,
+            name=name + "_dw")
+
+        pointwise_conv = self.conv_bn_layer(
+            input=depthwise_conv,
+            filter_size=1,
+            num_filters=int(num_filters2 * scale),
+            stride=1,
+            padding=0,
+            name=name + "_sep")
+        return pointwise_conv

fluid/PaddleCV/image_classification/models_name/mobilenet_v2.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle.fluid as fluid
+from paddle.fluid.initializer import MSRA
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = ['MobileNetV2']
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 60, 90],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class MobileNetV2():
+    def __init__(self):
+        self.params = train_parameters
+
+    def net(self, input, class_dim=1000, scale=1.0):
+
+        bottleneck_params_list = [
+            (1, 16, 1, 1),
+            (6, 24, 2, 2),
+            (6, 32, 3, 2),
+            (6, 64, 4, 2),
+            (6, 96, 3, 1),
+            (6, 160, 3, 2),
+            (6, 320, 1, 1),
+        ]
+
+        #conv1 
+        input = self.conv_bn_layer(
+            input,
+            num_filters=int(32 * scale),
+            filter_size=3,
+            stride=2,
+            padding=1,
+            if_act=True,
+            name='conv1_1')
+
+        # bottleneck sequences
+        i = 1
+        in_c = int(32 * scale)
+        for layer_setting in bottleneck_params_list:
+            t, c, n, s = layer_setting
+            i += 1
+            input = self.invresi_blocks(
+                input=input,
+                in_c=in_c,
+                t=t,
+                c=int(c * scale),
+                n=n,
+                s=s,
+                name='conv' + str(i))
+            in_c = int(c * scale)
+        #last_conv
+        input = self.conv_bn_layer(
+            input=input,
+            num_filters=int(1280 * scale) if scale > 1.0 else 1280,
+            filter_size=1,
+            stride=1,
+            padding=0,
+            if_act=True,
+            name='conv9')
+
+        input = fluid.layers.pool2d(
+            input=input,
+            pool_size=7,
+            pool_stride=1,
+            pool_type='avg',
+            global_pooling=True)
+
+        output = fluid.layers.fc(input=input,
+                                 size=class_dim,
+                                 act='softmax',
+                                 param_attr=ParamAttr(name='fc10_weights'),
+                                 bias_attr=ParamAttr(name='fc10_offset'))
+        return output
+
+    def conv_bn_layer(self,
+                      input,
+                      filter_size,
+                      num_filters,
+                      stride,
+                      padding,
+                      channels=None,
+                      num_groups=1,
+                      if_act=True,
+                      name=None,
+                      use_cudnn=True):
+        conv = fluid.layers.conv2d(
+            input=input,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=padding,
+            groups=num_groups,
+            act=None,
+            use_cudnn=use_cudnn,
+            param_attr=ParamAttr(name=name + '_weights'),
+            bias_attr=False)
+        bn_name = name + '_bn'
+        bn = fluid.layers.batch_norm(
+            input=conv,
+            param_attr=ParamAttr(name=bn_name + "_scale"),
+            bias_attr=ParamAttr(name=bn_name + "_offset"),
+            moving_mean_name=bn_name + '_mean',
+            moving_variance_name=bn_name + '_variance')
+        if if_act:
+            return fluid.layers.relu6(bn)
+        else:
+            return bn
+
+    def shortcut(self, input, data_residual):
+        return fluid.layers.elementwise_add(input, data_residual)
+
+    def inverted_residual_unit(self,
+                               input,
+                               num_in_filter,
+                               num_filters,
+                               ifshortcut,
+                               stride,
+                               filter_size,
+                               padding,
+                               expansion_factor,
+                               name=None):
+        num_expfilter = int(round(num_in_filter * expansion_factor))
+
+        channel_expand = self.conv_bn_layer(
+            input=input,
+            num_filters=num_expfilter,
+            filter_size=1,
+            stride=1,
+            padding=0,
+            num_groups=1,
+            if_act=True,
+            name=name + '_expand')
+
+        bottleneck_conv = self.conv_bn_layer(
+            input=channel_expand,
+            num_filters=num_expfilter,
+            filter_size=filter_size,
+            stride=stride,
+            padding=padding,
+            num_groups=num_expfilter,
+            if_act=True,
+            name=name + '_dwise',
+            use_cudnn=False)
+
+        linear_out = self.conv_bn_layer(
+            input=bottleneck_conv,
+            num_filters=num_filters,
+            filter_size=1,
+            stride=1,
+            padding=0,
+            num_groups=1,
+            if_act=False,
+            name=name + '_linear')
+        if ifshortcut:
+            out = self.shortcut(input=input, data_residual=linear_out)
+            return out
+        else:
+            return linear_out
+
+    def invresi_blocks(self, input, in_c, t, c, n, s, name=None):
+        first_block = self.inverted_residual_unit(
+            input=input,
+            num_in_filter=in_c,
+            num_filters=c,
+            ifshortcut=False,
+            stride=s,
+            filter_size=3,
+            padding=1,
+            expansion_factor=t,
+            name=name + '_1')
+
+        last_residual_block = first_block
+        last_c = c
+
+        for i in range(1, n):
+            last_residual_block = self.inverted_residual_unit(
+                input=last_residual_block,
+                num_in_filter=last_c,
+                num_filters=c,
+                ifshortcut=True,
+                stride=1,
+                filter_size=3,
+                padding=1,
+                expansion_factor=t,
+                name=name + '_' + str(i + 1))
+        return last_residual_block

fluid/PaddleCV/image_classification/models_name/resnet.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle
+import paddle.fluid as fluid
+import math
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = ["ResNet", "ResNet50", "ResNet101", "ResNet152"]
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 60, 90],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class ResNet():
+    def __init__(self, layers=50):
+        self.params = train_parameters
+        self.layers = layers
+
+    def net(self, input, class_dim=1000):
+        layers = self.layers
+        supported_layers = [50, 101, 152]
+        assert layers in supported_layers, \
+            "supported layers are {} but input layer is {}".format(supported_layers, layers)
+
+        if layers == 50:
+            depth = [3, 4, 6, 3]
+        elif layers == 101:
+            depth = [3, 4, 23, 3]
+        elif layers == 152:
+            depth = [3, 8, 36, 3]
+        num_filters = [64, 128, 256, 512]
+
+        conv = self.conv_bn_layer(
+            input=input,
+            num_filters=64,
+            filter_size=7,
+            stride=2,
+            act='relu',
+            name="conv1")
+        conv = fluid.layers.pool2d(
+            input=conv,
+            pool_size=3,
+            pool_stride=2,
+            pool_padding=1,
+            pool_type='max')
+
+        for block in range(len(depth)):
+            for i in range(depth[block]):
+                if layers in [101, 152] and block == 2:
+                    if i == 0:
+                        conv_name = "res" + str(block + 2) + "a"
+                    else:
+                        conv_name = "res" + str(block + 2) + "b" + str(i)
+                else:
+                    conv_name = "res" + str(block + 2) + chr(97 + i)
+                conv = self.bottleneck_block(
+                    input=conv,
+                    num_filters=num_filters[block],
+                    stride=2 if i == 0 and block != 0 else 1,
+                    name=conv_name)
+
+        pool = fluid.layers.pool2d(
+            input=conv, pool_size=7, pool_type='avg', global_pooling=True)
+        stdv = 1.0 / math.sqrt(pool.shape[1] * 1.0)
+        out = fluid.layers.fc(input=pool,
+                              size=class_dim,
+                              act='softmax',
+                              param_attr=fluid.param_attr.ParamAttr(
+                                  initializer=fluid.initializer.Uniform(-stdv,
+                                                                        stdv)))
+        return out
+
+    def conv_bn_layer(self,
+                      input,
+                      num_filters,
+                      filter_size,
+                      stride=1,
+                      groups=1,
+                      act=None,
+                      name=None):
+        conv = fluid.layers.conv2d(
+            input=input,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=(filter_size - 1) // 2,
+            groups=groups,
+            act=None,
+            param_attr=ParamAttr(name=name + "_weights"),
+            bias_attr=False,
+            name=name + '.conv2d.output.1')
+        if name == "conv1":
+            bn_name = "bn_" + name
+        else:
+            bn_name = "bn" + name[3:]
+        return fluid.layers.batch_norm(
+            input=conv,
+            act=act,
+            name=bn_name + '.output.1',
+            param_attr=ParamAttr(name=bn_name + '_scale'),
+            bias_attr=ParamAttr(bn_name + '_offset'),
+            moving_mean_name=bn_name + '_mean',
+            moving_variance_name=bn_name + '_variance', )
+
+    def shortcut(self, input, ch_out, stride, name):
+        ch_in = input.shape[1]
+        if ch_in != ch_out or stride != 1:
+            return self.conv_bn_layer(input, ch_out, 1, stride, name=name)
+        else:
+            return input
+
+    def bottleneck_block(self, input, num_filters, stride, name):
+        conv0 = self.conv_bn_layer(
+            input=input,
+            num_filters=num_filters,
+            filter_size=1,
+            act='relu',
+            name=name + "_branch2a")
+        conv1 = self.conv_bn_layer(
+            input=conv0,
+            num_filters=num_filters,
+            filter_size=3,
+            stride=stride,
+            act='relu',
+            name=name + "_branch2b")
+        conv2 = self.conv_bn_layer(
+            input=conv1,
+            num_filters=num_filters * 4,
+            filter_size=1,
+            act=None,
+            name=name + "_branch2c")
+
+        short = self.shortcut(
+            input, num_filters * 4, stride, name=name + "_branch1")
+
+        return fluid.layers.elementwise_add(
+            x=short, y=conv2, act='relu', name=name + ".add.output.5")
+
+
+def ResNet50():
+    model = ResNet(layers=50)
+    return model
+
+
+def ResNet101():
+    model = ResNet(layers=101)
+    return model
+
+
+def ResNet152():
+    model = ResNet(layers=152)
+    return model

fluid/PaddleCV/image_classification/models_name/se_resnext.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle
+import paddle.fluid as fluid
+import math
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = [
+    "SE_ResNeXt", "SE_ResNeXt50_32x4d", "SE_ResNeXt101_32x4d",
+    "SE_ResNeXt152_32x4d"
+]
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "dropout_seed": None,
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [40, 80, 100],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class SE_ResNeXt():
+    def __init__(self, layers=50):
+        self.params = train_parameters
+        self.layers = layers
+
+    def net(self, input, class_dim=1000):
+        layers = self.layers
+        supported_layers = [50, 101, 152]
+        assert layers in supported_layers, \
+            "supported layers are {} but input layer is {}".format(supported_layers, layers)
+        if layers == 50:
+            cardinality = 32
+            reduction_ratio = 16
+            depth = [3, 4, 6, 3]
+            num_filters = [128, 256, 512, 1024]
+
+            conv = self.conv_bn_layer(
+                input=input,
+                num_filters=64,
+                filter_size=7,
+                stride=2,
+                act='relu',
+                name='conv1', )
+            conv = fluid.layers.pool2d(
+                input=conv,
+                pool_size=3,
+                pool_stride=2,
+                pool_padding=1,
+                pool_type='max')
+        elif layers == 101:
+            cardinality = 32
+            reduction_ratio = 16
+            depth = [3, 4, 23, 3]
+            num_filters = [128, 256, 512, 1024]
+
+            conv = self.conv_bn_layer(
+                input=input,
+                num_filters=64,
+                filter_size=7,
+                stride=2,
+                act='relu',
+                name="conv1", )
+            conv = fluid.layers.pool2d(
+                input=conv,
+                pool_size=3,
+                pool_stride=2,
+                pool_padding=1,
+                pool_type='max')
+        elif layers == 152:
+            cardinality = 64
+            reduction_ratio = 16
+            depth = [3, 8, 36, 3]
+            num_filters = [128, 256, 512, 1024]
+
+            conv = self.conv_bn_layer(
+                input=input,
+                num_filters=64,
+                filter_size=3,
+                stride=2,
+                act='relu',
+                name='conv1')
+            conv = self.conv_bn_layer(
+                input=conv,
+                num_filters=64,
+                filter_size=3,
+                stride=1,
+                act='relu',
+                name='conv2')
+            conv = self.conv_bn_layer(
+                input=conv,
+                num_filters=128,
+                filter_size=3,
+                stride=1,
+                act='relu',
+                name='conv3')
+            conv = fluid.layers.pool2d(
+                input=conv, pool_size=3, pool_stride=2, pool_padding=1, \
+                pool_type='max')
+        n = 1 if layers == 50 or layers == 101 else 3
+        for block in range(len(depth)):
+            n += 1
+            for i in range(depth[block]):
+                conv = self.bottleneck_block(
+                    input=conv,
+                    num_filters=num_filters[block],
+                    stride=2 if i == 0 and block != 0 else 1,
+                    cardinality=cardinality,
+                    reduction_ratio=reduction_ratio,
+                    name=str(n) + '_' + str(i + 1))
+
+        pool = fluid.layers.pool2d(
+            input=conv, pool_size=7, pool_type='avg', global_pooling=True)
+        drop = fluid.layers.dropout(
+            x=pool, dropout_prob=0.5, seed=self.params['dropout_seed'])
+        stdv = 1.0 / math.sqrt(drop.shape[1] * 1.0)
+        out = fluid.layers.fc(
+            input=drop,
+            size=class_dim,
+            act='softmax',
+            param_attr=ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name='fc6_weights'),
+            bias_attr=ParamAttr(name='fc6_offset'))
+        return out
+
+    def shortcut(self, input, ch_out, stride, name):
+        ch_in = input.shape[1]
+        if ch_in != ch_out or stride != 1:
+            filter_size = 1
+            return self.conv_bn_layer(
+                input, ch_out, filter_size, stride, name='conv' + name + '_prj')
+        else:
+            return input
+
+    def bottleneck_block(self,
+                         input,
+                         num_filters,
+                         stride,
+                         cardinality,
+                         reduction_ratio,
+                         name=None):
+        conv0 = self.conv_bn_layer(
+            input=input,
+            num_filters=num_filters,
+            filter_size=1,
+            act='relu',
+            name='conv' + name + '_x1')
+        conv1 = self.conv_bn_layer(
+            input=conv0,
+            num_filters=num_filters,
+            filter_size=3,
+            stride=stride,
+            groups=cardinality,
+            act='relu',
+            name='conv' + name + '_x2')
+        conv2 = self.conv_bn_layer(
+            input=conv1,
+            num_filters=num_filters * 2,
+            filter_size=1,
+            act=None,
+            name='conv' + name + '_x3')
+        scale = self.squeeze_excitation(
+            input=conv2,
+            num_channels=num_filters * 2,
+            reduction_ratio=reduction_ratio,
+            name='fc' + name)
+
+        short = self.shortcut(input, num_filters * 2, stride, name=name)
+
+        return fluid.layers.elementwise_add(x=short, y=scale, act='relu')
+
+    def conv_bn_layer(self,
+                      input,
+                      num_filters,
+                      filter_size,
+                      stride=1,
+                      groups=1,
+                      act=None,
+                      name=None):
+        conv = fluid.layers.conv2d(
+            input=input,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=(filter_size - 1) // 2,
+            groups=groups,
+            act=None,
+            bias_attr=False,
+            param_attr=ParamAttr(name=name + '_weights'), )
+        bn_name = name + "_bn"
+        return fluid.layers.batch_norm(
+            input=conv,
+            act=act,
+            param_attr=ParamAttr(name=bn_name + '_scale'),
+            bias_attr=ParamAttr(bn_name + '_offset'),
+            moving_mean_name=bn_name + '_mean',
+            moving_variance_name=bn_name + '_variance')
+
+    def squeeze_excitation(self,
+                           input,
+                           num_channels,
+                           reduction_ratio,
+                           name=None):
+        pool = fluid.layers.pool2d(
+            input=input, pool_size=0, pool_type='avg', global_pooling=True)
+        stdv = 1.0 / math.sqrt(pool.shape[1] * 1.0)
+        squeeze = fluid.layers.fc(
+            input=pool,
+            size=num_channels // reduction_ratio,
+            act='relu',
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=name + '_sqz_weights'),
+            bias_attr=ParamAttr(name=name + '_sqz_offset'))
+        stdv = 1.0 / math.sqrt(squeeze.shape[1] * 1.0)
+        excitation = fluid.layers.fc(
+            input=squeeze,
+            size=num_channels,
+            act='sigmoid',
+            param_attr=fluid.param_attr.ParamAttr(
+                initializer=fluid.initializer.Uniform(-stdv, stdv),
+                name=name + '_exc_weights'),
+            bias_attr=ParamAttr(name=name + '_exc_offset'))
+        scale = fluid.layers.elementwise_mul(x=input, y=excitation, axis=0)
+        return scale
+
+
+def SE_ResNeXt50_32x4d():
+    model = SE_ResNeXt(layers=50)
+    return model
+
+
+def SE_ResNeXt101_32x4d():
+    model = SE_ResNeXt(layers=101)
+    return model
+
+
+def SE_ResNeXt152_32x4d():
+    model = SE_ResNeXt(layers=152)
+    return model

fluid/PaddleCV/image_classification/models_name/shufflenet_v2.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle.fluid as fluid
+from paddle.fluid.initializer import MSRA
+from paddle.fluid.param_attr import ParamAttr
+
+__all__ = [
+    'ShuffleNetV2', 'ShuffleNetV2_x0_5', 'ShuffleNetV2_x1_0',
+    'ShuffleNetV2_x1_5', 'ShuffleNetV2_x2_0'
+]
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 60, 90],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class ShuffleNetV2():
+    def __init__(self, scale=1.0):
+        self.params = train_parameters
+        self.scale = scale
+
+    def net(self, input, class_dim=1000):
+        scale = self.scale
+        stage_repeats = [4, 8, 4]
+
+        if scale == 0.5:
+            stage_out_channels = [-1, 24, 48, 96, 192, 1024]
+        elif scale == 1.0:
+            stage_out_channels = [-1, 24, 116, 232, 464, 1024]
+        elif scale == 1.5:
+            stage_out_channels = [-1, 24, 176, 352, 704, 1024]
+        elif scale == 2.0:
+            stage_out_channels = [-1, 24, 224, 488, 976, 2048]
+        else:
+            raise ValueError("""{} groups is not supported for
+                       1x1 Grouped Convolutions""".format(num_groups))
+
+        #conv1
+
+        input_channel = stage_out_channels[1]
+        conv1 = self.conv_bn_layer(
+            input=input,
+            filter_size=3,
+            num_filters=input_channel,
+            padding=1,
+            stride=2,
+            name='stage1_conv')
+        pool1 = fluid.layers.pool2d(
+            input=conv1,
+            pool_size=3,
+            pool_stride=2,
+            pool_padding=1,
+            pool_type='max')
+        conv = pool1
+        # bottleneck sequences
+        for idxstage in range(len(stage_repeats)):
+            numrepeat = stage_repeats[idxstage]
+            output_channel = stage_out_channels[idxstage + 2]
+            for i in range(numrepeat):
+                if i == 0:
+                    conv = self.inverted_residual_unit(
+                        input=conv,
+                        num_filters=output_channel,
+                        stride=2,
+                        benchmodel=2,
+                        name=str(idxstage + 2) + '_' + str(i + 1))
+                else:
+                    conv = self.inverted_residual_unit(
+                        input=conv,
+                        num_filters=output_channel,
+                        stride=1,
+                        benchmodel=1,
+                        name=str(idxstage + 2) + '_' + str(i + 1))
+
+        conv_last = self.conv_bn_layer(
+            input=conv,
+            filter_size=1,
+            num_filters=stage_out_channels[-1],
+            padding=0,
+            stride=1,
+            name='conv5')
+        pool_last = fluid.layers.pool2d(
+            input=conv_last,
+            pool_size=7,
+            pool_stride=1,
+            pool_padding=0,
+            pool_type='avg')
+
+        output = fluid.layers.fc(input=pool_last,
+                                 size=class_dim,
+                                 act='softmax',
+                                 param_attr=ParamAttr(
+                                     initializer=MSRA(), name='fc6_weights'),
+                                 bias_attr=ParamAttr(name='fc6_offset'))
+        return output
+
+    def conv_bn_layer(self,
+                      input,
+                      filter_size,
+                      num_filters,
+                      stride,
+                      padding,
+                      num_groups=1,
+                      use_cudnn=True,
+                      if_act=True,
+                      name=None):
+        #         print(num_groups)
+        conv = fluid.layers.conv2d(
+            input=input,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=padding,
+            groups=num_groups,
+            act=None,
+            use_cudnn=use_cudnn,
+            param_attr=ParamAttr(
+                initializer=MSRA(), name=name + '_weights'),
+            bias_attr=False)
+        bn_name = name + '_bn'
+        if if_act:
+            return fluid.layers.batch_norm(
+                input=conv,
+                act='relu',
+                param_attr=ParamAttr(name=bn_name + "_scale"),
+                bias_attr=ParamAttr(name=bn_name + "_offset"),
+                moving_mean_name=bn_name + '_mean',
+                moving_variance_name=bn_name + '_variance')
+        else:
+            return fluid.layers.batch_norm(
+                input=conv,
+                param_attr=ParamAttr(name=bn_name + "_scale"),
+                bias_attr=ParamAttr(name=bn_name + "_offset"),
+                moving_mean_name=bn_name + '_mean',
+                moving_variance_name=bn_name + '_variance')
+
+    def channel_shuffle(self, x, groups):
+        batchsize, num_channels, height, width = x.shape[0], x.shape[
+            1], x.shape[2], x.shape[3]
+        channels_per_group = num_channels // groups
+
+        # reshape
+        x = fluid.layers.reshape(
+            x=x, shape=[batchsize, groups, channels_per_group, height, width])
+
+        x = fluid.layers.transpose(x=x, perm=[0, 2, 1, 3, 4])
+
+        # flatten
+        x = fluid.layers.reshape(
+            x=x, shape=[batchsize, num_channels, height, width])
+
+        return x
+
+    def inverted_residual_unit(self,
+                               input,
+                               num_filters,
+                               stride,
+                               benchmodel,
+                               name=None):
+        assert stride in [1, 2], \
+            "supported stride are {} but your stride is {}".format([1,2], stride)
+
+        oup_inc = num_filters // 2
+        inp = input.shape[1]
+
+        if benchmodel == 1:
+            x1, x2 = fluid.layers.split(
+                input,
+                num_or_sections=[input.shape[1] // 2, input.shape[1] // 2],
+                dim=1)
+            #             x1 = input[:, :(input.shape[1]//2), :, :]
+            #             x2 = input[:, (input.shape[1]//2):, :, :]
+
+            conv_pw = self.conv_bn_layer(
+                input=x2,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True,
+                name='stage_' + name + '_conv1')
+
+            conv_dw = self.conv_bn_layer(
+                input=conv_pw,
+                num_filters=oup_inc,
+                filter_size=3,
+                stride=stride,
+                padding=1,
+                num_groups=oup_inc,
+                if_act=False,
+                name='stage_' + name + '_conv2')
+
+            conv_linear = self.conv_bn_layer(
+                input=conv_dw,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True,
+                name='stage_' + name + '_conv3')
+
+            out = fluid.layers.concat([x1, conv_linear], axis=1)
+
+        else:
+            #branch1
+            conv_dw_1 = self.conv_bn_layer(
+                input=input,
+                num_filters=inp,
+                filter_size=3,
+                stride=stride,
+                padding=1,
+                num_groups=inp,
+                if_act=False,
+                name='stage_' + name + '_conv4')
+
+            conv_linear_1 = self.conv_bn_layer(
+                input=conv_dw_1,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True,
+                name='stage_' + name + '_conv5')
+
+            #branch2
+            conv_pw_2 = self.conv_bn_layer(
+                input=input,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True,
+                name='stage_' + name + '_conv1')
+
+            conv_dw_2 = self.conv_bn_layer(
+                input=conv_pw_2,
+                num_filters=oup_inc,
+                filter_size=3,
+                stride=stride,
+                padding=1,
+                num_groups=oup_inc,
+                if_act=False,
+                name='stage_' + name + '_conv2')
+
+            conv_linear_2 = self.conv_bn_layer(
+                input=conv_dw_2,
+                num_filters=oup_inc,
+                filter_size=1,
+                stride=1,
+                padding=0,
+                num_groups=1,
+                if_act=True,
+                name='stage_' + name + '_conv3')
+            out = fluid.layers.concat([conv_linear_1, conv_linear_2], axis=1)
+
+        return self.channel_shuffle(out, 2)
+
+
+def ShuffleNetV2_x0_5():
+    model = ShuffleNetV2(scale=0.5)
+    return model
+
+
+def ShuffleNetV2_x1_0():
+    model = ShuffleNetV2(scale=1.0)
+    return model
+
+
+def ShuffleNetV2_x1_5():
+    model = ShuffleNetV2(scale=1.5)
+    return model
+
+
+def ShuffleNetV2_x2_0():
+    model = ShuffleNetV2(scale=2.0)
+    return model

fluid/PaddleCV/image_classification/models_name/vgg.py

+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+import paddle
+import paddle.fluid as fluid
+
+__all__ = ["VGGNet", "VGG11", "VGG13", "VGG16", "VGG19"]
+
+train_parameters = {
+    "input_size": [3, 224, 224],
+    "input_mean": [0.485, 0.456, 0.406],
+    "input_std": [0.229, 0.224, 0.225],
+    "learning_strategy": {
+        "name": "piecewise_decay",
+        "batch_size": 256,
+        "epochs": [30, 60, 90],
+        "steps": [0.1, 0.01, 0.001, 0.0001]
+    }
+}
+
+
+class VGGNet():
+    def __init__(self, layers=16):
+        self.params = train_parameters
+        self.layers = layers
+
+    def net(self, input, class_dim=1000):
+        layers = self.layers
+        vgg_spec = {
+            11: ([1, 1, 2, 2, 2]),
+            13: ([2, 2, 2, 2, 2]),
+            16: ([2, 2, 3, 3, 3]),
+            19: ([2, 2, 4, 4, 4])
+        }
+        assert layers in vgg_spec.keys(), \
+            "supported layers are {} but input layer is {}".format(vgg_spec.keys(), layers)
+
+        nums = vgg_spec[layers]
+        conv1 = self.conv_block(input, 64, nums[0], name="conv1_")
+        conv2 = self.conv_block(conv1, 128, nums[1], name="conv2_")
+        conv3 = self.conv_block(conv2, 256, nums[2], name="conv3_")
+        conv4 = self.conv_block(conv3, 512, nums[3], name="conv4_")
+        conv5 = self.conv_block(conv4, 512, nums[4], name="conv5_")
+
+        fc_dim = 4096
+        fc_name = ["fc6", "fc7", "fc8"]
+        fc1 = fluid.layers.fc(
+            input=conv5,
+            size=fc_dim,
+            act='relu',
+            param_attr=fluid.param_attr.ParamAttr(name=fc_name[0] + "_weights"),
+            bias_attr=fluid.param_attr.ParamAttr(name=fc_name[0] + "_offset"))
+        fc1 = fluid.layers.dropout(x=fc1, dropout_prob=0.5)
+        fc2 = fluid.layers.fc(
+            input=fc1,
+            size=fc_dim,
+            act='relu',
+            param_attr=fluid.param_attr.ParamAttr(name=fc_name[1] + "_weights"),
+            bias_attr=fluid.param_attr.ParamAttr(name=fc_name[1] + "_offset"))
+        fc2 = fluid.layers.dropout(x=fc2, dropout_prob=0.5)
+        out = fluid.layers.fc(
+            input=fc2,
+            size=class_dim,
+            act='softmax',
+            param_attr=fluid.param_attr.ParamAttr(name=fc_name[2] + "_weights"),
+            bias_attr=fluid.param_attr.ParamAttr(name=fc_name[2] + "_offset"))
+
+        return out
+
+    def conv_block(self, input, num_filter, groups, name=None):
+        conv = input
+        for i in range(groups):
+            conv = fluid.layers.conv2d(
+                input=conv,
+                num_filters=num_filter,
+                filter_size=3,
+                stride=1,
+                padding=1,
+                act='relu',
+                param_attr=fluid.param_attr.ParamAttr(
+                    name=name + str(i + 1) + "_weights"),
+                bias_attr=fluid.param_attr.ParamAttr(
+                    name=name + str(i + 1) + "_offset"))
+        return fluid.layers.pool2d(
+            input=conv, pool_size=2, pool_type='max', pool_stride=2)
+
+
+def VGG11():
+    model = VGGNet(layers=11)
+    return model
+
+
+def VGG13():
+    model = VGGNet(layers=13)
+    return model
+
+
+def VGG16():
+    model = VGGNet(layers=16)
+    return model
+
+
+def VGG19():
+    model = VGGNet(layers=19)
+    return model

fluid/PaddleCV/image_classification/reader.py

@@ -54,7 +54,7 @@ def random_crop(img, size, scale=[0.08, 1.0], ratio=[3. / 4., 4. / 3.]):
     scale_min = min(scale[0], bound)
 
     target_area = img.size[0] * img.size[1] * np.random.uniform(scale_min,
-                                                             scale_max)
+                                                                scale_max)
     target_size = math.sqrt(target_area)
     w = int(target_size * w)
     h = int(target_size * h)
@@ -169,7 +169,12 @@ def _reader_creator(file_list,
 def train(data_dir=DATA_DIR):
     file_list = os.path.join(data_dir, 'train_list.txt')
     return _reader_creator(
-        file_list, 'train', shuffle=True, color_jitter=False, rotate=False, data_dir=data_dir)
+        file_list,
+        'train',
+        shuffle=True,
+        color_jitter=False,
+        rotate=False,
+        data_dir=data_dir)
 
 
 def val(data_dir=DATA_DIR):

fluid/PaddleCV/image_classification/run.sh

+#Hyperparameters config
+python train.py \
+       --model=SE_ResNeXt50_32x4d \
+       --batch_size=32 \
+       --total_images=1281167 \
+       --class_dim=1000 \
+       --image_shape=3,224,224 \
+       --model_save_dir=output/ \
+       --with_mem_opt=False \
+       --lr_strategy=piecewise_decay \
+       --lr=0.1
+#      >log_SE_ResNeXt50_32x4d.txt 2>&1 &
+
+#AlexNet:
+#python train.py \
+#       --model=AlexNet \
+#       --batch_size=256 \
+#       --total_images=1281167 \
+#       --class_dim=1000 \
+#       --image_shape=3,224,224 \
+#       --model_save_dir=output/ \
+#       --with_mem_opt=False \
+#       --lr_strategy=piecewise_decay \
+#	--num_epochs=120 \
+#       --lr=0.01
+
+#VGG11:
+#python train.py \
+#       --model=VGG11 \
+#       --batch_size=512 \
+#       --total_images=1281167 \
+#       --class_dim=1000 \
+#       --image_shape=3,224,224 \
+#       --model_save_dir=output/ \
+#       --with_mem_opt=False \
+#       --lr_strategy=piecewise_decay \
+#       --num_epochs=120 \
+#       --lr=0.1
+
+
+#MobileNet v1:
+#python train.py \
+#       --model=MobileNet \
+#       --batch_size=256 \
+#       --total_images=1281167 \
+#       --class_dim=1000 \
+#       --image_shape=3,224,224 \
+#       --model_save_dir=output/ \
+#       --with_mem_opt=False \
+#       --lr_strategy=piecewise_decay \
+#	--num_epochs=120 \
+#       --lr=0.1
+
+
+#ResNet50:
+#python train.py \
+#       --model=ResNet50 \
+#       --batch_size=256 \
+#       --total_images=1281167 \
+#       --class_dim=1000 \
+#       --image_shape=3,224,224 \
+#       --model_save_dir=output/ \
+#       --with_mem_opt=False \
+#       --lr_strategy=piecewise_decay \
+#	--num_epochs=120 \
+#       --lr=0.1
+
+#ResNet101:
+#python train.py \
+#       --model=ResNet101 \
+#       --batch_size=256 \
+#       --total_images=1281167 \
+#       --class_dim=1000 \
+#       --image_shape=3,224,224 \
+#       --model_save_dir=output/ \
+#       --with_mem_opt=False \
+#       --lr_strategy=piecewise_decay \
+#	--num_epochs=120 \
+#       --lr=0.1
+

fluid/PaddleCV/image_classification/train.py

@@ -13,8 +13,13 @@ import paddle.dataset.flowers as flowers
 import models
 import reader
 import argparse
-from models.learning_rate import cosine_decay
+import functools
+import subprocess
+import utils
+from utils.learning_rate import cosine_decay
 from utility import add_arguments, print_arguments
+import models
+import models_name
 
 parser = argparse.ArgumentParser(description=__doc__)
 add_arg = functools.partial(add_arguments, argparser=parser)
@@ -34,20 +39,25 @@ add_arg('lr_strategy',      str,   "piecewise_decay",    "Set the learning rate
 add_arg('model',            str,   "SE_ResNeXt50_32x4d", "Set the network to use.")
 add_arg('enable_ce',        bool,  False,                "If set True, enable continuous evaluation job.")
 add_arg('data_dir',         str,   "./data/ILSVRC2012",  "The ImageNet dataset root dir.")
-# yapf: enable
+add_arg('model_category',   str,   "models",             "Whether to use models_name or not, valid value:'models','models_name'" )
+# yapf: enabl
 
-model_list = [m for m in dir(models) if "__" not in m]
+
+def set_models(model):
+    global models
+    if model == "models":
+        models = models
+    else:
+        models = models_name
 
 
 def optimizer_setting(params):
     ls = params["learning_strategy"]
-
     if ls["name"] == "piecewise_decay":
         if "total_images" not in params:
             total_images = 1281167
         else:
             total_images = params["total_images"]
-
         batch_size = ls["batch_size"]
         step = int(total_images / batch_size + 1)
 
@@ -60,6 +70,7 @@ def optimizer_setting(params):
                 boundaries=bd, values=lr),
             momentum=0.9,
             regularization=fluid.regularizer.L2Decay(1e-4))
+
     elif ls["name"] == "cosine_decay":
         if "total_images" not in params:
             total_images = 1281167
@@ -76,7 +87,29 @@ def optimizer_setting(params):
             learning_rate=cosine_decay(
                 learning_rate=lr, step_each_epoch=step, epochs=num_epochs),
             momentum=0.9,
-            regularization=fluid.regularizer.L2Decay(1e-4))
+            regularization=fluid.regularizer.L2Decay(4e-5))
+    elif ls["name"] == "exponential_decay":
+        if "total_images" not in params:
+            total_images = 1281167
+        else:
+            total_images = params["total_images"]
+        batch_size = ls["batch_size"]
+        step = int(total_images / batch_size +1)
+        lr = params["lr"]
+        num_epochs = params["num_epochs"]
+        learning_decay_rate_factor=ls["learning_decay_rate_factor"]
+        num_epochs_per_decay = ls["num_epochs_per_decay"]
+        NUM_GPUS = 1
+
+        optimizer = fluid.optimizer.Momentum(
+            learning_rate=fluid.layers.exponential_decay(
+                learning_rate = lr * NUM_GPUS,
+                decay_steps = step * num_epochs_per_decay / NUM_GPUS,
+                decay_rate = learning_decay_rate_factor),
+            momentum=0.9,
+
+            regularization = fluid.regularizer.L2Decay(4e-5))
+
     else:
         lr = params["lr"]
         optimizer = fluid.optimizer.Momentum(
@@ -86,29 +119,16 @@ def optimizer_setting(params):
 
     return optimizer
 
+def net_config(image, label, model, args):
+    model_list = [m for m in dir(models) if "__" not in m]
+    assert args.model in model_list,"{} is not lists: {}".format(
+        args.model, model_list)
 
-def train(args):
-    # parameters from arguments
     class_dim = args.class_dim
     model_name = args.model
-    checkpoint = args.checkpoint
-    pretrained_model = args.pretrained_model
-    with_memory_optimization = args.with_mem_opt
-    model_save_dir = args.model_save_dir
-    image_shape = [int(m) for m in args.image_shape.split(",")]
-
-    assert model_name in model_list, "{} is not in lists: {}".format(args.model,
-                                                                     model_list)
-
-    image = fluid.layers.data(name='image', shape=image_shape, dtype='float32')
-    label = fluid.layers.data(name='label', shape=[1], dtype='int64')
-
-    # model definition
-    model = models.__dict__[model_name]()
 
     if args.enable_ce:
         assert model_name == "SE_ResNeXt50_32x4d"
-        fluid.default_startup_program().random_seed = 1000
         model.params["dropout_seed"] = 100
         class_dim = 102
 
@@ -132,42 +152,101 @@ def train(args):
         acc_top1 = fluid.layers.accuracy(input=out, label=label, k=1)
         acc_top5 = fluid.layers.accuracy(input=out, label=label, k=5)
 
-    test_program = fluid.default_main_program().clone(for_test=True)
+    return avg_cost, acc_top1, acc_top5
+
+
+def build_program(is_train, main_prog, startup_prog, args):
+    image_shape = [int(m) for m in args.image_shape.split(",")]
+    model_name = args.model
+    model_list = [m for m in dir(models) if "__" not in m]
+    assert model_name in model_list, "{} is not in lists: {}".format(args.model,
+                                                                     model_list)
+    model = models.__dict__[model_name]()
+    with fluid.program_guard(main_prog, startup_prog):
+        py_reader = fluid.layers.py_reader(
+            capacity=16,
+            shapes=[[-1] + image_shape, [-1, 1]],
+            lod_levels=[0, 0],
+            dtypes=["float32", "int64"],
+            use_double_buffer=True)
+        with fluid.unique_name.guard():
+            image, label = fluid.layers.read_file(py_reader)
+            avg_cost, acc_top1, acc_top5 = net_config(image, label, model, args)
+            avg_cost.persistable = True
+            acc_top1.persistable = True
+            acc_top5.persistable = True
+            if is_train:
+                params = model.params
+                params["total_images"] = args.total_images
+                params["lr"] = args.lr
+                params["num_epochs"] = args.num_epochs
+                params["learning_strategy"]["batch_size"] = args.batch_size
+                params["learning_strategy"]["name"] = args.lr_strategy
+
+                optimizer = optimizer_setting(params)
+                optimizer.minimize(avg_cost)
+
+    return py_reader, avg_cost, acc_top1, acc_top5
+
+
+def train(args):
+    # parameters from arguments
+    model_name = args.model
+    checkpoint = args.checkpoint
+    pretrained_model = args.pretrained_model
+    with_memory_optimization = args.with_mem_opt
+    model_save_dir = args.model_save_dir
 
-    # parameters from model and arguments
-    params = model.params
-    params["total_images"] = args.total_images
-    params["lr"] = args.lr
-    params["num_epochs"] = args.num_epochs
-    params["learning_strategy"]["batch_size"] = args.batch_size
-    params["learning_strategy"]["name"] = args.lr_strategy
+    startup_prog = fluid.Program()
+    train_prog = fluid.Program()
+    test_prog = fluid.Program()
 
-    # initialize optimizer
-    optimizer = optimizer_setting(params)
-    opts = optimizer.minimize(avg_cost)
+    if args.enable_ce:
+        startup_prog.random_seed = 1000
+        train_prog.random_seed = 1000
+
+    train_py_reader, train_cost, train_acc1, train_acc5 = build_program(
+        is_train=True,
+        main_prog=train_prog,
+        startup_prog=startup_prog,
+        args=args)
+    test_py_reader, test_cost, test_acc1, test_acc5 = build_program(
+        is_train=False,
+        main_prog=test_prog,
+        startup_prog=startup_prog,
+        args=args)
+    test_prog = test_prog.clone(for_test=True)
 
     if with_memory_optimization:
-        fluid.memory_optimize(fluid.default_main_program())
+        fluid.memory_optimize(train_prog)
+        fluid.memory_optimize(test_prog)
 
     place = fluid.CUDAPlace(0) if args.use_gpu else fluid.CPUPlace()
     exe = fluid.Executor(place)
-    exe.run(fluid.default_startup_program())
+    exe.run(startup_prog)
 
     if checkpoint is not None:
-        fluid.io.load_persistables(exe, checkpoint)
+        fluid.io.load_persistables(exe, checkpoint, main_program=train_prog)
 
     if pretrained_model:
 
         def if_exist(var):
             return os.path.exists(os.path.join(pretrained_model, var.name))
 
-        fluid.io.load_vars(exe, pretrained_model, predicate=if_exist)
+        fluid.io.load_vars(
+            exe, pretrained_model, main_program=train_prog, predicate=if_exist)
 
-    train_batch_size = args.batch_size
-    test_batch_size = 16
+    visible_device = os.getenv('CUDA_VISIBLE_DEVICES')
+    if visible_device:
+        device_num = len(visible_device.split(','))
+    else:
+        device_num = subprocess.check_output(['nvidia-smi', '-L']).count('\n')
 
+    train_batch_size = args.batch_size / device_num
+    test_batch_size = 8
     if not args.enable_ce:
-        train_reader = paddle.batch(reader.train(), batch_size=train_batch_size)
+        train_reader = paddle.batch(
+            reader.train(), batch_size=train_batch_size, drop_last=True)
         test_reader = paddle.batch(reader.val(), batch_size=test_batch_size)
     else:
         # use flowers dataset for CE and set use_xmap False to avoid disorder data
@@ -176,89 +255,105 @@ def train(args):
         random.seed(0)
         np.random.seed(0)
         train_reader = paddle.batch(
-            flowers.train(use_xmap=False), batch_size=train_batch_size)
+            flowers.train(use_xmap=False),
+            batch_size=train_batch_size,
+            drop_last=True)
         test_reader = paddle.batch(
             flowers.test(use_xmap=False), batch_size=test_batch_size)
 
-    feeder = fluid.DataFeeder(place=place, feed_list=[image, label])
-
+    train_py_reader.decorate_paddle_reader(train_reader)
+    test_py_reader.decorate_paddle_reader(test_reader)
     train_exe = fluid.ParallelExecutor(
-        use_cuda=True if args.use_gpu else False, loss_name=avg_cost.name)
+        main_program=train_prog,
+        use_cuda=bool(args.use_gpu),
+        loss_name=train_cost.name)
+
+    train_fetch_list = [train_cost.name, train_acc1.name, train_acc5.name]
+    test_fetch_list = [test_cost.name, test_acc1.name, test_acc5.name]
 
-    fetch_list = [avg_cost.name, acc_top1.name, acc_top5.name]
+    params = models.__dict__[args.model]().params
 
-    gpu = os.getenv("CUDA_VISIBLE_DEVICES") or ""
-    gpu_nums = len(gpu.split(","))
     for pass_id in range(params["num_epochs"]):
+
+        train_py_reader.start()
+
         train_info = [[], [], []]
         test_info = [[], [], []]
         train_time = []
-        for batch_id, data in enumerate(train_reader()):
-            t1 = time.time()
-            loss, acc1, acc5 = train_exe.run(fetch_list, feed=feeder.feed(data))
-            t2 = time.time()
-            period = t2 - t1
-            loss = np.mean(np.array(loss))
-            acc1 = np.mean(np.array(acc1))
-            acc5 = np.mean(np.array(acc5))
-            train_info[0].append(loss)
-            train_info[1].append(acc1)
-            train_info[2].append(acc5)
-            train_time.append(period)
-            if batch_id % 10 == 0:
-                print("Pass {0}, trainbatch {1}, loss {2}, \
-                       acc1 {3}, acc5 {4} time {5}"
-                                                   .format(pass_id, \
-                       batch_id, loss, acc1, acc5, \
-                       "%2.2f sec" % period))
-                sys.stdout.flush()
+        batch_id = 0
+        try:
+            while True:
+                t1 = time.time()
+                loss, acc1, acc5 = train_exe.run(fetch_list=train_fetch_list)
+                t2 = time.time()
+                period = t2 - t1
+                loss = np.mean(np.array(loss))
+                acc1 = np.mean(np.array(acc1))
+                acc5 = np.mean(np.array(acc5))
+                train_info[0].append(loss)
+                train_info[1].append(acc1)
+                train_info[2].append(acc5)
+                train_time.append(period)
+                if batch_id % 10 == 0:
+                    print("Pass {0}, trainbatch {1}, loss {2}, \
+                        acc1 {3}, acc5 {4} time {5}"
+                          .format(pass_id, batch_id, loss, acc1, acc5,
+                                  "%2.2f sec" % period))
+                    sys.stdout.flush()
+                batch_id += 1
+        except fluid.core.EOFException:
+            train_py_reader.reset()
 
         train_loss = np.array(train_info[0]).mean()
         train_acc1 = np.array(train_info[1]).mean()
         train_acc5 = np.array(train_info[2]).mean()
         train_speed = np.array(train_time).mean() / train_batch_size
-        cnt = 0
-        for test_batch_id, data in enumerate(test_reader()):
-            t1 = time.time()
-            loss, acc1, acc5 = exe.run(test_program,
-                                       fetch_list=fetch_list,
-                                       feed=feeder.feed(data))
-            t2 = time.time()
-            period = t2 - t1
-            loss = np.mean(loss)
-            acc1 = np.mean(acc1)
-            acc5 = np.mean(acc5)
-            test_info[0].append(loss * len(data))
-            test_info[1].append(acc1 * len(data))
-            test_info[2].append(acc5 * len(data))
-            cnt += len(data)
-            if test_batch_id % 10 == 0:
-                print("Pass {0},testbatch {1},loss {2}, \
-                       acc1 {3},acc5 {4},time {5}"
-                                                  .format(pass_id, \
-                       test_batch_id, loss, acc1, acc5, \
-                       "%2.2f sec" % period))
-                sys.stdout.flush()
-
-        test_loss = np.sum(test_info[0]) / cnt
-        test_acc1 = np.sum(test_info[1]) / cnt
-        test_acc5 = np.sum(test_info[2]) / cnt
+
+        test_py_reader.start()
+
+        test_batch_id = 0
+        try:
+            while True:
+                t1 = time.time()
+                loss, acc1, acc5 = exe.run(program=test_prog,
+                                           fetch_list=test_fetch_list)
+                t2 = time.time()
+                period = t2 - t1
+                loss = np.mean(loss)
+                acc1 = np.mean(acc1)
+                acc5 = np.mean(acc5)
+                test_info[0].append(loss)
+                test_info[1].append(acc1)
+                test_info[2].append(acc5)
+                if test_batch_id % 10 == 0:
+                    print("Pass {0},testbatch {1},loss {2}, \
+                        acc1 {3},acc5 {4},time {5}"
+                          .format(pass_id, test_batch_id, loss, acc1, acc5,
+                                  "%2.2f sec" % period))
+                    sys.stdout.flush()
+                test_batch_id += 1
+        except fluid.core.EOFException:
+            test_py_reader.reset()
+
+        test_loss = np.array(test_info[0]).mean()
+        test_acc1 = np.array(test_info[1]).mean()
+        test_acc5 = np.array(test_info[2]).mean()
 
         print("End pass {0}, train_loss {1}, train_acc1 {2}, train_acc5 {3}, "
-              "test_loss {4}, test_acc1 {5}, test_acc5 {6}".format(pass_id, \
-              train_loss, train_acc1, train_acc5, test_loss, test_acc1, \
-              test_acc5))
+              "test_loss {4}, test_acc1 {5}, test_acc5 {6}".format(
+                  pass_id, train_loss, train_acc1, train_acc5, test_loss,
+                  test_acc1, test_acc5))
         sys.stdout.flush()
 
         model_path = os.path.join(model_save_dir + '/' + model_name,
                                   str(pass_id))
         if not os.path.isdir(model_path):
             os.makedirs(model_path)
-        fluid.io.save_persistables(exe, model_path)
+        fluid.io.save_persistables(exe, model_path, main_program=train_prog)
 
         # This is for continuous evaluation only
         if args.enable_ce and pass_id == args.num_epochs - 1:
-            if gpu_nums == 1:
+            if device_num == 1:
                 # Use the mean cost/acc for training
                 print("kpis	train_cost	%s" % train_loss)
                 print("kpis	train_acc_top1	%s" % train_acc1)
@@ -270,18 +365,24 @@ def train(args):
                 print("kpis	train_speed	%s" % train_speed)
             else:
                 # Use the mean cost/acc for training
-                print("kpis	train_cost_card%s	%s" % (gpu_nums, train_loss))
-                print("kpis	train_acc_top1_card%s	%s" % (gpu_nums, train_acc1))
-                print("kpis	train_acc_top5_card%s	%s" % (gpu_nums, train_acc5))
+                print("kpis	train_cost_card%s	%s" % (device_num, train_loss))
+                print("kpis	train_acc_top1_card%s	%s" %
+                      (device_num, train_acc1))
+                print("kpis	train_acc_top5_card%s	%s" %
+                      (device_num, train_acc5))
                 # Use the mean cost/acc for testing
-                print("kpis	test_cost_card%s	%s" % (gpu_nums, test_loss))
-                print("kpis	test_acc_top1_card%s	%s" % (gpu_nums, test_acc1))
-                print("kpis	test_acc_top5_card%s	%s" % (gpu_nums, test_acc5))
-                print("kpis	train_speed_card%s	%s" % (gpu_nums, train_speed))
+                print("kpis	test_cost_card%s	%s" % (device_num, test_loss))
+                print("kpis	test_acc_top1_card%s	%s" % (device_num, test_acc1))
+                print("kpis	test_acc_top5_card%s	%s" % (device_num, test_acc5))
+                print("kpis	train_speed_card%s	%s" % (device_num, train_speed))
 
 
 def main():
     args = parser.parse_args()
+    models_now = args.model_category
+    assert models_now in ["models", "models_name"], "{} is not in lists: {}".format(
+            models_now, ["models", "models_name"])
+    set_models(models_now)
     print_arguments(args)
     train(args)
 

fluid/PaddleCV/image_classification/utils/__init__.py

+from .learning_rate import cosine_decay, lr_warmup

fluid/PaddleCV/image_classification/models/learning_rate.py → fluid/PaddleCV/image_classification/utils/learning_rate.py

@@ -27,8 +27,8 @@ def lr_warmup(learning_rate, warmup_steps, start_lr, end_lr):
         Argument learning_rate can be float or a Variable
         lr = lr + (warmup_rate * step / warmup_steps)
     """
-    assert(isinstance(end_lr, float))
-    assert(isinstance(start_lr, float))
+    assert (isinstance(end_lr, float))
+    assert (isinstance(start_lr, float))
     linear_step = end_lr - start_lr
     with fluid.default_main_program()._lr_schedule_guard():
         lr = fluid.layers.tensor.create_global_var(
@@ -42,9 +42,10 @@ def lr_warmup(learning_rate, warmup_steps, start_lr, end_lr):
 
         with fluid.layers.control_flow.Switch() as switch:
             with switch.case(global_step < warmup_steps):
-                decayed_lr = start_lr + linear_step * (global_step / warmup_steps)
+                decayed_lr = start_lr + linear_step * (global_step /
+                                                       warmup_steps)
                 fluid.layers.tensor.assign(decayed_lr, lr)
             with switch.default():
                 fluid.layers.tensor.assign(learning_rate, lr)
 
-        return lr
+        return lr
\ No newline at end of file

fluid/PaddleNLP/language_model/lstm/.run_ce.sh

@@ -7,5 +7,6 @@ python  train.py \
     --data_path data/simple-examples/data/ \
     --model_type small \
     --use_gpu True \
+    --rnn_model static \
     --enable_ce | python _ce.py
 

fluid/PaddleNLP/language_model/lstm/args.py

@@ -26,7 +26,12 @@ def parse_args():
         "--model_type",
         type=str,
         default="small",
-        help="model_type [test|small|med|big]")
+        help="model_type [test|small|medium|large]")
+    parser.add_argument(
+        "--rnn_model",
+        type=str,
+        default="static",
+        help="model_type [static|padding|cudnn]")
     parser.add_argument(
         "--data_path", type=str, help="all the data for train,valid,test")
     parser.add_argument('--para_init', action='store_true')

fluid/PaddleNLP/language_model/lstm/lm_model.py

@@ -28,7 +28,8 @@ def lm_model(hidden_size,
              num_layers=2,
              num_steps=20,
              init_scale=0.1,
-             dropout=None):
+             dropout=None, 
+             rnn_model='static'):
     def padding_rnn(input_embedding, len=3, init_hidden=None, init_cell=None):
         weight_1_arr = []
         weight_2_arr = []
@@ -243,7 +244,7 @@ def lm_model(hidden_size,
         input=x,
         size=[vocab_size, hidden_size],
         dtype='float32',
-        is_sparse=True,
+        is_sparse=False,
         param_attr=fluid.ParamAttr(
             name='embedding_para',
             initializer=fluid.initializer.UniformInitializer(
@@ -255,9 +256,22 @@ def lm_model(hidden_size,
             x_emb,
             dropout_prob=dropout,
             dropout_implementation='upscale_in_train')
-
-    rnn_out, last_hidden, last_cell = padding_rnn(
-        x_emb, len=num_steps, init_hidden=init_hidden, init_cell=init_cell)
+    
+    if rnn_model == "padding":
+        rnn_out, last_hidden, last_cell = padding_rnn(
+            x_emb, len=num_steps, init_hidden=init_hidden, init_cell=init_cell)
+    elif rnn_model == "static":
+        rnn_out, last_hidden, last_cell = encoder_static(
+            x_emb, len=num_steps, init_hidden=init_hidden, init_cell=init_cell)
+    elif rnn_model == "cudnn":
+        x_emb = layers.transpose( x_emb, perm=[1, 0, 2])
+        rnn_out, last_hidden, last_cell = layers.lstm( x_emb, init_hidden, init_cell,  num_steps, hidden_size, num_layers, \
+                is_bidirec=False, \
+                default_initializer=fluid.initializer.UniformInitializer(low=-init_scale, high=init_scale) )
+        rnn_out = layers.transpose( rnn_out, perm=[1, 0, 2])
+    else:
+        print( "type not support")
+        return
     rnn_out = layers.reshape(rnn_out, shape=[-1, num_steps, hidden_size])
 
 

fluid/PaddleNLP/language_model/lstm/train.py

@@ -77,6 +77,7 @@ def save_para_npz(train_prog, train_exe):
 def train():
     args = parse_args()
     model_type = args.model_type
+    rnn_model = args.rnn_model
     logger = logging.getLogger("lm")
     logger.setLevel(logging.INFO)
     formatter = logging.Formatter(
@@ -157,7 +158,8 @@ def train():
         num_layers=num_layers,
         num_steps=num_steps,
         init_scale=init_scale,
-        dropout=dropout)
+        dropout=dropout, 
+        rnn_model=rnn_model)
     # clone from default main program and use it as the validation program
     main_program = fluid.default_main_program()
     inference_program = fluid.default_main_program().clone(for_test=True)
@@ -206,18 +208,19 @@ def train():
 
     def eval(data):
         # when eval the batch_size set to 1
-        eval_data_iter = reader.get_data_iter(data, 1, num_steps)
+        eval_data_iter = reader.get_data_iter(data, batch_size, num_steps)
         total_loss = 0.0
         iters = 0
-        init_hidden = np.zeros((num_layers, 1, hidden_size), dtype='float32')
-        init_cell = np.zeros((num_layers, 1, hidden_size), dtype='float32')
+        init_hidden = np.zeros((num_layers, batch_size, hidden_size), dtype='float32')
+        init_cell = np.zeros((num_layers, batch_size, hidden_size), dtype='float32')
         for batch_id, batch in enumerate(eval_data_iter):
             input_data_feed = prepare_input(
                 batch, init_hidden, init_cell, epoch_id, with_lr=False)
             fetch_outs = exe.run(
                 inference_program,
                 feed=input_data_feed,
-                fetch_list=[loss.name, last_hidden.name, last_cell.name])
+                fetch_list=[loss.name, last_hidden.name, last_cell.name],
+                use_program_cache=True)
 
             cost_train = np.array(fetch_outs[0])
             init_hidden = np.array(fetch_outs[1])
@@ -284,7 +287,7 @@ def train():
 
         if epoch_id == max_epoch - 1 and args.enable_ce:
             print("ptblm\tlstm_language_model_duration\t%s" %
-                  (total_time / max_epoch))
+                        (total_time / max_epoch))
             print("ptblm\tlstm_language_model_loss\t%s" % ppl[0])
 
         model_path = os.path.join("model_new/", str(epoch_id))

fluid/PaddleRec/ctr/reader.py

@@ -35,7 +35,7 @@ class CriteoDataset(Dataset):
                             else:
                                 dense_feature.append((float(features[idx]) - self.cont_min_[idx - 1]) / self.cont_diff_[idx - 1])
                         for idx in self.categorical_range_:
-                            sparse_feature.append([hash("%d_%s" % (idx, features[idx])) % self.hash_dim_])
+                            sparse_feature.append([hash(str(idx) + features[idx]) % self.hash_dim_])
 
                         label = [int(features[0])]
                         yield [dense_feature] + sparse_feature + [label]

fluid/PaddleRec/gru4rec/train.py

@@ -69,82 +69,54 @@ def train():
     
     # Initialize executor
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-    training_role = os.getenv("TRAINING_ROLE", "TRAINER")
-    if training_role == "PSERVER":
-        place = fluid.CPUPlace()
     exe = fluid.Executor(place)
+    exe.run(fluid.default_startup_program())
     if parallel:
         train_exe = fluid.ParallelExecutor(
-            use_cuda=use_cuda, loss_name=avg_cost.name)
+            use_cuda=use_cuda,
+            loss_name=avg_cost.name)
     else:
         train_exe = exe
     
-    def train_loop(main_program):
-        """ train network """
-        pass_num = args.pass_num
-        model_dir = args.model_dir
-        fetch_list = [avg_cost.name]
+    pass_num = args.pass_num
+    model_dir = args.model_dir
+    fetch_list = [avg_cost.name]
 
-        exe.run(fluid.default_startup_program())
-        total_time = 0.0
-        for pass_idx in six.moves.xrange(pass_num):
-            epoch_idx = pass_idx + 1
-            print("epoch_%d start" % epoch_idx)
+    total_time = 0.0
+    for pass_idx in six.moves.xrange(pass_num):
+        epoch_idx = pass_idx + 1
+        print("epoch_%d start" % epoch_idx)
 
-            t0 = time.time()
-            i = 0
-            newest_ppl = 0
-            for data in train_reader():
-                i += 1
-                lod_src_wordseq = utils.to_lodtensor([dat[0] for dat in data],
-                                                     place)
-                lod_dst_wordseq = utils.to_lodtensor([dat[1] for dat in data],
-                                                     place)
-                ret_avg_cost = train_exe.run(main_program,
-                                            feed={  "src_wordseq": lod_src_wordseq,
-                                                    "dst_wordseq": lod_dst_wordseq},
-                                            fetch_list=fetch_list)
-                avg_ppl = np.exp(ret_avg_cost[0])
-                newest_ppl = np.mean(avg_ppl)
-                if i % args.print_batch == 0:
-                    print("step:%d ppl:%.3f" % (i, newest_ppl))
+        t0 = time.time()
+        i = 0
+        newest_ppl = 0
+        for data in train_reader():
+            i += 1
+            lod_src_wordseq = utils.to_lodtensor([dat[0] for dat in data],
+                                                 place)
+            lod_dst_wordseq = utils.to_lodtensor([dat[1] for dat in data],
+                                                 place)
+            ret_avg_cost = train_exe.run(feed={  
+                                                "src_wordseq": lod_src_wordseq,
+                                                "dst_wordseq": lod_dst_wordseq},
+                                        fetch_list=fetch_list)
+            avg_ppl = np.exp(ret_avg_cost[0])
+            newest_ppl = np.mean(avg_ppl)
+            if i % args.print_batch == 0:
+                print("step:%d ppl:%.3f" % (i, newest_ppl))
 
-            t1 = time.time()
-            total_time += t1 - t0
-            print("epoch:%d num_steps:%d time_cost(s):%f" %
-                  (epoch_idx, i, total_time / epoch_idx))
-            save_dir = "%s/epoch_%d" % (model_dir, epoch_idx)
-            feed_var_names = ["src_wordseq", "dst_wordseq"]
-            fetch_vars = [avg_cost, acc]
-            fluid.io.save_inference_model(save_dir, feed_var_names, fetch_vars, exe)
-            print("model saved in %s" % save_dir)
-        #exe.close()
-        print("finish training")
+        t1 = time.time()
+        total_time += t1 - t0
+        print("epoch:%d num_steps:%d time_cost(s):%f" %
+              (epoch_idx, i, total_time / epoch_idx))
+        save_dir = "%s/epoch_%d" % (model_dir, epoch_idx)
+        feed_var_names = ["src_wordseq", "dst_wordseq"]
+        fetch_vars = [avg_cost, acc]
+        fluid.io.save_inference_model(save_dir, feed_var_names, fetch_vars, exe)
+        print("model saved in %s" % save_dir)
+    #exe.close()
+    print("finish training")
 
-    if args.is_local:
-        print("run local training")
-        train_loop(fluid.default_main_program())
 
-    else:
-        print("run distribute training")
-        port = os.getenv("PADDLE_PORT", "6174")
-        pserver_ips = os.getenv("PADDLE_PSERVERS")  # ip,ip...
-        eplist = []
-        for ip in pserver_ips.split(","):
-            eplist.append(':'.join([ip, port]))
-        pserver_endpoints = ",".join(eplist)  # ip:port,ip:port...
-        trainers = int(os.getenv("PADDLE_TRAINERS_NUM", "0"))
-        current_endpoint = os.getenv("POD_IP") + ":" + port
-        trainer_id = int(os.getenv("PADDLE_TRAINER_ID", "0"))
-        t = fluid.DistributeTranspiler()
-        t.transpile(trainer_id, pservers=pserver_endpoints, trainers=trainers)
-        if training_role == "PSERVER":
-            pserver_prog = t.get_pserver_program(current_endpoint)
-            pserver_startup = t.get_startup_program(current_endpoint,
-                                                    pserver_prog)
-            exe.run(pserver_startup)
-            exe.run(pserver_prog)
-        elif training_role == "TRAINER":
-            train_loop(t.get_trainer_program())
 if __name__ == "__main__":
     train()

fluid/PaddleRec/multiview_simnet/README.cn.md

@@ -15,8 +15,13 @@
 ```bash
 python train.py
 ```
+## 
+如下
+如下命令行可以获得预测工具的具体选项，`python infer -h`内容可以参考说明
+```bash
+python infer.py
+```
 ## 未来的工作
 - 多种pairwise的损失函数会被加入到这个项目中。对于不同视角的特征，用户-项目之间的匹配关系可以使用不同的损失函数进行联合优化。整个模型会在真实数据中进行验证。
-- 推理工具会被加入
 - Parallel Executor选项会被加入
 - 分布式训练能力会被加入

fluid/PaddleRec/multiview_simnet/README.md

@@ -15,8 +15,13 @@ The command line options for training can be listed by `python train.py -h`
 python train.py
 ```
 
+## Infer
+The command line options for inference can be listed by `python infer.py -h`
+```bash
+python infer.py
+```
+
 ## Future work
 - Multiple types of pairwise loss will be added in this project. For different views of features between a user and an item, multiple losses will be supported. The model will be verified in real world dataset.
-- infer will be added
 - Parallel Executor will be added in this project
 - Distributed Training will be added

fluid/PaddleRec/multiview_simnet/infer.py

+# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import os
+import sys
+import time
+import six
+import numpy as np
+import math
+import argparse
+import logging
+import paddle.fluid as fluid
+import paddle
+import time
+import reader as reader
+from nets import MultiviewSimnet, SimpleEncoderFactory
+
+logging.basicConfig(format="%(asctime)s - %(levelname)s - %(message)s")
+logger = logging.getLogger("fluid")
+logger.setLevel(logging.INFO)
+
+
+def parse_args():
+    parser = argparse.ArgumentParser("multi-view simnet")
+    parser.add_argument(
+        "--train_file", type=str, help="Training file")
+    parser.add_argument(
+        "--valid_file", type=str, help="Validation file")
+    parser.add_argument(
+        "--epochs", type=int, default=10, help="Number of epochs for training")
+    parser.add_argument(
+        "--model_dir", type=str, default='model_output', help="Model output folder")
+    parser.add_argument(
+        "--query_slots", type=int, default=1, help="Number of query slots")
+    parser.add_argument(
+        "--title_slots", type=int, default=1, help="Number of title slots")
+    parser.add_argument(
+        "--query_encoder", type=str, default="bow", help="Encoder module for slot encoding")
+    parser.add_argument(
+        "--title_encoder", type=str, default="bow", help="Encoder module for slot encoding")
+    parser.add_argument(
+        "--query_encode_dim", type=int, default=128, help="Dimension of query encoder output")
+    parser.add_argument(
+        "--title_encode_dim", type=int, default=128, help="Dimension of title encoder output")
+    parser.add_argument(
+        "--batch_size", type=int, default=128, help="Batch size for training")
+    parser.add_argument(
+        "--embedding_dim", type=int, default=128, help="Default Dimension of Embedding")
+    parser.add_argument(
+        "--sparse_feature_dim", type=int, default=1000001, help="Sparse feature hashing space for index processing")
+    parser.add_argument(
+        "--hidden_size", type=int, default=128, help="Hidden dim")
+    return parser.parse_args()
+
+
+def start_infer(args, model_path):
+    dataset = reader.SyntheticDataset(args.sparse_feature_dim, args.query_slots,
+                                      args.title_slots)
+    test_reader = paddle.batch(
+        paddle.reader.shuffle(
+            dataset.valid(), buf_size=args.batch_size * 100),
+        batch_size=args.batch_size)
+    place = fluid.CPUPlace()
+    exe = fluid.Executor(place)
+    
+    with fluid.scope_guard(fluid.core.Scope()):
+        infer_program, feed_target_names, fetch_vars = fluid.io.load_inference_model(
+                args.model_dir, exe)
+        t0 = time.time()
+        step_id = 0
+        feeder = fluid.DataFeeder(program=infer_program, feed_list=feed_target_names, place=place)
+        for batch_id, data in enumerate(test_reader()):
+            step_id += 1
+            loss_val, correct_val = exe.run(infer_program,
+                                            feed=feeder.feed(data),
+                                            fetch_list=fetch_vars)
+            logger.info("TRAIN --> pass: {} batch_id: {} avg_cost: {}, acc: {}"
+                        .format(step_id, batch_id, loss_val,
+                                float(correct_val) / args.batch_size))
+
+def main():
+    args = parse_args()
+    start_infer(args, args.model_dir)
+
+if __name__ == "__main__":
+    main()

fluid/PaddleRec/tagspace/README.md

@@ -7,17 +7,27 @@
 ├── README.md            # 文档
 ├── train.py             # 训练脚本
 ├── infer.py             # 预测脚本
+├── net.py               # 网络结构
+├── text2paddle.py       # 文本数据转paddle数据
+├── cluster_train.py     # 多机训练
+├── cluster_train.sh     # 多机训练脚本
 ├── utils                # 通用函数
-├── small_train.txt      # 小样本训练集
-└── small_test.txt       # 小样本测试集
+├── vocab_text.txt       # 小样本文本字典
+├── vocab_tag.txt        # 小样本类别字典
+├── train_data           # 小样本训练目录
+└── test_data            # 小样本测试目录
 
 ```
 
 
 ## 简介
 
-TagSpace模型的介绍可以参阅论文[#TagSpace: Semantic Embeddings from Hashtags](https://research.fb.com/publications/tagspace-semantic-embeddings-from-hashtags/)，在本例中，我们实现了TagSpace的模型。
-## 数据下载
+TagSpace模型的介绍可以参阅论文[#TagSpace: Semantic Embeddings from Hashtags](https://research.fb.com/publications/tagspace-semantic-embeddings-from-hashtags/)。
+
+Tagspace模型学习文本及标签的embedding表示，应用于工业级的标签推荐，具体应用场景有feed新闻标签推荐。
+
+
+## 数据下载及预处理
 
 [ag news dataset](https://github.com/mhjabreel/CharCNN/tree/master/data/ag_news_csv)
 
@@ -27,23 +37,54 @@ TagSpace模型的介绍可以参阅论文[#TagSpace: Semantic Embeddings from Ha
 "3","Wall St. Bears Claw Back Into the Black (Reuters)","Reuters - Short-sellers, Wall Street's dwindling\band of ultra-cynics, are seeing green again."
 ```
 
-## 训练
-'--use_cuda 1' 表示使用gpu, 缺省表示使用cpu 
+将文本数据转为paddle数据，先将数据放到训练数据目录和测试数据目录
+```
+mv train.csv raw_big_train_data
+mv test.csv raw_big_test_data
+```
+
+运行脚本text2paddle.py 生成paddle输入格式
+```
+python text2paddle.py raw_big_train_data/ raw_big_test_data/ train_big_data test_big_data big_vocab_text.txt big_vocab_tag.txt
+```
+
+## 单机训练
+'--use_cuda 1' 表示使用gpu, 0表示使用cpu, '--parallel 1' 表示使用多卡
+
+小数据训练（样例中的数据已经准备，可跳过上一节的数据准备，直接运行命令）
 
 GPU 环境
-运行命令 `CUDA_VISIBLE_DEVICES=0 python train.py train_file test_file --use_cuda 1` 开始训练模型。
 ```
-CUDA_VISIBLE_DEVICES=0 python train.py small_train.txt small_test.txt --use_cuda 1
+CUDA_VISIBLE_DEVICES=0 python train.py  --use_cuda 1
 ```
 CPU 环境
-运行命令 `python train.py train_file test_file` 开始训练模型。
 ```
-python train.py small_train.txt small_test.txt
+python train.py 
+```
+
+全量数据单机单卡训练
+```
+CUDA_VISIBLE_DEVICES=0 python train.py --use_cuda 1 --train_dir train_big_data/ --vocab_text_path big_vocab_text.txt --vocab_tag_path big_vocab_tag.txt --model_dir big_model --batch_size 500
+```
+全量数据单机多卡训练
+
+```
+python train.py --train_dir train_big_data/ --vocab_text_path big_vocab_text.txt --vocab_tag_path big_vocab_tag.txt --model_dir big_model --batch_size 500 --parallel 1
 ```
 
 ## 预测
+小数据预测
+```
+python infer.py
+```
 
+全量数据预测
 ```
-CUDA_VISIBLE_DEVICES=0 python infer.py model/ 1 10 small_train.txt small_test.txt --use_cuda 1
+python infer.py --model_dir big_model --vocab_tag_path big_vocab_tag.txt --test_dir test_big_data/
 ```
 
+## 本地模拟多机
+运行命令
+```
+sh cluster_train.py
+```

fluid/PaddleRec/tagspace/cluster_train.py

+import os
+import sys
+import time
+import six
+import numpy as np
+import math
+import argparse
+import paddle
+import paddle.fluid as fluid
+import time
+import utils
+import net
+
+SEED = 102
+
+def parse_args():
+    parser = argparse.ArgumentParser("TagSpace benchmark.")
+    parser.add_argument(
+        '--neg_size', type=int, default=3, help='neg/pos ratio')
+    parser.add_argument(
+        '--train_dir', type=str, default='train_data', help='train file address')
+    parser.add_argument(
+        '--vocab_text_path', type=str, default='vocab_text.txt', help='vocab_text file address')
+    parser.add_argument(
+        '--vocab_tag_path', type=str, default='vocab_tag.txt', help='vocab_text file address')
+    parser.add_argument(
+        '--is_local', type=int, default=1, help='whether local')
+    parser.add_argument(
+        '--model_dir', type=str, default='model_', help='model dir')
+    parser.add_argument(
+        '--batch_size', type=int, default=5, help='num of batch size')
+    parser.add_argument(
+        '--print_batch', type=int, default=10, help='num of print batch')
+    parser.add_argument(
+        '--pass_num', type=int, default=10, help='num of epoch')
+    parser.add_argument(
+        '--use_cuda', type=int, default=0, help='whether use gpu')
+    parser.add_argument(
+        '--base_lr', type=float, default=0.01, help='learning rate')
+    parser.add_argument(
+        '--num_devices', type=int, default=1, help='Number of GPU devices')
+    parser.add_argument(
+        '--role', type=str, default='pserver', help='trainer or pserver')
+    parser.add_argument(
+        '--endpoints', type=str, default='127.0.0.1:6000', help='The pserver endpoints, like: 127.0.0.1:6000, 127.0.0.1:6001')
+    parser.add_argument(
+        '--current_endpoint', type=str, default='127.0.0.1:6000', help='The current_endpoint')
+    parser.add_argument(
+        '--trainer_id', type=int, default=0, help='trainer id ,only trainer_id=0 save model')
+    parser.add_argument(
+        '--trainers', type=int, default=1, help='The num of trianers, (default: 1)')
+    args = parser.parse_args()
+    return args
+
+def get_cards(args):
+    return args.num_devices
+
+def train():
+    """ do training """
+    args = parse_args()
+    train_dir = args.train_dir
+    vocab_text_path = args.vocab_text_path
+    vocab_tag_path = args.vocab_tag_path
+    use_cuda = True if args.use_cuda else False
+    batch_size = args.batch_size
+    neg_size = args.neg_size
+    vocab_text_size, vocab_tag_size, train_reader = utils.prepare_data(
+        file_dir=train_dir, vocab_text_path=vocab_text_path, 
+        vocab_tag_path=vocab_tag_path, neg_size=neg_size, 
+        batch_size=batch_size * get_cards(args), 
+        buffer_size=batch_size*100, is_train=True)
+    """ train network """
+    # Train program
+    avg_cost, correct, cos_pos = net.network(vocab_text_size, vocab_tag_size, neg_size=neg_size)
+
+    # Optimization to minimize lost
+    sgd_optimizer = fluid.optimizer.SGD(learning_rate=args.base_lr)
+    sgd_optimizer.minimize(avg_cost)
+
+    def train_loop(main_program):
+        # Initialize executor
+        place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
+        exe = fluid.Executor(place)
+        pass_num = args.pass_num
+        model_dir = args.model_dir
+        fetch_list = [avg_cost.name]
+        exe.run(fluid.default_startup_program())
+        total_time = 0.0
+        for pass_idx in range(pass_num):
+            epoch_idx = pass_idx + 1
+            print("epoch_%d start" % epoch_idx)
+            t0 = time.time()
+            for batch_id, data in enumerate(train_reader()):
+                lod_text_seq = utils.to_lodtensor([dat[0] for dat in data], place)
+                lod_pos_tag = utils.to_lodtensor([dat[1] for dat in data], place)
+                lod_neg_tag = utils.to_lodtensor([dat[2] for dat in data], place)
+                loss_val, correct_val = exe.run(
+                        feed={
+                            "text": lod_text_seq,
+                            "pos_tag": lod_pos_tag,
+                            "neg_tag": lod_neg_tag},
+                        fetch_list=[avg_cost.name, correct.name])
+                if batch_id % args.print_batch == 0:
+                    print("TRAIN --> pass: {} batch_num: {} avg_cost: {}, acc: {}"
+                            .format(pass_idx, (batch_id+10) * batch_size, np.mean(loss_val),
+                                    float(np.sum(correct_val)) / batch_size))
+            t1 = time.time()
+            total_time += t1 - t0
+            print("epoch:%d num_steps:%d time_cost(s):%f" %
+                  (epoch_idx, batch_id, total_time / epoch_idx))
+            save_dir = "%s/epoch_%d" % (model_dir, epoch_idx)
+            feed_var_names = ["text", "pos_tag"]
+            fetch_vars = [cos_pos]
+            fluid.io.save_inference_model(save_dir, feed_var_names, fetch_vars, exe)
+        print("finish training")
+    
+    if args.is_local:
+        print("run local training")
+        train_loop(fluid.default_main_program())
+    else:
+        print("run distribute training")
+        t = fluid.DistributeTranspiler()
+        t.transpile(args.trainer_id, pservers=args.endpoints, trainers=args.trainers)
+        if args.role == "pserver":
+            print("run psever")
+            pserver_prog = t.get_pserver_program(args.current_endpoint)
+            pserver_startup = t.get_startup_program(args.current_endpoint,
+                                                    pserver_prog)
+            exe = fluid.Executor(fluid.CPUPlace())
+            exe.run(pserver_startup)
+            exe.run(pserver_prog)
+        elif args.role == "trainer":
+            print("run trainer")
+            train_loop(t.get_trainer_program())
+
+if __name__ == "__main__":
+    train()

fluid/PaddleRec/tagspace/cluster_train.sh

+#!/bin/bash
+
+#export GLOG_v=30
+#export GLOG_logtostderr=1
+
+# start pserver0
+python cluster_train.py \
+    --train_dir train_data \
+    --model_dir cluster_model \
+    --batch_size 5 \
+    --is_local 0 \
+    --role pserver \
+    --endpoints 127.0.0.1:6000,127.0.0.1:6001 \
+    --current_endpoint 127.0.0.1:6000 \
+    --trainers 2 \
+    > pserver0.log 2>&1 &
+
+# start pserver1
+python cluster_train.py \
+    --train_dir train_data \
+    --model_dir cluster_model \
+    --batch_size 5 \
+    --is_local 0 \
+    --role pserver \
+    --endpoints 127.0.0.1:6000,127.0.0.1:6001 \
+    --current_endpoint 127.0.0.1:6001 \
+    --trainers 2 \
+    > pserver1.log 2>&1 &
+
+# start trainer0
+#CUDA_VISIBLE_DEVICES=1 python cluster_train.py \
+python cluster_train.py \
+    --train_dir train_data \
+    --model_dir cluster_model \
+    --batch_size 5 \
+    --print_batch 10 \
+    --use_cuda 0 \
+    --is_local 0 \
+    --role trainer \
+    --endpoints 127.0.0.1:6000,127.0.0.1:6001 \
+    --trainers 2 \
+    --trainer_id 0 \
+    > trainer0.log 2>&1 &
+
+# start trainer1
+#CUDA_VISIBLE_DEVICES=2 python cluster_train.py \
+python cluster_train.py \
+    --train_dir train_data \
+    --model_dir cluster_model \
+    --batch_size 5 \
+    --print_batch 10 \
+    --use_cuda 0 \
+    --is_local 0 \
+    --role trainer \
+    --endpoints 127.0.0.1:6000,127.0.0.1:6001 \
+    --trainers 2 \
+    --trainer_id 1 \
+    > trainer1.log 2>&1 &

fluid/PaddleRec/tagspace/infer.py

 import sys
+import argparse
 import time
 import math
 import unittest
@@ -9,6 +10,25 @@ import paddle.fluid as fluid
 import paddle
 import utils
 
+def parse_args():
+    parser = argparse.ArgumentParser("gru4rec benchmark.")
+    parser.add_argument(
+        '--test_dir', type=str, default='test_data', help='test file address')
+    parser.add_argument(
+        '--vocab_tag_path', type=str, default='vocab_tag.txt', help='vocab path')
+    parser.add_argument(
+        '--start_index', type=int, default='1', help='start index')
+    parser.add_argument(
+        '--last_index', type=int, default='10', help='end index')
+    parser.add_argument(
+        '--model_dir', type=str, default='model_', help='model dir')
+    parser.add_argument(
+        '--use_cuda', type=int, default='0', help='whether use cuda')
+    parser.add_argument(
+        '--batch_size', type=int, default='5', help='batch_size')
+    args = parser.parse_args()
+    return args
+
 def infer(test_reader, vocab_tag, use_cuda, model_path):
     """ inference function """
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
@@ -21,7 +41,7 @@ def infer(test_reader, vocab_tag, use_cuda, model_path):
         step_id = 0
         true_num = 0 
         all_num = 0
-        size = len(vocab_tag)
+        size = vocab_tag
         value = []
         for data in test_reader():
             step_id += 1
@@ -47,30 +67,18 @@ def infer(test_reader, vocab_tag, use_cuda, model_path):
         t1 = time.time()
 
 if __name__ == "__main__":
-    if len(sys.argv) != 6:
-        print(
-            "Usage: %s model_dir start_epoch last_epoch(inclusive) train_file test_file"
-        )
-        exit(0)
-    train_file = ""
-    test_file = ""
-    model_dir = sys.argv[1]
-    try:
-        start_index = int(sys.argv[2])
-        last_index = int(sys.argv[3])
-        train_file = sys.argv[4]
-        test_file = sys.argv[5]
-    except:
-        print(
-            "Usage: %s model_dir start_ipoch last_epoch(inclusive) train_file test_file"
-        )
-        exit(-1)
-    vocab_text, vocab_tag, train_reader, test_reader = utils.prepare_data(
-        train_file,
-        test_file,
-        batch_size=1,
-        buffer_size=1000,
-        word_freq_threshold=0)
+    args = parse_args()
+    start_index = args.start_index
+    last_index = args.last_index
+    test_dir = args.test_dir
+    model_dir = args.model_dir
+    batch_size = args.batch_size
+    vocab_tag_path = args.vocab_tag_path
+    use_cuda = True if args.use_cuda else False
+    print("start index: ", start_index, " last_index:" ,last_index)
+    vocab_text, vocab_tag, test_reader = utils.prepare_data(
+        test_dir, "", vocab_tag_path, batch_size=1,
+        neg_size=0, buffer_size=1000, is_train=False)
 
     for epoch in xrange(start_index, last_index + 1):
         epoch_path = model_dir + "/epoch_" + str(epoch)

fluid/PaddleRec/tagspace/net.py

+import paddle.fluid as fluid
+import paddle.fluid.layers.nn as nn
+import paddle.fluid.layers.tensor as tensor
+import paddle.fluid.layers.control_flow as cf
+import paddle.fluid.layers.io as io
+
+def network(vocab_text_size, vocab_tag_size, emb_dim=10, hid_dim=1000, win_size=5, margin=0.1, neg_size=5):
+    """ network definition """
+    text = io.data(name="text", shape=[1], lod_level=1, dtype='int64')
+    pos_tag = io.data(name="pos_tag", shape=[1], lod_level=1, dtype='int64')
+    neg_tag = io.data(name="neg_tag", shape=[1], lod_level=1, dtype='int64')
+    text_emb = nn.embedding(
+            input=text, size=[vocab_text_size, emb_dim], param_attr="text_emb")
+    pos_tag_emb = nn.embedding(
+            input=pos_tag, size=[vocab_tag_size, emb_dim], param_attr="tag_emb")
+    neg_tag_emb = nn.embedding(
+            input=neg_tag, size=[vocab_tag_size, emb_dim], param_attr="tag_emb")
+
+    conv_1d = fluid.nets.sequence_conv_pool(
+            input=text_emb,
+            num_filters=hid_dim,
+            filter_size=win_size,
+            act="tanh",
+            pool_type="max",
+            param_attr="cnn")
+    text_hid = fluid.layers.fc(input=conv_1d, size=emb_dim, param_attr="text_hid")
+    cos_pos = nn.cos_sim(pos_tag_emb, text_hid)
+    mul_text_hid = fluid.layers.sequence_expand_as(x=text_hid, y=neg_tag_emb)
+    mul_cos_neg = nn.cos_sim(neg_tag_emb, mul_text_hid)
+    cos_neg_all = fluid.layers.sequence_reshape(input=mul_cos_neg, new_dim=neg_size)
+    #choose max negtive cosine
+    cos_neg = nn.reduce_max(cos_neg_all, dim=1, keep_dim=True)
+    #calculate hinge loss
+    loss_part1 = nn.elementwise_sub(
+            tensor.fill_constant_batch_size_like(
+                input=cos_pos,
+                shape=[-1, 1],
+                value=margin,
+                dtype='float32'),
+            cos_pos)
+    loss_part2 = nn.elementwise_add(loss_part1, cos_neg)
+    loss_part3 = nn.elementwise_max(
+        tensor.fill_constant_batch_size_like(
+            input=loss_part2, shape=[-1, 1], value=0.0, dtype='float32'),
+        loss_part2)
+    avg_cost = nn.mean(loss_part3)
+    less = tensor.cast(cf.less_than(cos_neg, cos_pos), dtype='float32')
+    correct = nn.reduce_sum(less)
+    return avg_cost, correct, cos_pos

fluid/PaddleRec/tagspace/test_data/small_test.csv

+3,7824 5278 529 14 4358 4179 148 46 30 8862 39 349 12 10920 10109 303 637 4155 1 
+0,472 91 2409 1039 27 63 79 4 10570 2488 9 0 297 169 104 2272 853 1625 3 2495 9 3536 3350 10941 201 1729 25 1489 172 0 43 516 1699 9 16 1476 1308 1 
+0,24 27 62 2571 20 3 4589 3582 14 0 254 4 6939 124 3 3366 2 1707 3 5095 4 2685 478 10142 84 30 807 8443 4 8067 7994 0 974 1675 4 5232 1 
+0,24 165 7 2449 2024 102 1054 6450 1322 26 1139 12 3 8297 4 4618 9813 2057 6 4812 8459 42 502 52 2087 120 0 106 29 1225 6906 29 454 5 1293 52 2051 7911 29 1494 51 9915 1293 29 3279 6 4884 29 5303 1 
+0,24 1399 829 7 10776 443 354 737 39 3293 4 0 8323 5548 95 3099 0 390 7 43 1136 2 2703 474 1500 23 8744 4680 9811 1 
+0,11 260 945 9 6345 6 7610 3941 1800 1000 3 2369 6981 581 12 0 10045 9577 4 8964 0 119 1104 4 260 9946 3064 46 436 5462 51 12 0 365 240 3954 6 1850 4467 2 200 40 557 3954 25 1662 9961 35 157 566 249 10 978 5 593 3601 29 210 34 9532 50 0 9483 4 4779 365 11 4978 6 8167 4 40 550 2 792 3 419 249 78 2804 21 520 1763 2 0 3941 50 19 3064 19 7 10134 15 78 10977 121 32 2 34 102 11433 2 9325 3580 211 78 619 121 548 19 1848 6 3506 3325 5007 2 277 562 5362 155 71 102 0 398 923 25 3371 5 185 643 1 
+0,1 4325 2931 2182 8604 2447 4 0 2983 6 3029 2242 13 2167 9 825 113 4 931 9539 5 250 187 2 3 401 162 931 2175 2149 70 20 3114 62 4310 11 530 47 266 2931 31 1148 53 3234 5 1366 5 4142 20 369 54 18 52 89 4417 5222 505 0 2983 6 3029 2249 562 511 14 130 866 27 130 3114 10553 522 554 2 0 1548 7 3117 2892 0 339 389 10260 2931 7 4564 160 64 14 214 352 11290 4 2983 6 57 4 0 2679 10631 392 2248 3029 2912 6152 677 193 4650 14 4310 18 19 31 10869 1 
+0,1 1902 6861 6 4516 9170 32 3 662 4 2081 2687 9 2195 392 6 3739 1237 78 732 600 68 222 57 661 293 1027 6 1467 6507 7217 257 9177 1237 5 0 1902 992 9 2731 932 4 11189 206 257 1237 1646 155 600 404 1902 6 4991 0 2996 257 3330 7 9 9475 542 55 1478 3 662 1068 1467 0 2996 257 3331 601 0 1902 992 4 185 2517 2062 6 564 1467 3 4516 9169 4 0 2996 257 3331 4 185 3301 3739 0 811 1902 992 85 4 868 794 2 34 9002 7949 3739 85 34 10232 23 0 2195 392 15 185 2482 985 38 9474 1 
+0,7908 1391 5573 50 3508 39 16 995 5398 612 31 1547 1 
+0,117 16 43 48 103 0 850 7 2030 862 995 927 25 5263 297 161 5415 3169 6 5645 8133 6 874 19 3598 5723 1 
+0,1185 1163 17 27 229 4 193 275 360 960 2926 58 37 4335 98 36 4278 37 4277 2993 36 6 3871 97 37 3842 236 36 8 86 18 46 29 3543 3 38 567 1283 1320 51 2 169 526 0 1146 4 0 545 1104 1 
+0,1185 1163 17 602 746 58 37 5751 236 36 8 81 611 1531 3 38 522 1087 2 2090 371 1082 905 72 2104 4926 6 1847 3 2519 372 3284 114 8372 2562 16 2059 2902 2876 1196 114 1 
+0,3791 17 1906 7421 4297 3 1898 365 725 1689 355 602 746 58 2 143 8 81 12 0 516 4 3 38 871 175 5 499 7 258 3139 9 481 2740 399 1 
+0,27 239 5624 8548 4 4449 84 25 90 985 5 7034 325 55 5477 1002 3404 1537 1 
+0,6296 1107 71 10213 5272 8 10791 2 553 203 23 8537 242 1427 1 
+0,2418 2712 4407 16 791 493 416 2 21 3300 449 3652 188 16 4419 103 1 
+0,4072 4238 4100 842 32 985 2612 9773 3146 0 4974 1830 4 2933 7 199 1281 4001 3054 2612 9824 140 34 2173 9235 5443 20 4001 7 4974 10689 9 3149 4 103 1 
+0,603 608 1347 3955 1381 2899 25 10474 2126 23 57 4 270 7 1381 5310 9 0 43 72 15 368 4 3 192 4 5018 1283 4719 11 4719 10127 0 234 9 618 1381 3461 2 2041 8 1755 5018 1283 1229 1 
+0,435 510 361 504 469 9065 3196 10892 23 0 4901 2701 5 1645 325 1658 287 455 9701 10658 30 8261 71 8466 25 10818 3 10834 1701 7238 3582 7164 7024 4723 138 4 0 1267 5 4655 1 
+0,527 6333 1016 911 627 3206 12 3 47 380 42 62 9902 967 1449 20 3 174 492 1 
+0,129 0 157 4 0 47 0 3209 239 105 2 32 16 580 9385 248 5657 1 
+0,3938 167 382 544 9 38 307 163 5368 184 256 3226 1 
+0,111 193 8113 8975 9269 2 154 3388 6 3578 9137 1 
+0,116 25 8258 16 377 15 3 234 9 9063 94 2 167 5 8 0 800 3251 641 541 0 2239 10913 19 7 89 8 6697 4240 1674 3 9891 4 118 339 400 21 520 800 12 1162 5527 99 40 10214 2 3 38 661 4 183 5165 211 40 1114 69 0 507 1 
+0,129 2357 4098 607 1043 24 577 16 363 96 6 9472 8164 2519 202 3163 1886 4283 97 7501 10501 21 4443 3659 5 641 541 7 1159 11123 10701 9 2028 1 
+0,2925 4697 142 2201 3318 2 673 138 4 0 182 3732 8582 4 0 558 3367 1 
+2,219 271 124 0 112 144 5 0 570 998 3454 6 192 3 92 159 5 3 72 4 186 1008 268 823 909 1 
+2,3973 6933 14 26 10558 1088 5540 3156 12 514 8254 6 3372 21 3490 4 1727 19 7 1238 2 1024 933 7803 6879 4459 0 356 4666 5 880 8478 5178 521 44 4879 757 7 76 2 9506 10 52 7 9670 3 4879 4628 213 52 68 3628 198 26 808 4 0 482 281 15 6253 7532 1 
+2,577 0 528 2531 631 2334 7840 4816 513 354 0 471 331 767 2 4566 5106 4090 4298 35 8426 4 0 618 5 26 4605 926 0 9788 11 331 166 182 4 120 46 32 1137 14 4355 27 7286 23 4298 1027 19 2161 1461 5523 287 52 68 200 0 8014 14 0 1497 1 
+2,1606 4982 120 944 1620 55 277 8 3 447 110 7 1020 1620 2743 0 110 132 75 12 26 2448 1676 1361 5839 54 13 2449 73 162 266 573 89 1124 8 4474 0 471 331 2308 1041 14 3983 1194 109 1009 42 52 89 2 3636 5 353 99 1620 89 3 2739 14 168 6 52 1 
+2,117 87 1744 1328 87 7457 937 6117 6546 219 6775 6 944 7284 510 1285 95 110 2 200 1038 5943 7 649 8 120 2 277 12 26 10409 245 4789 1 
+2,11 1358 952 802 601 57 137 4 0 1035 940 231 20 972 0 835 1403 289 73 136 110 12 168 675 20 4353 4024 6 3069 1615 1 
+1,327 378 7833 1172 11 2733 4 3 343 54 770 39 3 1031 8614 12 369 25 4715 0 179 7 612 8634 3583 9 870 0 1781 3410 1 
+1,24 11 343 54 582 1909 1375 1362 4144 10678 9371 180 31 3553 9 3 437 1528 4 51 2 297 682 104 132 0 1737 2029 2 694 10 52 31 1735 6 89 21 9109 4423 1316 813 11 1036 378 1 
+1,836 100 6391 6 3369 10391 164 0 179 4 253 15 41 45 196 5 4461 2787 6 1546 1135 122 53 0 917 292 33 3 106 39 0 1231 4 349 6 12 203 3 3689 1686 616 2607 2 157 0 443 5 40 484 179 1 
+1,1919 684 27 4903 1062 149 414 686 4540 4953 6 9278 14 526 4575 21 520 474 1188 8 33 9 678 2597 2768 3513 232 5 5588 14 3 628 2 0 1207 2965 10 13 2443 12 5101 2022 497 57 1125 52 18 154 792 5 7334 310 3670 164 26 9428 5 409 1 
+1,7693 6 726 3736 4 719 995 107 123 251 7 2506 55 1127 11202 5 92 67 143 1 
+1,24 27 5659 893 5422 891 10321 9 212 1367 6 1394 196 358 2 4412 8948 1369 30 216 10330 4 0 7823 1397 6363 6 7427 7458 1872 7816 1 
+1,3050 17 27 677 1914 388 317 18 33 538 6 0 88 128 85 76 8749 475 1183 248 19 55 454 122 8099 5 538 12 16 258 9903 306 3103 147 1 
+1,1377 375 1910 196 9589 5 375 7 258 10049 4 2026 443 197 11333 3340 23 0 359 2 0 657 5586 8 41 45 6 4162 8044 2 10536 21 474 1419 8652 5 0 598 134 18 33 1 
+2,24 1939 747 209 3 133 2594 53 0 1013 340 2 154 0 1353 1895 157 3 1012 137 1274 2205 33 972 1201 4350 6 0 111 332 1382 66 226 1 
+3,17 5846 2709 30 1463 35 122 2 676 9133 29 200 0 9328 335 3672 8620 5466 6 738 4420 40 1245 42 19 55 34 3 2230 425 12 2208 6 1290 1472 3 11116 783 8 35 2793 1 
+1,24 464 299 328 89 918 77 103 380 101 152 7 1555 164 0 2250 4 40 8529 179 85 32 90 118 0 3647 4 1294 1091 816 10 135 10279 32 181 2 0 10635 4 3661 4788 9207 6 9182 1 
+0,6401 91 1630 5 7627 91 3 295 56 2630 634 2609 5192 442 49 61 169 1016 3175 8 95 39 171 7 580 344 190 392 1288 9508 3 57 47 5451 8 16 2109 2641 399 213 4 300 256 1025 1536 1 
+2,24 6235 6375 4776 5 0 222 274 145 5 0 1240 340 5058 0 2828 3817 2 3 170 66 224 44 0 1594 3056 8 33 11 108 802 2828 601 3 5173 1125 4 828 6 960 5 0 1035 1561 862 282 1 
+2,24 7235 1047 1361 6308 713 0 826 1585 137 83 0 1064 1379 1587 8 33 110 12 3 3646 674 10354 1 
+1,533 272 17 962 1157 5 514 1121 8 33 5 3 659 319 8 287 2 424 215 1159 101 261 251 6 1445 476 10308 5567 566 53 0 110 1 
+0,2319 2340 17 830 58 37 1893 236 36 0 92 7 199 470 342 18 8 28 19 25 215 0 363 157 381 470 143 5 171 6 265 16 1784 449 286 9 0 127 40 47 416 2 5413 438 5 0 5331 1 
+1,2839 17 1355 8 28 543 3 634 126 1978 9 10858 9 1401 4 974 21 9049 392 5 0 88 128 6 18 52 55 222 8 814 930 1144 1 
+2,3818 216 255 4746 9 2414 3081 0 262 73 1514 670 5 0 92 54 25 76 124 3 216 2004 248 26 11180 14 0 2811 41 45 330 1 
+1,125 27 10860 4 56 1874 2905 2306 5642 1454 206 5 2269 29 3339 2 227 689 119 162 9 3689 8811 2435 14 4814 56 9181 8 0 1051 10139 56 476 18 1 
+2,80 93 17 1939 747 918 60 740 3138 5 268 73 170 889 2 154 0 1353 1895 157 3 1012 137 1274 2205 12 3 66 226 108 44 0 1261 111 332 1382 5 0 212 467 33 1 
+1,125 27 8731 5 0 359 4 0 4137 31 772 42 2120 10671 64 396 82 4994 6 1716 561 867 879 75 65 7 10561 134 6 1133 18 1 
+0,4344 91 11 41 2339 344 239 1340 2 548 3 234 2 8873 0 2877 567 6 2566 518 1227 1 
+3,412 17 11 702 2516 2 3 133 65 363 83 0 1718 8 28 132 2713 3637 8 207 348 42 5410 67 96 6 5577 41 45 301 867 2 4859 458 50 0 782 4 0 92 7 199 221 1 
+0,429 219 6809 892 320 53 94 54 64 776 5145 6 780 49 11311 799 52 10399 52 55 277 562 50 19 1 
+1,125 605 7 4339 6749 6 7664 941 690 209 3 2204 1147 2 3310 16 936 9220 12 0 2439 19 29 479 0 496 1193 10890 1287 4 1398 7 7137 690 1 
+1,744 100 3934 14 100 7 212 1889 358 8 1300 540 480 1385 334 432 2 3087 26 2238 83 41 45 6 269 196 6 1795 26 1454 60 4 3 484 628 5 253 1 
+3,80 93 17 41 45 4354 3242 96 10758 8 28 1027 3697 8452 0 592 15 207 348 3563 61 245 3349 1079 5 3615 1 
+3,80 93 17 11 702 2055 2573 83 0 1718 8 28 39 3 218 8 3338 53 41 45 2442 1140 614 4 3 627 5 881 3409 2 2112 0 545 3089 774 9 0 162 1 
+2,4128 693 468 41 45 753 535 1075 4025 4324 4876 26 79 8 28 102 52 172 0 10662 4 1376 1600 6 1202 1357 15 26 1075 7 2125 1 
+2,1376 1600 3412 1202 1357 15 0 48 1075 7 2125 9 3 41 45 79 10 29 495 2 2158 0 3231 23 499 119 162 1 
+2,24 126 7138 3915 124 59 112 5 0 311 364 4487 28 110 3915 0 241 1563 417 2 1832 243 73 1482 5 0 653 1101 43 5 73 1482 226 2254 909 6836 6195 4 4399 54 11010 14 5877 254 5 3809 1714 0 1142 5 73 1566 682 6877 6977 4 684 209 1995 5 73 1566 2819 1 
+0,1057 1388 1055 7198 25 5297 16 191 808 1881 969 5 565 3753 169 4 16 191 1768 2668 1520 663 6 440 4621 968 1 
+1,125 11 88 749 1000 21 1216 292 9 5669 104 987 2 154 4890 1851 2555 4927 1102 3625 39 48 5476 4 1584 31 10936 3254 4691 6 515 49 61 2817 94 1 
+0,1910 5 227 1182 150 4 6828 3617 51 567 9656 15 368 4 3402 2 8342 481 8899 1 
+1,1060 1045 831 2421 908 1149 786 2430 6 212 1367 196 1810 290 312 28 2 154 1851 4 299 328 6 1704 599 6 135 1325 2 350 4 94 215 2595 1 
+1,3022 2953 745 3976 4 219 1916 3304 5 3184 28 15 0 2678 285 2538 23 3 546 3244 1555 2471 6 737 53 228 9 3 1814 12 0 3545 54 25 3552 180 9 103 8 2479 2647 647 1 
+2,24 11 6115 5976 2771 976 7352 6309 2 0 826 6636 8 28 9 3 63 346 1494 5 0 741 2370 2040 1 
+1,327 378 6851 326 938 1172 1391 32 1978 3 343 5 0 3415 6 1313 761 4 21 351 47 266 4920 10 4909 40 10951 4369 2486 75 65 1 
+0,471 3989 844 408 0 62 291 171 7 295 500 2482 1237 105 2 878 49 8 16 910 307 163 6 1243 300 0 62 7 1969 267 18 1 
+0,1576 4237 7599 1908 462 7760 1396 6711 1908 91 58 172 195 10 0 6166 2982 5763 6 6689 2687 4 324 1059 58 32 1141 21 759 12 1908 2 309 1908 7 5480 208 1466 15 21 2606 368 4 35 208 345 7829 1908 7 5480 208 175 208 371 68 1 
+0,24 270 7 2283 915 25 10830 48 38 419 3467 263 1022 2171 1629 0 201 354 18 28 1 
+1,428 3400 179 9990 4 1616 3669 15 5762 1462 3 3162 2133 1 
+0,24 27 2987 9695 5 5089 18 28 0 152 610 7 3430 8 2064 9 4803 1824 1077 308 3287 32 2748 0 8513 8 126 242 1095 2 1707 2775 9 3 1261 4 8483 8761 3262 1 
+2,24 3012 2211 255 1138 28 5 0 1050 937 1313 761 507 1274 3 75 4763 857 2 553 0 2370 285 7 1761 23 3688 2617 50 0 1657 5557 7 1529 361 1 
+1,17 171 25 18 141 1699 25 90 192 9 934 661 349 8 0 413 1051 1488 869 6 775 141 3399 10 0 1276 25 90 3166 431 946 244 1349 1897 18 8 81 1 
+2,11 4255 1099 514 1329 23 1866 5200 6188 6590 54 31 1141 2 1212 0 79 7 11353 927 0 3248 518 1 
+1,125 11 75 10994 126 8777 2 3205 413 689 1340 2 1549 26 1542 2 396 3 785 279 8 0 4315 150 84 52 1987 13 11233 11306 7973 260 2058 1275 18 1 
+1,2390 6420 6 4316 7275 30 549 9 112 9 2318 407 8 106 133 5 151 1 
+1,17 636 8445 149 329 25 3 4620 231 44 101 152 5 829 4 2816 113 2 2255 113 335 362 817 3 2130 296 8 81 280 1 
+2,74 17 497 0 1671 2789 0 250 177 13 3 425 60 2060 10906 1828 1 
+3,412 17 11 702 304 5408 391 40 65 7 133 65 363 83 0 1718 8 81 12 485 8188 3 6545 341 766 1026 6 41 45 381 2077 1121 2 2188 1118 8 0 2518 1 
+3,607 1043 117 0 258 4 3 728 4 671 1703 953 97 25 3525 0 516 4 3 372 986 765 19 89 2140 1124 2 5020 20 47 157 1796 82 3 1814 12 960 2926 58 5 0 1726 448 143 9 173 4711 921 4757 1 
+1,533 962 1157 10517 2 553 9179 3524 261 251 15 35 267 5 3 224 10 3274 9981 5254 109 23 505 1816 5 0 917 465 1779 8442 0 345 15 8501 6 3091 1 
+1,125 951 956 8797 543 171 4 9621 57 4 35 1119 8 11188 51 5230 647 5 3 645 2 8874 3 718 1483 1355 25 90 9155 12 9 321 103 1 
+3,24 1630 4 955 1180 97 2757 632 113 28 39 0 2519 523 147 3 207 9 0 63 158 6 18 1078 6 2634 1785 3959 6872 13 210 513 2 1753 0 62 7 297 66 115 153 5 10985 437 1 
+1,24 1597 2478 336 1878 2309 7 2439 11094 13 3296 2 3 466 1563 55 108 180 11021 335 466 134 6 154 180 1220 2508 2 0 88 128 26 9159 6 57 4 26 2756 18 81 1 
+1,74 205 2416 126 1318 2334 6898 6 687 2873 713 0 265 5 0 59 311 364 445 1502 81 3 2754 2458 9 3 127 10 89 1416 1711 566 5 0 653 1622 1338 1706 1603 4 0 1927 6 431 687 848 1412 2 0 1447 3612 3 106 39 8942 1548 219 271 5 0 460 445 124 20 848 1 
+0,27 2146 1418 20 3 3197 4 3686 1682 85 265 3149 4 987 5 4868 10085 20 174 2520 275 2 2704 190 4918 42 85 1413 0 2704 275 2 200 51 138 4 0 3255 20 1301 3466 190 3150 9 667 4 981 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 11 433 1623 37 326 36 37 572 36 1 
+1,1377 17 11 88 128 25 8357 3 1686 616 197 3 10490 935 1007 294 6 0 3292 1737 4 0 681 1511 4 6649 433 7 4820 2 2341 18 81 1 
+3,4089 21 1721 8 185 1698 862 9615 46 140 34 50 2 2152 185 1130 1 
+0,117 255 671 10249 2724 239 953 97 18 19 124 69 34 394 3 765 9 3540 2410 20 0 157 4 250 15 19 89 172 991 40 47 1 
+3,80 93 1586 2974 2974 1767 30 448 49 10156 50 0 221 2014 2686 6 56 202 143 5296 187 2 3 3466 1026 20 4147 4122 296 81 1 
+2,59 336 2940 6816 195 215 2071 274 4 26 1699 12 2921 9560 3190 26 5003 6 11323 87 353 0 8721 698 0 10534 87 1 
+2,4167 536 24 536 613 4325 6573 6372 18 81 52 1114 22 69 243 245 876 1126 5 0 4426 4 3 1522 10 8745 0 3992 1037 22 7 5280 1 
+2,4167 536 24 7677 6359 6653 25 802 60 4 0 536 5398 274 4 70 22 7 4908 83 4364 213 4 3 1553 9362 0 3992 1037 18 81 1 
+3,80 93 17 7657 58 37 27 376 305 373 341 397 91 410 415 402 4275 236 286 223 423 418 36 4275 236 37 27 36 0 176 41 45 489 395 523 8 81 147 3 22 113 1269 5 559 207 1018 16 585 47 1252 6 18 19 105 2 1242 0 1726 448 634 143 1526 16 400 293 1 
+1,5760 100 204 632 27 2507 4 6746 31 3249 9 142 1517 195 42 184 2607 2 816 253 2 495 2 8668 3 3565 540 480 6 26 1007 2 9057 3 628 5 0 484 179 6 157 1 
+3,273 17 41 45 381 96 1444 5 833 9 0 43 72 5 441 389 15 3 3622 145 51 5 704 1433 10545 0 107 18 5 3 218 10 3679 3 2737 904 4 999 904 9415 13 362 1 
+2,428 1182 1977 216 492 26 234 2 26 127 22 7 43 417 998 59 1142 144 8 81 310 5 0 398 653 21 1772 4 6311 22 7 1807 1724 1073 0 43 417 5088 23 0 88 743 2895 1 
+3,4259 1103 96 3 8366 9 182 4 0 725 2180 30 1744 950 1593 1368 1619 306 308 0 10488 239 10 147 3 2212 5 63 158 383 81 6 4938 0 1020 4 0 47 293 1 
+0,2839 204 632 2426 171 13 1683 60 3 9953 581 2 8697 198 8 87 190 1529 87 399 10105 255 2219 2113 83 156 10237 244 4 4048 2925 7882 7937 18 292 81 1 
+3,6116 7783 1122 2432 8 11404 0 11317 4 0 8264 2180 1 
+1,1606 120 1349 1897 31 5474 102 52 1436 8 1798 75 95 10 87 46 55 616 3 1278 23 413 689 2 1400 87 13 3709 11 10331 10485 118 276 132 26 4463 109 8 26 63 816 2 0 413 1051 1 
+2,74 17 379 311 666 336 4349 2301 29 1774 0 151 177 39 189 1216 83 3 48 47 990 1669 31 3264 8 81 3 811 813 17 1 
+1,80 93 3041 560 9 3 63 1324 1527 81 15 3 703 5 381 96 1414 485 2 396 3119 458 50 2077 14 214 9 0 807 1151 577 3353 96 9125 2 441 162 3438 0 3910 3011 2923 3569 3 809 703 5 833 884 981 3 10520 23 1535 704 96 1 
+2,6708 6707 4 205 5478 0 168 1083 81 972 1942 6569 4 4364 2 108 0 112 144 5 0 2818 9652 1428 1 
+1,125 315 29 494 0 231 5 10528 715 2 16 3565 298 233 42 1011 0 563 4 633 5170 6 3385 1212 946 244 7105 7229 6751 18 1 
+0,11 1221 44 168 2048 1283 13 4952 51 15 3 3197 5228 1032 255 1320 9 0 6691 22 7 4644 1 
+3,434 239 1965 13 9275 10 370 1084 8 16 3266 668 2705 6 434 9446 395 29 200 19 49 8575 12 8867 5 0 201 84 2099 10 809 1227 590 2 592 367 167 3 5454 338 164 1 
+0,574 18 81 19 25 3093 3 2119 4 6234 604 399 1250 15 368 4 16 773 2 4548 16 2138 5 7545 429 3 1134 4 0 3469 574 29 1654 50 170 2817 604 5400 2 16 483 781 2500 1364 1550 4 1 
+0,7093 105 2 425 344 190 392 873 10 1337 500 206 544 3 1147 2570 5 392 2576 1275 22 1095 2 1302 263 3 1320 1 
+0,130 29 2033 0 247 4 16 959 527 9 255 65 2 3335 114 4922 211 76 1244 13 2695 249 4785 2 3405 0 959 527 9 178 609 117 1723 232 275 32 4715 0 1650 2 2139 35 1 
+0,17 1961 38 3467 64 1822 263 1629 20 0 2283 201 1297 1301 0 2228 2 1867 3467 9 0 5301 899 270 18 8 28 1 
+2,639 30 10486 5363 197 0 250 1932 846 6015 6 426 10 124 0 420 2004 118 48 103 380 1 
+0,11 38 574 7345 3281 9467 5717 436 2056 293 898 6 38 5562 1229 10 1413 19 2 145 1777 307 525 14 1014 8 3616 10116 1 
+0,7161 569 7 4316 852 723 8 1524 3036 569 7 7123 6603 3126 670 5 3 1541 1865 1033 1604 569 7 19 8569 2 0 1465 569 78 760 3 10611 1033 2845 361 13 3 662 1003 51 2 742 552 569 52 18 1033 11 6854 10396 13 1757 4 8851 569 78 18 569 2845 361 13 3 662 1003 569 52 18 1033 78 1156 19 68 569 69 521 4131 4874 569 78 18 1033 2845 361 13 3 662 1003 569 52 18 2251 204 632 
+3,7058 97 473 2 3550 3 682 113 999 5 1012 9796 2665 20 537 1924 5865 1920 97 9 118 44 73 153 1233 3 38 1755 2240 197 0 1107 11 338 29 2553 1 
+2,74 11 1775 64 249 9893 1883 2380 22 7 2060 10644 0 895 4 0 11232 3132 3916 4096 22 11193 1492 10 1097 292 3406 4 160 7074 5961 22 7 5371 10932 8 0 3141 1 
+1,11 370 4 67 25 867 16 3622 627 1491 1692 14 3 159 173 11 790 768 5 111 332 1118 2762 509 9 1705 119 162 25 2007 14 3 159 7814 1573 3 1072 51 825 2002 8 109 22 7 337 1 
+2,87 165 1749 166 4959 78 55 1024 848 1080 51 211 102 78 31 4540 78 1311 366 79 549 8704 6 10 722 892 792 549 87 1 
+0,165 1158 22 69 0 43 2 222 809 6 19 124 22 69 34 0 580 1506 42 1575 22 7 1063 2287 2260 1074 1346 2110 9978 697 30 302 2 200 1858 5 0 2724 2132 1 
+1,412 2824 632 407 977 441 1551 2217 8 81 12 4646 2 4623 2651 6 18 57 89 105 10 55 34 404 5 1827 56 8368 10 64 0 878 4 142 10633 40 162 1 
+0,1057 574 1619 574 25 333 1819 2 32 0 10138 894 582 20 11 4271 832 3264 1577 2 3 2104 19 1247 5 3 41 45 4761 228 574 3116 10 4271 25 141 654 2 563 16 582 10 19 8119 4635 811 544 23 2420 3049 7831 6 2127 19 5 500 1 
+3,80 93 11 1778 296 3698 2503 9134 301 32 562 9 1244 5 0 1701 50 0 1521 4 1821 5544 2787 1 
+2,27 6037 2598 22 7 3436 9 26 2246 25 554 2 3 5483 4 142 14 99 59 3722 5 151 1 
+2,837 10 9062 5 6 263 1404 2844 25 4539 3 514 5358 4 2160 2090 22 7 76 167 4567 503 1 
+2,1 11 88 128 146 22 7 316 79 8409 82 3 357 106 9 0 1404 20 443 82 205 9 3 11314 108 5731 5730 87 847 901 12 1256 87 18 3897 4103 1879 87 757 22 7 99 121 68 4468 87 1 
+1,744 100 27 420 718 508 22 7 645 2 157 0 443 5 0 540 1925 484 179 4 253 1071 2 32 656 81 1 
+1,4331 1046 1365 101 7055 7515 4452 2 92 1271 81 2 8662 21 587 508 8 0 1430 5 3159 227 1192 225 3238 9078 4 3369 5477 2 10814 1 
+1,125 1365 6 227 2378 444 1491 543 1094 135 4 734 2 731 0 135 808 7 2134 5 1046 7 3159 233 4 227 1192 15 133 1365 810 64 147 2 34 1861 1 
+2,2842 165 31 255 8382 106 8 0 1770 1458 9 0 471 331 15 3 728 4 10583 3469 64 172 10292 2 81 110 22 7 137 83 0 827 1049 1 
+1,24 149 329 2853 6830 165 7 1238 2 167 4639 1 
+2,24 4385 4018 6622 0 3585 254 4 312 1544 613 54 554 0 6537 2 441 508 2770 770 81 676 134 18 573 31 3783 1 
+2,4333 204 632 1631 219 271 13 76 549 2 664 0 321 112 665 124 20 957 1953 532 19 13 597 382 2 1549 155 52 29 664 5809 6289 0 2403 9352 54 124 441 2228 665 5 2808 211 426 64 76 0 1 
+1,1957 47 266 7852 23 1064 7307 25 90 1272 14 0 7551 22 1724 168 5 6619 5 1487 1366 9 678 528 1 
+2,24 4197 3883 246 3 145 2179 546 82 0 4009 7081 5 0 1013 340 8 10480 834 7640 7 63 3516 4 0 137 884 0 828 471 331 3 139 186 108 44 0 849 827 1049 8 81 110 1 
+2,7585 3806 1989 6629 1348 6973 1714 0 43 417 59 144 9 0 88 743 2895 102 52 209 168 0 112 144 5 146 22 7 546 5505 2190 8 81 5 151 1 
+1,538 172 105 9 73 161 3384 5 0 642 683 109 4405 0 1719 4 0 3620 1 
+2,24 497 214 57 1772 4 0 176 1514 1399 829 880 79 13 243 890 9 1313 761 0 1052 1167 1391 945 18 81 1 
+1,101 152 5 4211 18 10 3484 4 3 1278 650 163 64 1796 0 390 7 142 14 1306 1 
+2,2839 17 171 4539 3 5087 5358 4 2160 39 0 2735 4 16 316 239 4394 4149 8949 111 1649 2 1831 26 79 9957 2 35 3360 4 2699 0 151 177 2723 1455 1 
+1,27 9706 411 318 2 1226 73 161 38 2336 10673 887 5 0 642 683 109 1025 152 610 134 10643 9 3 2166 5 0 5334 1291 132 1928 22 7 594 237 1 
+1,412 260 369 977 441 1837 2217 109 12 4647 2 4623 2651 6 309 5247 9830 5496 2574 8467 71 3312 2 646 87 1451 71 1471 87 1 
+1,1374 229 22 7 1417 1159 0 6754 8971 6249 334 7379 6003 6292 18 52 2211 3423 11419 1 
+2,6323 204 632 3069 3085 506 2 0 1382 12 1300 1483 8 26 3518 9188 3336 5 26 1705 6 3 3232 3646 1538 9739 263 26 674 4795 1 
+3,638 17 11 702 2106 5 1553 10384 8 70 15 182 485 3627 82 895 61 302 41 45 2077 301 6 3671 2 0 3725 0 41 45 637 1394 85 767 1301 1130 1 
+2,3019 9365 2968 7089 1552 321 5379 889 42 0 1587 2654 3 2451 2632 2792 2 63 3134 1641 7918 6 3 9330 10748 23 38 674 4871 4103 4377 2 257 3 401 145 1240 340 9 3 1 
+2,74 17 685 1318 1890 1608 6 1607 1639 32 970 14 21 151 1747 9 21 465 59 1359 2921 885 53 35 713 8912 1153 3 5317 10 25 10682 6 8076 0 59 1261 1 
+2,2842 11 849 827 1049 32 89 2803 3378 446 40 283 83 5052 10182 61 3004 1615 1 
+1,7438 3192 5 227 1192 84 5291 5 769 4 48 1365 810 3304 589 8 430 632 678 552 39 0 227 2378 6 1365 134 473 8 1687 429 3 1134 7731 25 502 172 10 19 29 76 1 
+2,414 948 1545 255 492 14 551 3695 102 52 1273 495 2 477 0 43 126 2 108 0 59 146 7 99 263 5 1740 1 
+1,17 375 7 101 4027 4083 1509 2826 4928 142 102 46 222 2 456 5 0 127 7 43 417 4718 237 188 21 3397 106 5388 8 70 1 
+1,607 1915 745 117 3 2191 10 116 58 7 586 183 394 1114 69 15 557 15 302 0 62 9768 16 3306 370 1303 2 197 5736 6 3787 1016 911 198 23 0 991 10264 4 5629 6 3756 1016 911 1 
+1,4319 14 0 7094 7601 7550 4 7030 1117 10 4738 6 2504 10683 782 5 1899 6623 1 
+3,412 17 750 96 2757 2 3 38 173 4 1343 3 1072 8 70 39 3 38 1028 20 906 1007 83 269 67 9112 6 15 0 88 128 18 2077 89 2747 5 2477 583 2713 704 1433 1 
+2,74 17 205 22 7 48 176 1219 32 802 60 4 0 151 177 6 8103 2 0 685 94 9 3 1522 44 713 3268 1153 10 25 11048 0 252 22 626 2 35 4511 1 
+2,74 17 27 11356 219 271 1547 26 384 59 112 144 5 151 2778 26 1667 8 70 2 0 460 666 998 3454 6 10674 9 0 63 1246 1784 72 5 0 2591 1 
+1,3996 17 27 677 6613 294 920 21 411 4469 857 5 0 1599 2406 70 42 14 214 339 135 1621 64 709 5 0 2551 10 5491 164 26 168 1 
+2,607 1043 577 0 43 8913 14 4269 1194 192 2 222 82 195 502 5416 2 34 3 174 11411 9 0 1907 165 29 979 8 0 4956 4 21 278 1003 110 1 
+3,1643 91 11 1945 6 2305 1174 13 302 2 456 70 2 3539 1282 2063 275 23 9262 202 5500 2 8355 1250 10 2434 2 5226 35 1254 2 485 1 
+0,4254 25 3414 82 120 19 582 13 0 580 367 365 1310 5 551 462 3 1765 72 1448 2740 5 0 303 22 7 7419 2369 3042 68 34 3270 9 2813 2002 12 182 8047 762 9 186 2265 1 
+1,7256 1195 17 4216 3205 907 4905 70 48 810 46 89 304 15 721 4 259 9 49 61 339 389 595 46 1339 2 3566 766 5 717 349 12 0 107 1 
+1,412 17 260 1551 2217 977 5 3 10200 1474 2 142 4432 14 994 3684 5 111 332 111 1605 6 433 181 35 43 228 4453 70 1265 3 173 142 1504 1 
+3,2319 2340 1042 539 171 2363 951 956 690 6046 0 2630 927 4 171 22 7 580 2001 190 1288 720 70 3 289 268 627 5 43 779 487 207 8 3 821 708 5 16 10943 1419 1 
+3,80 93 17 7080 408 37 27 376 305 373 341 397 91 410 415 402 4130 98 286 223 423 418 36 4130 98 37 27 36 8 70 18 0 41 45 2934 945 25 2007 21 5004 53 623 1975 1267 10455 3 257 9551 404 5 16 7483 9403 1 
+3,3917 1888 58 22 7 1422 4 3258 25 1732 3 730 3204 4 1469 3258 2 2493 3917 3964 58 22 7 2146 2 494 0 62 798 5 3 268 153 202 8385 1 
+2,935 288 181 3 807 3433 1085 5 0 59 252 8 70 15 530 47 266 2616 5593 6487 7415 31 1702 39 3773 909 4 189 43 664 5 0 243 5729 9875 1 
+1,1636 587 773 2 157 0 413 1051 1488 869 1663 14 1306 4 11245 583 946 244 1349 1897 22 7 173 5219 591 2 0 10653 4315 150 1 
+0,639 30 138 778 2468 60 160 211 46 99 1811 2 32 138 5371 9974 1439 9 1094 10 200 19 658 9 792 2 5034 122 6 9013 203 963 160 30 138 1508 60 160 2 200 2468 49 1848 9 138 10820 10348 71 20 8309 8 4945 9105 5605 69 19 34 778 155 455 94 55 979 51 12 3 10875 192 4 1486 2056 166 455 27 662 4 2468 140 563 57 71 48 42 78 1273 8251 9985 32 203 99 1 
+0,4292 835 369 8807 30 482 1152 3 4942 1708 10 3599 3 2215 7 3330 6 8849 3183 19 83 835 7 3229 551 6 3330 2023 1 
+3,24 2885 523 4267 2405 58 70 18 16 207 442 50 531 113 5 0 258 158 416 2 799 12 3 38 220 163 10 1765 0 62 7 1650 2 5286 2 1230 5 1234 984 1 
+0,2842 11 2085 13 184 60 8 287 3 220 68 417 3707 34 5012 42 160 13 141 1129 10 19 68 32 1777 5176 165 22 7 118 3 1769 4 114 851 2784 732 4 0 1794 765 15 0 3158 4 3 220 11 1 
+1,6802 1190 11 7148 359 31 5041 265 82 23 0 1020 4 0 127 8 70 39 2963 907 1154 2 475 245 2787 3701 8 790 2172 5 3 1853 3155 4 6813 2 490 35 4717 9 0 247 4 1 
+1,327 378 1930 1172 2365 4 0 217 48 104 9490 54 970 5 1039 188 0 5656 3619 5 57 4 0 127 7 2253 8436 9868 1217 42 184 280 19 4947 2 548 235 61 5123 2461 2480 7662 1039 147 70 1 
+0,1203 2972 58 25 2029 2 10248 16 1948 7688 250 575 1044 5543 508 6 13 8647 3455 0 653 12 16 4068 741 4733 508 2177 9 0 157 4 688 741 1 
+2,74 1898 3236 778 2946 3086 1338 2977 8993 3247 189 59 72 814 112 144 292 1 
+0,1388 2610 3 38 3839 2675 109 3875 4054 5630 10 3593 51 53 57 5381 99 4 35 3839 1554 4040 49 2081 61 0 9977 9340 2513 4 120 13 4991 13 0 122 1542 348 422 14 0 398 72 1 
+2,3791 6211 2946 3086 3067 2977 124 0 43 112 144 9 0 1927 14 0 151 59 252 8 70 1 
+3,80 93 17 6845 408 0 176 41 45 4932 8 81 720 3 1570 113 627 5 559 207 416 2 439 3216 42 370 1084 2 3439 5347 1685 383 2 1774 641 541 874 6 400 442 1 
+2,2961 467 7621 7 48 1719 1149 1197 7527 2350 6 2867 6264 6131 1404 29 1116 0 1645 1889 9 0 741 283 1 
+2,24 11 6719 7264 4411 976 4069 6471 23 0 2604 1001 81 1 
+0,6674 8 0 4956 4 16 169 47 2204 3499 12 0 1378 5982 4 7156 130 13 337 2 3630 3 38 4547 77 47 183 2180 759 12 0 107 1 
+3,571 17 41 45 4431 32 473 2 3429 1730 53 461 7 236 4031 465 1875 2 2819 8134 1016 784 197 289 3 486 6 268 451 486 5 21 773 2 265 4640 10 25 1534 0 5589 41 45 4482 163 512 134 18 8 70 1 
+1,2955 204 530 2426 314 101 2423 2385 18 70 10 752 13 1131 2 4658 2 3 1528 4 1430 197 1046 6 16 2184 1511 4 227 1192 1 
+0,7607 389 132 270 105 2 626 0 1532 2 201 134 732 46 22 928 4824 3644 0 1298 10 3267 1590 5 1161 8762 1080 82 16 1103 
+1,1391 5 2857 22 7 359 6025 404 3685 883 2 859 51 3 8802 70 304 2 1017 0 5083 4 3909 4359 2157 1 
+2,11 2320 4 6412 296 0 258 7847 1359 8206 109 429 121 387 71 387 76 2162 160 22 7 3 217 38 3204 5 1029 1 
+0,6189 2456 7563 7 2755 44 5168 9 1155 658 3276 404 5 5191 4911 1 
+0,571 949 958 463 25 2106 16 5786 6193 7026 392 142 114 53 4507 1153 12 3 10661 229 4 603 6 413 126 300 5 1746 4 10397 0 671 9 3 741 247 21 463 875 18 14 0 463 379 508 292 5 461 70 1 
+1,327 378 1930 1172 414 2964 172 3 216 1490 4 26 677 1539 70 12 678 337 8605 6 57 9068 8388 1120 2067 216 10577 5 0 501 244 7 1931 5 3 1294 237 10680 1 
+1,125 636 1067 1912 9437 3031 149 329 680 10 101 947 2424 152 7 318 2 3744 825 161 196 23 499 6 496 85 4963 0 259 8 1836 6 8988 413 689 1 
+1,4296 480 1383 432 2164 8036 2 157 21 2238 5 0 484 269 179 4 253 1 
+3,571 17 4145 58 37 27 376 305 373 341 397 91 410 415 402 2356 98 286 223 423 418 36 2356 98 37 27 36 8 70 18 16 559 383 560 1214 4546 984 9 1090 10 1766 2612 1256 2452 6 395 404 2 2231 0 5389 1 
+3,637 134 195 172 105 2 11078 265 2254 1730 307 14 461 7 236 4031 465 1875 2 154 2703 0 2033 799 10 10563 290 0 127 1 
+1,80 93 1373 3627 82 2713 509 3351 70 2 4564 566 2141 400 1526 0 1055 8583 2076 51 73 115 113 274 4 116 58 7 281 1215 586 183 394 4 202 117 1211 1332 0 1042 539 3400 1981 2065 3777 169 71 226 289 113 2 169 5621 3786 1 
+2,2946 3086 1338 2977 4 0 1927 11394 3 3685 39 457 0 112 144 5 0 288 22 7 1307 3236 998 72 814 14 0 7864 59 6098 5 151 8 70 1 
+3,571 17 4145 58 37 27 376 305 373 341 397 91 410 415 402 2356 98 286 223 423 418 36 2356 98 37 27 36 8 70 18 16 559 383 560 8 4546 984 9 1090 10 1766 2612 1256 2452 6 395 404 2 2231 0 5389 1 
+2,74 17 205 22 7 176 48 1318 32 10368 0 59 252 39 10940 35 127 2 401 276 4 4801 5 3 4960 6 5330 2495 12 579 1442 9009 1 
+0,7604 5112 5201 64 357 4854 5 676 898 1601 126 5465 54 2631 12 902 1143 5112 30 49 362 2 167 53 2232 14 676 6 32 902 5081 6 30 517 362 2 732 46 1273 222 2 2484 123 3 7455 254 596 6667 6641 21 3121 3538 4 10792 8070 301 23 350 4 5466 54 209 368 5 0 212 6930 7681 4 5790 2916 1 
+0,2824 530 250 6690 7114 7504 27 2634 4 56 168 156 371 210 32 973 187 2 3 1026 20 4176 58 1 
+0,604 7257 2 34 3 10634 3748 4 3 727 9161 2198 27 201 1297 1683 10290 3449 10639 23 3491 201 1322 2 10206 5495 593 899 211 132 19 68 454 3 4958 9211 20 3 1370 10935 1 
+1,744 17 41 45 444 709 49 61 682 1200 1268 1481 8 70 5 3 3629 3101 53 3 1170 10950 10 13 3 10253 9 1300 480 1383 334 432 0 388 18 1 
+2,24 7794 2 167 0 356 719 8207 338 9 0 2976 2898 216 3420 3133 9572 16 1761 2 490 274 12 2111 2082 133 4 0 1217 1674 9 0 79 1 
+2,74 17 27 5304 1083 1258 5006 205 7 176 288 7 1671 2789 79 2 11184 171 8 70 6 673 0 119 346 1 
+2,6530 7633 25 680 3851 4228 8778 2 34 11340 4 7053 122 2 1253 10924 957 2422 6 4075 3990 4033 1 
+0,17 4059 58 6695 236 342 4 0 262 73 41 45 2220 10276 114 7800 8 70 720 3 3742 559 386 39 348 3 3364 9497 2005 188 16 10651 11352 384 158 1 
+2,4333 204 530 7804 4028 89 8361 10839 99 44 26 122 10097 10807 4 3 1959 4706 684 2578 11086 1862 1 
+0,17 7905 32 5324 44 139 104 7997 5 2269 15 4 81 9228 134 18 3 38 159 10 5359 719 1230 5 1691 980 6 0 3208 4 0 3719 2061 1 
+2,2318 407 22 7 3505 5 0 59 252 181 3 1443 6 5529 1112 109 15 46 124 3 1142 6 77 1995 665 847 64 436 4935 14 214 3 1142 144 5 4490 22 7 3462 8914 1 
+1,4331 1046 11 2184 4638 5 1046 5291 23 0 87 9218 87 592 20 1046 2 10926 227 1192 6 5779 4 35 4480 5407 188 0 2403 1231 101 2423 2385 4 752 31 2696 15 595 8 1 
+2,24 11 3903 5884 3529 9 0 623 386 4 678 257 1458 518 797 3630 238 575 2338 7584 1075 7855 6231 2 3 57 47 297 66 104 768 70 1 
+2,11 1910 195 775 3 5162 222 274 9 0 1182 4689 79 22 7 1555 4 1584 9 0 43 3053 664 2453 8 86 42 16 142 2507 29 222 2 6130 0 135 5550 2 200 21 4471 4 0 1028 10143 
+2,353 3 10827 568 2179 957 2422 3365 48 3361 15 2973 355 1921 635 170 66 5981 737 274 102 2422 209 1202 6304 1492 1678 19 11004 243 982 6 8476 19 44 10016 635 9644 944 1375 1 
+0,6396 6853 831 6806 1894 2018 10 19 284 32 478 1215 299 328 22 7 1304 5015 6 3556 815 0 212 299 1884 22 7 1252 1158 22 69 10 657 82 3 2136 1294 9927 1315 1 
+3,2916 863 6 381 395 342 747 855 747 6759 98 1063 1938 840 13 5 2111 2 1652 1770 1708 342 4012 97 1 
+3,431 2605 5767 1305 2 0 1222 5 0 43 779 4 0 47 12 487 2686 4 27 5699 5640 39 3 8864 9223 53 407 2693 19 27 66 819 153 53 0 2155 75 47 1 
+0,6923 7720 13 4915 51 2 3592 60 57 4 0 199 8665 897 5 499 3 5710 825 161 190 1406 1470 9740 99 0 696 22 7 4536 6 437 1499 1 
+2,260 2445 4591 32 4455 1317 10333 944 6159 15 7301 6282 2 1127 144 2247 14 0 3763 1355 177 1 
+2,151 205 693 468 70 110 19 31 414 948 22 7 815 2 10696 9 0 88 128 15 52 124 0 112 144 5 0 146 22 7 99 263 438 3078 86 1127 3 1831 9 0 56 14 0 177 1 
+1,7939 3822 0 692 267 181 3 1515 7995 4 1775 243 26 561 109 6 1030 10445 2 157 4669 1 
+3,6266 4 11274 8028 6 1466 4 239 9928 6 8532 2 4661 3 3614 4 230 2327 13 9971 4504 20 232 1133 1 
+2,11 2465 23 1171 355 6607 22 7 420 2738 79 70 5 3 4908 3364 137 115 226 11 137 31 0 8331 4 2855 101 6941 6712 6943 4700 7600 54 31 8 1255 5 0 6608 359 9 0 4966 664 1 
+3,2293 2407 832 172 105 2 9854 16 2293 2407 1177 828 1945 927 12 0 1020 4 0 62 22 7 1287 6 265 15 232 15 1866 588 1 
+3,17 41 45 5347 32 892 3 1553 4931 8 35 2793 40 1745 283 12 9528 8 81 1314 5374 191 5 0 63 65 4 0 283 1 
+1,125 11 63 216 8042 4 7853 7082 54 9197 2 6068 7 1803 9639 39 1582 579 107 1510 29 504 14 0 736 1 
+2,74 2400 10700 406 1581 1941 6 3068 852 3412 502 3142 146 22 7 176 2721 3021 3979 5 0 4866 22 9089 23 0 59 1327 630 8 70 1 
+3,299 328 0 1564 731 2 246 0 56 5 44 3 2027 29 439 1266 118 289 5709 57 786 4841 874 1 
+1,4079 17 11 1451 4 475 892 182 2787 82 2172 295 2 1190 7 359 9 3 63 106 8 86 15 476 1819 349 2 157 3 2729 358 20 2963 5009 1 
+2,5837 539 246 3 48 145 997 82 3061 6662 5 0 1013 340 6 0 826 1585 1552 60 0 623 11211 145 14 0 1497 9 0 131 60 70 110 10278 3 115 139 979 23 291 108 44 0 575 1896 2998 1 
+2,74 165 31 3688 109 10 0 2634 4 0 2039 4 2615 54 760 2694 6 3124 3 306 508 53 3 685 1442 1522 64 2650 327 4243 39 1129 506 23 6069 851 64 99 160 78 732 1 
+3,683 4 324 97 109 1759 82 888 4 529 14 6456 696 4536 290 0 7179 15 0 413 939 696 611 2 9502 16 1424 4 1 
+0,4340 2875 30 8772 2749 23 910 10345 84 1720 933 3953 6 7150 383 64 51 6 933 1203 6 463 64 1251 71 198 7885 123 1 
+3,5939 4907 11389 2019 140 5384 166 3 2762 9301 42 4252 6 6351 6638 32 10806 8 19 5937 35 168 150 2362 1675 2517 2189 23 60 4 150 
+2,24 3040 3479 8674 1038 6207 31 3377 81 15 781 613 14 1645 2362 1 
+3,273 11 370 4 67 977 2 3 38 173 1210 1343 3 1072 109 1214 5120 1697 50 9563 5 100 0 5523 2559 4 314 5177 239 3084 6 0 92 7 1765 1543 8734 1 
+2,7195 205 27 3161 2073 348 0 492 396 122 2 0 4511 4 0 59 252 5234 155 1285 4798 10 2546 266 722 68 477 38 333 1 
+1,2333 2384 4303 2278 215 1460 9 3 2792 12 260 501 244 1641 2850 6 26 2246 6123 2850 14 2278 22 7 11307 28 24 1 
+3,80 93 17 1630 4 116 58 29 200 35 1055 202 143 1085 8 86 39 0 47 7 182 1215 586 183 394 2141 657 2454 586 874 1301 73 3777 153 1 
+0,1597 267 1038 6146 8 28 1258 516 3 38 156 208 62 1418 20 0 634 107 84 123 16 193 1337 4466 1113 2 1127 478 345 61 116 58 1 
+2,1185 1163 414 6924 6312 795 550 2 9922 2 168 1497 1027 52 510 10740 9306 2 0 8943 9916 39 2555 238 8234 2896 6515 9 0 43 72 1 
+2,11 122 4 26 10698 813 0 1542 75 110 14 0 4214 59 2854 2320 1 
+0,2957 6995 308 62 6745 6729 8 81 172 0 345 4 4332 3 559 1026 4 481 365 4501 1577 2 0 218 981 8031 459 6 5145 5 0 88 128 64 15 362 2 34 4486 4 602 746 58 7 4973 2369 3042 6 7134 66 226 102 19 506 2 10420 481 365 871 8332 1094 365 175 3569 323 113 4 120 4332 10437 2 15 176 4 2103 8192 1 
+3,325 22 7 2059 1076 4241 1165 25 147 3 159 578 487 207 42 680 67 96 1154 16 898 1264 0 864 4 3 4962 5 2227 370 10980 2 2112 16 1103 1988 1 
+3,11 150 4 4135 4260 3807 97 6 3907 3007 6863 32 473 2 3 8622 8773 5184 980 8 0 1310 4 339 3907 3007 1323 2 4260 3807 1 
+0,963 0 2228 393 4 168 2609 371 25 1516 3 10189 5 0 56 426 54 277 309 0 2609 30 3099 973 14 3 1304 2117 187 2 6988 7493 677 1440 4 1659 14 0 2446 9837 62 1 
+0,11 258 1155 3340 3588 10 55 154 5 208 6 908 71 5450 25 90 1847 5 313 1 
+0,7248 2284 58 18 70 19 29 1652 3 227 939 10484 4 2046 6 3127 3699 1264 16 56 4808 1419 20 1034 94 1 
+1,11 5035 4 21 126 409 2614 5 100 32 1154 2 9081 180 601 1571 552 3713 56 444 3744 23 0 484 179 4 253 1 
+3,4123 6058 4241 387 32 720 1032 255 159 207 42 0 420 1076 22 7 2462 6543 6290 582 383 30 256 9361 20 2780 3674 9 587 1682 1414 2 3339 5 6 60 4 325 1 
+1,7608 401 146 543 4 10203 3 3223 5 3965 4014 8806 8334 4399 7 142 2624 1 
+2,2304 3490 4 26 704 31 9093 2626 21 2539 1103 5460 211 278 5 3 8809 150 2856 2321 280 3 234 2 4522 26 417 448 1544 9709 1 
+3,412 17 750 96 918 3 617 159 1210 1571 3 1072 8 86 10857 293 20 2163 598 5 100 6 617 654 10 347 984 449 5 171 6 605 25 76 90 1534 1032 20 293 704 1433 1 
+3,27 4611 5576 3451 4 618 652 3094 442 5 833 9 0 63 699 162 10464 654 10 0 390 7 994 2702 13 10747 1 
+0,17 428 1660 11329 179 1696 9552 138 73 1866 103 380 20 3 2504 789 8618 20 0 6714 25 90 985 1235 5 7305 7 1166 3413 2550 18 8 81 1 
+3,80 93 17 3901 4205 408 37 27 376 305 373 341 397 91 410 415 402 3880 98 286 223 423 418 36 3880 98 37 27 36 29 265 50 186 570 588 71 459 113 4 16 235 584 6 337 217 3 194 4 16 2557 243 3 5290 0 381 395 62 18 8 81 1 
+0,80 93 17 3007 1689 3881 97 8 86 18 19 85 488 0 92 7 43 8883 481 1680 12 3 3160 648 3726 1023 10 1415 981 2 9568 10282 2126 1 
+0,1 37 451 36 37 451 36 37 451 36 129 3931 6155 37 451 36 37 451 36 571 17 7741 873 9789 97 37 6050 236 36 8 86 147 3 3742 386 9 11096 194 158 416 2 10746 984 6 1252 191 9590 545 158 85 34 419 865 23 0 11102 37 451 36 1 
+0,129 5770 6756 80 93 24 11 393 4 406 54 167 8 0 156 1557 173 1146 1473 25 210 9029 0 393 405 4735 51 2012 833 9838 23 7167 4176 2127 0 973 2446 14 2814 113 4 0 41 45 1 
+2,74 17 5777 7290 124 26 63 112 144 14 0 151 177 86 39 457 21 1216 83 26 8886 23 0 146 7 460 364 4487 1 
+1,27 218 8 0 1871 2316 3535 1968 1522 29 4513 14 214 48 1093 49 94 148 56 134 1 
+0,17 353 3 2027 4 228 3140 1761 8 70 209 3 2088 20 6340 952 3092 41 45 67 303 3892 97 4 10227 0 1166 3413 53 0 482 1 
+0,117 57 4 0 43 2541 4410 4 0 406 12 6283 1872 8 0 156 48 216 2774 399 32 473 2 200 1817 49 3088 2 0 8279 6 11313 9496 1 
+0,117 5696 326 938 0 685 10164 7330 2509 21 2083 52 358 5872 837 8655 38 654 582 0 9111 179 150 31 206 8 3 905 422 1 
+0,24 7225 5278 0 390 7 170 104 242 2542 6 2521 32 9604 505 594 3167 2 32 3 5601 12 203 427 367 1 
+2,74 17 1883 2380 5541 314 4764 4326 4093 2 477 0 43 126 5 323 103 2 10659 0 288 7 59 1740 99 346 112 144 8 86 1 
+0,11 1275 4 48 295 992 2187 1507 30 76 9710 5054 9 0 2196 2107 6 135 3219 1860 11000 367 20 35 371 3 512 2440 228 1309 86 5 3 10898 2458 2 0 2048 240 1 
+2,74 7894 31 3568 23 0 258 869 2 246 0 10145 872 5458 1821 292 15 401 49 1219 64 3414 60 4 0 177 9 3319 3268 1153 1 
+2,24 7412 6088 1843 2 1220 2128 880 39 1524 0 5159 19 9542 8 26 9135 2969 5974 26 2062 6 3564 1457 3 865 4 1256 1 
+2,24 11 6842 7128 534 29 34 901 57 65 789 61 2239 916 5 1568 5208 0 6864 1959 7 43 216 23 776 555 335 135 357 5390 2545 1 
+1,11 810 5599 23 0 9395 1210 0 2378 4 359 4 7795 86 2778 0 643 44 2 5170 1046 123 77 4 16 810 64 709 5 991 443 310 2378 476 148 77 3189 770 1 
+3,24 1932 476 5437 2279 3876 408 58 8 86 1972 0 1096 523 3296 5 2780 6 4707 264 5202 1 
+0,1173 1065 1115 3 142 2433 1854 10 138 897 405 16 5308 140 34 3732 2 8803 4 175 644 3939 320 6737 7218 3049 6 8998 9 0 520 5349 3501 43 7200 1 
+2,165 85 34 1486 2 1024 40 65 22 7 311 364 5395 15 2732 0 356 9221 4 99 72 851 55 5600 133 5431 169 63 10854 9 0 43 72 417 165 85 34 1486 2 2795 12 1018 4851 3406 4 9265 2658 165 1 
+0,24 842 30 10209 3 728 4 173 1150 10666 53 0 6676 4261 5 21 773 2 11244 9947 4 0 9945 11345 1 
+0,24 27 79 4 242 13 3701 3 2259 9883 3667 4 0 6728 7279 40 162 2 596 0 2044 4 4694 806 3737 6 135 8644 5 3038 2269 1857 1 
+2,24 1906 23 0 8375 4 734 2 200 0 753 535 79 1202 1357 1475 14 1715 86 8 3 1477 106 14 0 4166 2927 10 1097 12 26 668 1979 0 5048 1 
+3,171 13 5017 5 10115 5 0 1674 9 7952 0 176 67 2685 10944 4 0 314 67 239 3084 3 634 652 317 31 2696 15 595 5 3 218 86 20 0 314 306 354 4057 11 1 
+0,5772 25 1115 3 1521 424 9 1162 161 3137 9 16 7348 9995 595 46 10240 3 623 616 3374 1 
+3,1583 537 4074 7308 13 5438 2279 3876 1972 0 1096 523 25 3296 5 2780 6 4707 264 5202 1 
+2,74 204 820 2426 634 6650 6918 6134 86 3427 77 92 2701 8 189 234 2 457 0 288 22 7 5726 112 144 14 0 151 177 0 384 4 0 558 1821 438 9 171 1 
+2,2948 2881 867 189 8904 624 145 6 1516 0 75 441 4 0 2290 330 12 3 186 115 115 186 115 73 108 44 6907 7230 1 
+0,117 3 1443 2540 10 5081 3 2453 5 1523 308 270 7 2911 915 13 192 2 3682 20 498 6 9624 3 2717 626 8410 9166 12 3496 4 0 1203 10 140 11216 1549 1336 49 50 0 9286 4 593 914 163 1 
+2,4170 7463 123 160 13 141 1132 933 52 6 4490 2668 6516 6361 284 76 108 0 119 177 1 
+1,125 315 7 107 6 16 4897 5 0 298 233 7 906 9935 29 1771 8 28 9 717 349 84 5080 3 75 864 9 633 8855 2 3643 0 869 132 0 88 749 1825 5 1 
+1,744 17 100 867 2 3313 67 14 57 104 2450 1016 106 8 95 39 21 475 8 0 227 750 941 2588 209 3595 1287 2 3 38 661 21 317 14 0 150 2665 9023 18 1 
+3,80 7941 406 3431 51 2 1427 43 72 5029 1420 75 65 42 967 18 0 9910 3245 18 930 419 50 0 545 150 4 0 9669 143 1 
+3,7844 7 606 1737 4582 0 152 610 7 3531 872 1791 5 512 228 109 3446 0 43 72 3 150 25 4595 3 894 8183 5 0 4837 1221 44 327 2599 1 
+2,2359 89 3 8767 1085 15 3 3655 42 6259 10078 0 262 73 1494 5 0 1189 6305 20 3693 9 3 11137 6 320 9 255 5 0 939 3001 22 2253 323 2050 224 75 110 44 1 
+2,571 11 471 331 64 1131 2 11366 3021 2866 122 2 828 1602 11235 393 1569 31 762 3004 1615 54 25 2552 0 11279 10519 9 2866 965 4580 85 32 181 138 1806 9 0 7466 391 0 5066 2866 174 404 6 1615 210 10011 3919 7548 85 32 90 5478 2 3516 12 1 
+0,351 66 7012 1092 22 69 167 245 49 3135 61 40 42 278 351 85 10933 0 8346 9888 335 0 7320 11173 4 0 1166 1 
+0,117 3 216 1138 9 0 365 6 1484 2602 3 512 2440 228 11263 3 895 228 7 649 5 0 9535 4011 507 1807 2123 2 2123 399 7088 6 4011 30 76 5054 9 0 3219 4997 4 35 371 129 4087 6243 1 
+0,11 1506 4 1069 2562 470 371 54 3683 2 6965 13 4454 0 398 11029 2 718 4622 542 72 52 1273 878 8 238 8195 4 101 152 10429 367 54 928 865 35 718 8051 129 6937 7915 1 
+2,273 204 820 2426 3816 5800 2021 53 158 1455 5 433 330 1327 12 3 115 186 115 66 224 44 6844 7331 4 6262 292 8 86 110 1 
+1,1370 1506 7542 7959 10954 0 5217 4 11 4117 4 0 7459 9 5668 5 5313 1 
+1,17 845 742 161 576 1851 4 617 644 20 743 1007 1265 298 32 1154 2 4692 53 2285 0 41 98 5265 354 680 8 95 1 
+0,1643 91 1176 3 357 43 106 1269 5 400 4 116 1598 2499 69 5242 1701 44 287 0 156 208 705 7 4656 695 31 3 246 71 3 9208 1 
+2,74 17 1883 2380 5541 314 4764 4326 4093 2 477 0 43 126 5 323 103 2 108 0 288 7 59 1740 99 346 112 144 8 86 1 
+3,4156 2085 2662 67 303 2 556 51 44 883 3514 9705 23 5653 62 105 2 1216 80 93 17 27 4156 2085 2115 3892 97 0 393 48 56 67 62 2 556 531 170 104 9 2543 1434 23 3 3353 1 
+2,151 7839 7003 1073 0 63 126 2 8324 60 4 0 59 2464 630 86 102 52 31 1702 323 351 20 6935 5852 4 942 5 35 5585 4531 14 3757 2681 1 
+3,1 37 27 376 305 373 341 397 91 410 415 402 4246 98 286 223 423 418 36 4246 98 37 27 36 1756 132 0 4503 39 537 6592 4237 58 37 27 376 305 373 341 397 91 410 415 402 3997 98 286 223 423 418 36 3997 98 37 27 36 0 262 66 41 45 3624 9795 1785 8 95 18 19 85 479 7484 9 289 66 153 1 
+3,7749 2076 698 5 14 965 546 1449 3276 51 2805 6 2592 1386 10069 67 412 17 11 2805 246 3 133 65 173 83 0 702 95 15 10929 61 302 466 175 2180 301 1018 5149 50 0 1 
+3,117 0 202 22 7 43 106 4 1332 485 2463 3641 6 9201 400 14 3 9263 2117 12 0 370 2540 118 1210 311 530 113 293 61 225 19 1823 165 31 5 135 5602 2546 0 62 22 7 882 7576 6004 6 1 
+3,577 0 281 8209 586 183 394 4 116 291 203 6 67 8485 2 3 38 159 173 485 209 3 1147 122 195 1 
+2,847 22 491 975 1932 3 87 3139 150 87 57 4 0 48 71 77 1293 362 2 1435 1928 22 7 594 237 691 1002 1637 0 152 6 329 1069 474 249 9246 10 19 22 7 1238 2 3345 19 22 7 2853 1890 54 2429 675 0 127 1 
+0,4366 116 2266 7826 232 208 2046 116 2 16 1698 4604 8804 208 5063 10 30 1015 10184 117 1723 116 2266 1976 5 3 3616 5329 198 0 215 808 4 0 1023 116 2266 672 21 9534 1070 5550 9 1227 2 8021 395 71 399 2 3 1547 1 
+0,1 37 347 36 4111 37 347 36 11 8369 10247 2577 51 16 10654 
+2,24 469 106 39 5184 3 5571 1427 8 1556 4666 3937 4267 0 7520 510 21 50 279 6 296 0 238 1932 846 10878 39 52 713 3 2177 1729 2 111 7226 8 70 110 1 
+3,6148 3 9115 13 795 557 42 19 7 0 674 923 2 277 1 
+3,11 258 301 23 0 56 945 4 3905 1459 8307 120 600 32 99 2216 1362 4621 191 30 1264 1058 241 387 57 2215 10 381 766 25 90 4523 248 75 47 160 1 
+2,74 205 7913 8 11376 675 5 0 1724 9 6486 5950 4215 1027 21 59 112 13 562 38 4215 10096 26 57 343 5033 4427 42 3503 513 5 505 26 3612 6 131 675 8 0 9805 59 1 
+0,2897 32 5501 11258 8902 53 10974 10 68 4731 990 71 3157 837 46 491 4896 8 3 38 1028 427 10319 2472 1 
+3,273 17 11 41 45 1904 6 2890 1578 105 2 527 4448 9666 2 10439 2209 10 3678 3 5061 197 0 990 6 10961 5 11441 42 2161 3172 50 0 3665 4 322 1840 187 2 2522 296 8 95 1 
+3,7807 7 88 1164 29 32 3 742 106 9106 8 0 1291 5 84 19 68 992 21 3308 8217 10495 318 1 
+1,6060 6035 315 17 27 576 906 8559 5 3 1681 5 298 11076 26 196 52 13 1463 9 717 211 118 3 547 552 5079 503 738 146 148 46 30 3217 315 1340 2 988 203 82 0 1760 1 
+1,24 2271 1535 509 67 96 3353 1433 32 90 3278 211 651 69 1099 10 2 75 2532 148 1 
+3,1057 1166 91 1055 5768 25 18 19 29 479 4076 91 6910 3 260 1207 6750 62 10 10019 0 199 156 725 434 729 5 171 9 56 297 1573 104 1 
+2,74 17 219 271 54 25 124 339 112 665 5 0 59 2132 18 95 52 31 3549 60 4 136 7 2258 364 3454 625 131 2 777 79 3448 687 2873 0 864 2 1832 1 
+3,17 27 41 45 2440 228 1309 8 95 10 5854 3898 58 5757 98 140 5214 12 16 645 2 479 0 2442 4 524 7789 3898 408 6860 3251 21 1716 2165 20 0 637 1642 1174 2 2457 0 338 3 1820 9 0 354 18 1 
+3,80 93 17 41 45 4354 96 10126 9 8338 8 81 39 3 8284 48 1527 1302 1513 60 4 10888 1027 967 184 1311 1806 2 0 11268 1105 0 710 807 1419 5 0 143 1 
+1,836 100 17 2388 1200 1268 1454 1071 184 2 34 5 982 4 0 2330 1579 3468 5 0 269 179 253 382 8 136 42 0 11372 4 35 294 0 4873 480 1383 334 432 64 3712 1 
+2,6859 536 24 2404 2119 7935 5818 23 1197 7622 6 4120 6528 23 5922 1141 339 47 8654 12 6920 8 95 1 
+0,1177 158 2189 4 10188 3687 5 0 88 128 49 61 4776 23 0 1503 47 187 2 308 303 4972 7356 442 20 217 73 161 44 0 398 1291 1 
+2,74 573 31 291 14 0 453 573 31 291 14 0 815 573 31 291 9 2265 2265 4 0 311 666 1602 781 31 291 687 2873 7 781 14 0 1248 532 1032 10800 75 110 219 271 124 333 9 0 384 6 131 72 5 21 998 282 14 455 177 1 
+0,117 3 218 0 308 303 10098 3 3160 618 9 0 520 811 307 163 3190 10 2189 4 805 320 500 6 10544 23 426 2189 29 627 2730 44 0 119 339 103 1 
+1,24 11 1198 29 32 57 1829 1370 285 9 35 1432 427 829 1909 5858 7553 427 966 12 5174 2 553 0 127 365 1100 2585 211 46 1273 34 658 3533 2 664 0 636 1432 7 1216 2 738 817 554 20 2277 5797 1 
+1,353 276 4 3651 3650 20 0 1560 8212 11413 20 2368 2278 8 1745 5 7539 7315 7480 18 0 345 4 0 2113 85 452 5 3 4676 4 389 1 
+0,11 4424 3137 55 10072 10241 3 616 3374 602 1518 133 723 4 455 3137 10073 262 5000 32 90 147 1 
+3,2395 2302 0 92 22 7 63 199 220 765 9810 109 18 10 19 85 1753 7920 5614 5680 5 16 2724 264 20 5667 2 9283 
+0,11 603 7481 591 29 2717 3 4619 15 19 3706 2 9376 6 4967 2659 16 1830 27 10823 1087 3492 29 8671 0 4619 5 8068 2 10785 9 10418 2537 1 
+2,3866 7902 25 1105 407 35 63 59 3236 112 144 5 48 276 457 0 146 22 7 186 9658 998 5239 1 
+1,411 810 492 6 709 3 692 15 52 8117 3 142 2563 197 538 6 0 1599 2406 411 388 1316 18 8 136 1 
+1,836 100 7056 5072 2 1300 540 480 1385 334 432 892 35 783 8 3 5298 628 6 3192 9194 5 253 8 136 1301 706 10 3 2708 2 0 869 5 0 484 179 55 1231 1214 8255 197 540 1271 11 3192 197 41 45 1 
+2,4337 6567 3427 3 92 159 2 1427 5 151 0 5625 59 112 9 171 248 16 5165 5 3760 59 252 8 136 102 
+1,24 429 0 162 157 1700 9958 9 315 2 4747 0 2650 743 5212 107 3459 5 298 0 88 749 6 1645 3526 30 5 3 8850 44 230 657 2 222 2 1826 0 515 5 21 633 127 1 
+3,6081 272 11 345 4 21 4478 563 0 317 456 2500 1314 10 101 261 251 124 40 162 22 7 424 319 5 272 0 781 4 0 2992 4 126 128 18 136 1 
+3,80 93 17 464 160 64 245 2527 10 600 491 184 1272 5 0 2544 4 0 2203 14 168 221 0 5312 4 2539 3604 14 0 151 177 284 154 4821 203 1 
+2,125 4114 1702 0 88 128 2820 1346 5 21 59 146 7 316 2136 346 137 241 0 384 386 5 3752 59 1322 9 0 870 2476 1 
+2,24 11 461 3854 473 136 2 733 5261 7004 5979 2 0 7048 6297 9 11250 5789 7209 427 155 0 3854 68 673 3 768 759 12 0 1196 3865 3248 157 3 811 337 2 0 2111 18 1 
+0,7202 5217 30 1782 8716 9 320 14 0 10721 4 3489 4976 10268 3406 2 9884 5187 4526 306 2545 3696 6 278 372 446 9 35 10647 1 
+2,412 125 3869 6875 18 52 29 494 202 132 3243 8 26 618 15 642 6721 1075 879 26 808 22 7 169 11378 2506 2 536 5 0 384 6 131 3053 1 
+1,7240 196 1418 20 8142 6 1656 1980 48 2216 1837 9409 391 0 10590 935 1223 109 515 6 3749 3 393 4 1006 1061 1977 6 142 134 18 1 
+2,3873 6457 8415 3 321 243 1788 1345 2 494 3 77 492 231 39 0 63 346 4 0 6581 6654 2865 5 413 939 1 
+2,6604 5 1191 499 94 3405 5354 796 35 11387 2 553 60 0 1257 0 1500 0 8754 0 9980 847 436 9752 0 5354 1553 102 46 222 503 8 1745 1 
+0,156 523 1166 91 58 18 8 86 10 19 29 479 4076 91 690 84 675 138 4 171 22 7 580 725 434 1817 9 50 1573 104 2 1733 3302 53 171 22 7 1726 448 143 1 
+0,844 4733 7524 91 58 720 3 3622 627 5 63 158 207 8 478 61 302 2169 188 16 43 158 15 3 183 62 42 485 10714 0 202 5 589 1332 
+2,7780 1876 6777 9400 6 9384 40 421 65 44 320 4307 7567 5 136 22 7 6577 73 3809 14 7538 1 
+1,17 2847 4398 4175 124 0 176 288 7 2445 112 144 14 0 177 8 136 11181 44 3 482 2206 4 232 357 2652 213 4 1442 11407 2 108 0 311 666 1 
+3,435 67 96 627 183 999 5 4438 704 1782 771 597 211 0 9756 13 9066 9045 641 541 1 
+2,5925 7964 1810 133 1462 44 441 2728 889 295 0 1035 642 295 2374 2834 421 0 3051 2846 3940 4253 139 226 8 136 110 1 
+1,6419 0 2633 692 1483 1746 10 16 238 642 683 294 7005 5924 85 494 122 26 8400 9 0 6766 374 594 237 1 
+1,27 728 4 4930 2456 25 3591 21 669 622 1302 5 0 1584 359 6276 515 14 214 739 94 639 64 321 71 441 4847 14 0 5887 467 2588 15 294 7586 6630 8008 3 1083 1 
+2,2847 22 4398 4175 1073 0 1246 1563 2 3554 9 0 288 22 7 311 666 63 346 14 0 59 252 292 8 95 1 
+1,11 2459 4 3 2419 237 489 5 375 10 2604 401 10220 3274 860 23 3 2419 2236 95 9 3 5597 4 5400 23 0 3292 390 1 
+1,685 11364 6890 7530 29 548 60 8 33 155 52 13 2 34 2206 4 26 144 1 
+2,532 8 0 1240 106 219 271 2429 550 2 1020 353 1544 138 530 3557 5 59 438 271 31 3 5091 10825 75 110 1856 26 5463 8407 3 5470 65 9 0 56 1832 79 1 
+3,171 29 2742 50 170 186 153 44 48 2 77 103 5047 350 4 1480 4 67 3515 2 154 10655 16 704 1543 5 0 279 4 1535 96 6 984 1 
+2,2334 6873 209 21 1333 1294 137 4496 1793 12 3378 613 2314 6514 39 3 1709 898 95 110 1 
+2,11 465 6601 6425 2218 77 9631 109 9 3 5102 46 181 5 2179 0 146 22 7 99 263 131 42 18 345 85 76 34 865 6 414 948 4 0 88 128 85 553 26 112 144 1 
+2,211 15 657 15 1832 9335 7485 6517 6 149 7127 30 3210 120 271 510 5 151 9072 120 1953 510 5 4160 5 3759 1 
+1,125 27 2336 2195 8437 5 764 3002 31 8830 20 616 1491 5 21 579 7571 8137 475 179 476 18 1 
+3,6059 3363 342 3901 4205 18 195 19 85 265 50 459 1016 3175 4 16 2254 161 1493 235 584 6 337 3 194 4 16 2557 3747 15 368 4 3 133 47 5290 1 
+1,6551 0 4939 138 1698 7799 6120 9463 951 956 22 7 3292 3893 3968 1736 10261 419 4 19 23 26 94 455 8755 9389 4 
+1,1377 17 27 343 6 48 288 64 492 2025 20 935 6 41 45 554 196 39 35 3720 1513 164 3 3182 8 136 3 41 45 388 679 18 1 
+0,1612 4108 7962 22 7 22 22 7650 2 6644 22 22 6 6945 7612 22 7 22 22 6482 22 22 4602 372 446 166 6489 6 7336 68 1225 122 21 3301 2544 1 
+2,126 10703 2968 6360 1249 14 0 1862 286 8 26 131 492 33 4517 3 4620 231 5 0 59 682 364 77 1792 10559 653 6 2437 6782 7963 4 171 2 494 0 112 1 
+3,11 370 4 67 1004 192 2 9327 9388 15 967 1252 10 16 159 1224 145 140 566 767 1 
+1,24 314 101 2423 2385 181 21 11221 816 2 2288 8 33 5047 9210 14 0 4927 4 0 259 2698 233 7 8147 267 3 65 132 2535 9 3 38 294 1 
+1,74 205 3019 210 0 406 1665 69 118 3 3948 1638 46 491 49 166 0 4213 1638 4113 3984 5183 3 77 2594 33 6 6205 6030 989 5 48 675 15 0 406 1135 2 35 1240 2728 5 441 276 139 226 44 325 1796 203 53 0 112 144 137 1 
+2,4109 1197 2960 8780 6576 5948 1987 0 4463 4 2337 7921 14 0 7535 5953 29 154 1225 21 8182 4 2770 2 1197 2960 1 
+1,125 27 565 4 6347 7102 7 5084 11 7559 6 255 1725 10099 20 0 778 1387 3118 64 5415 23 21 4198 5116 20 1417 6 9434 10569 369 18 1 
+1,74 205 11 1502 5 0 311 666 64 3506 118 3 2717 4 120 7 2 979 33 102 3 11032 229 4 10357 360 406 4294 2872 834 2315 6 870 336 2969 4010 29 495 2 815 35 438 53 0 1246 452 14 0 151 252 6451 146 8354 169 909 5 903 136 554 20 2872 7 72 4 374 2264 1 
+2,111 1649 22 7 4281 3063 10908 2 112 5 0 288 22 7 998 5239 5 3 38 92 159 72 3063 384 5 1400 133 103 380 355 325 22 7 4087 6966 5 3 72 4 77 1008 822 1 
+2,1885 5806 5831 506 23 291 2 11103 7109 5996 115 66 8 33 310 6434 246 133 421 4384 686 7763 2 2158 0 382 231 5 0 1898 43 1441 1 
+0,4259 1312 2628 4745 4921 5 0 5824 6 49 2026 1257 1858 30 1080 9 6383 213 4 483 1559 6384 2543 354 680 70 1 
+2,111 1649 22 7 285 8743 4281 3063 75 65 10574 243 92 159 72 1844 5 21 784 188 3 3605 1842 1527 5 684 1 
+2,2832 1799 9 77 63 779 3361 5 351 1008 8 33 2 355 2973 139 170 9828 3 352 47 266 159 4 2257 3420 446 568 3 386 5 0 176 1729 4 2303 2738 1 
+3,24 2324 67 96 84 32 90 3 4853 5 2788 99 41 45 10411 44 0 421 77 2028 30 2213 333 40 47 532 0 293 509 67 96 4605 0 49 1306 704 1433 5195 2 120 631 1800 232 302 2 34 3 8219 47 9 0 41 45 221 1 
+1,2348 1119 32 2177 48 22 906 22 579 7300 8603 9 33 57 2 34 1108 20 4129 7011 7854 6 0 63 20 4129 5965 2277 6354 1 
+2,24 7747 6678 6 5843 6632 550 1142 3916 4096 550 1995 532 414 948 550 5075 11 88 128 11264 16 1740 144 9379 2 321 33 110 0 182 248 0 406 124 819 14 0 8329 1052 1167 252 5 3760 532 46 387 76 34 1101 1032 1 
+2,3845 37 1677 36 5894 7194 7919 7 6987 7396 5617 37 1677 36 3845 37 1677 36 74 205 24 7070 7173 4 313 124 0 1477 33 5 66 552 823 1008 323 909 1 
+0,7687 204 5671 250 270 22 7 7644 10588 25 1712 51 2129 4 654 8 5372 4 87 1590 6657 87 10 599 1014 4678 0 643 1 
+2,74 17 126 834 2315 3587 23 1853 3712 2 108 0 1991 5300 59 146 7 311 666 282 8 33 5542 49 4980 726 5 3 10339 131 1 
+3,466 223 32 656 2 3451 51 2 21 4471 181 5 1582 20 4147 4122 22 7 612 483 10915 944 5991 54 18 313 31 87 930 281 1244 22 7 4865 3303 143 1 
+2,5853 4250 246 3 77 145 997 7075 5819 436 1260 6 0 461 1175 355 0 2327 3829 351 115 8 33 5 0 11090 4632 4 3 77 137 728 197 0 1618 940 726 1 
+2,6605 7762 329 3080 550 2129 4 145 563 42 701 22 69 10896 263 174 614 2 494 3102 4 19 3080 31 4794 5 0 1453 340 9 3378 4069 6824 15 0 1175 720 21 351 115 224 44 0 3829 1 
+2,74 17 428 4832 7165 7010 10435 5 6989 6449 5 339 5492 3617 8 33 2 108 2289 35 63 112 144 14 21 59 252 517 61 822 552 39 1744 203 2 35 43 1 
+1,27 343 6 48 288 64 492 2025 20 935 6 56 554 196 39 35 3720 1513 164 3 3182 8 136 3 56 388 679 18 1 
+1,935 101 4027 4083 13 2 816 1061 2 8871 443 1551 6 1993 733 1 
+1,11 318 85 777 0 2147 420 1440 10522 9 1113 2575 4 0 938 78 27 6 0 2381 1 
+2,126 2948 2881 1999 189 384 699 403 972 63 2721 7842 7967 115 170 115 66 5 0 131 4 0 7813 161 3073 2290 330 8 33 1 
+1,3899 444 5 375 148 10 77 94 64 709 6 48 716 8714 1861 102 35 10173 2234 1843 2 145 3 3182 5 0 1511 4 6547 1 
+1,836 100 17 27 41 45 5756 2806 9350 1980 1200 1268 1007 2134 5 0 484 269 179 4 253 382 8 28 39 1546 10462 0 2729 4 3 628 14 0 976 4 3 217 77 65 3407 1 
+1,1905 56 810 977 12 4404 269 721 30 192 2 279 1296 814 2590 5 1366 1 
+2,4396 22 7 177 11168 1336 2 0 48 9110 4 2445 0 9007 282 84 1153 0 3156 2 16 3430 6 0 4954 1814 0 5395 84 13 5 0 2103 1 
+2,1202 1357 54 2654 241 2 521 263 1788 2 8667 3 339 2752 231 39 2816 2070 53 3 108 510 118 10 5 0 4166 2927 14 6448 109 1 
+0,80 93 497 43 9304 19 387 1811 166 0 2731 3316 4 3 1320 1330 5 0 615 83 0 1982 1739 211 5 1723 11397 156 1123 2 5106 4690 6 812 13 2015 
+2,24 6982 4250 6 944 1620 1260 8 699 3518 2 453 0 1240 340 33 110 6 0 2751 828 471 331 355 0 461 1067 331 115 139 9 35 1533 1324 108 1 
+3,204 821 250 4344 91 7358 10186 2988 4177 97 172 105 86 204 820 2 3291 21 3096 170 1034 588 6 616 321 49 677 1098 15 0 62 9668 2 10467 
+0,165 68 241 1019 8056 9 77 1008 42 5583 517 61 21 2539 3639 3275 68 6 19 68 494 6 3700 2126 5 1729 1 
+2,1957 528 503 23 0 753 535 126 1202 1357 1746 52 25 5361 14 214 138 4 26 2019 155 9524 46 4834 1 
+1,1 27 388 3519 4684 53 3 2105 5 940 272 515 352 94 360 339 978 0 1874 2905 908 79 18 5 3 679 1 
+1,125 11 633 1644 29 1225 315 7 11336 107 6 906 8129 53 349 12 2704 558 8358 2435 14 2587 82 3 5113 3385 869 5 0 1511 4 298 1 
+1,27 456 13 416 2 34 304 5 1401 8 105 2 9358 0 393 4 3604 5 0 2083 7 9127 3376 9716 1 
+3,638 313 22 7 4505 1386 202 2076 8345 5263 2 0 351 161 661 28 1009 8 2245 1679 2145 20 4421 5 56 400 75 95 1 
+3,951 956 1364 2397 2917 4337 18 52 3145 951 956 25 3735 60 4 0 381 370 8784 8741 10 9738 9 5724 389 187 2 0 381 370 2076 3705 5 0 421 547 103 1 
+3,117 1550 4 702 1849 4 99 0 395 5 0 92 1285 13 2771 49 61 67 6204 67 2771 1210 1343 987 3 1072 9 0 43 72 40 65 1 
+3,272 22 7 101 261 251 25 172 10 52 29 141 1003 10419 0 636 6185 71 6041 15 0 669 3197 1 
+2,6398 3902 6236 6474 5346 82 3 2680 3239 5182 9900 2 108 0 6701 22 7 4038 5673 33 14 7316 7289 465 7399 1 
+3,24 11 5218 4 324 7 529 1106 14 3 73 268 113 578 904 5 0 194 158 0 10760 2117 5 217 48 103 0 107 147 81 1 
+2,6421 4223 4213 209 2005 198 0 3667 2 108 20 3 9722 5 0 5638 320 4 0 73 104 1193 2865 109 14 6255 6986 4223 
+1,56 2979 2836 2908 2837 942 1905 2216 1348 1626 1454 29 34 9233 977 12 259 3228 40 65 15 0 56 388 2660 0 43 894 2590 9 881 721 1999 188 0 259 5 375 6 304 14 3 1803 56 4173 1419 5 942 1 
+1,1384 61 73 161 6985 1000 3 1031 761 8 3 4761 2588 5 1190 22 7 9991 9934 134 18 33 15 4440 3697 1799 8 0 917 4 2936 2 1017 3 8713 906 3155 4 0 359 210 436 246 
+1,3026 946 244 1349 1897 3264 1284 582 0 431 107 31 3579 680 16 563 9 0 100 259 85 9500 0 615 83 1836 1 
+2,414 948 1011 3 38 1768 2750 6350 10 71 52 1011 3 3196 469 9374 9297 10667 180 23 9741 5 324 6 8405 5 227 689 6 
+2,74 7353 346 8983 85 32 90 3 1091 9 0 88 128 288 6269 9 3 224 0 406 4483 4801 20 993 2128 166 3 112 144 8649 6 166 3 79 1 
+3,431 1996 1213 1207 827 172 28 19 29 708 0 1103 5446 19 3097 2 2227 96 23 204 823 213 4 1535 67 96 1 
+3,412 430 821 111 3016 3793 2326 3793 3956 25 473 2 479 2833 6165 58 5774 3 56 206 1299 4 2606 4597 8824 3642 9 50 5742 104 5711 1 
+1,413 689 123 19 2722 141 1132 2 1116 3 934 661 1276 12 0 88 128 2 10275 9 1459 401 622 349 8 0 3205 150 22 7 1488 2797 700 1 
+1,1067 10979 6382 7748 13 4730 39 2436 1224 0 1550 4 26 10112 1 
+2,379 159 3380 4208 4247 22 7 11062 5597 23 0 288 22 7 59 1477 109 31 76 416 2 1471 0 260 79 123 1 
+3,412 17 11 702 4530 82 624 133 65 3438 83 0 1718 6 2805 8 28 5 11100 430 733 12 485 4896 8 41 45 301 6 3 5388 20 0 637 1394 612 789 5 0 65 1 
+0,111 3113 193 2201 30 8 0 234 10 2144 2 246 2197 225 19 9461 0 182 35 2793 497 660 13 0 1650 2 8172 0 3489 811 4 440 519 9858 3 216 
+0,638 17 6392 51 10 5030 5409 42 2802 50 185 5570 466 1096 342 3033 97 37 27 376 305 373 397 1693 11215 9174 9625 402 3778 1204 10353 11018 9548 9546 10352 306 36 3778 1204 37 27 36 123 19 25 1088 3 38 1727 4559 5097 3492 10 68 5090 138 4 185 458 503 1 
+1,1184 204 821 250 538 172 105 28 2 1226 888 4 38 1748 5533 5 0 642 683 879 21 1661 56 1791 1139 8 3620 10 0 1621 680 87 29 4727 0 717 800 1 
+3,80 93 17 7446 955 2326 97 37 27 376 305 373 341 397 91 410 415 402 954 236 286 223 423 418 36 954 236 37 27 36 8 28 18 19 4876 3 338 2 425 530 4 16 1323 2 950 1593 58 37 27 376 305 373 341 397 91 410 415 402 1368 98 286 223 423 418 36 1368 98 37 27 36 9 5685 104 1 
+3,6362 6 529 30 4639 20 3290 6 8498 4 0 38 2159 1 
+0,1643 91 602 7 11170 1465 481 365 670 84 25 10549 0 62 13 884 9686 191 3 858 188 122 2 676 283 1 
+0,270 22 7 77 47 773 2 1225 138 9292 285 2530 122 2 498 13 192 9 3 1443 9171 2398 268 102 1370 4958 5180 29 857 3 9874 10540 
+1,1377 375 204 821 250 27 650 5046 9 57 4 77 406 543 4 11135 3 1093 935 721 5 3 798 3411 1140 3 372 5 228 28 4 375 22 7 238 4793 
+2,74 205 7777 613 3061 2912 8018 52 31 0 1493 54 5511 0 5890 890 2912 123 52 22 7 0 57 54 8090 1025 3 11022 4 7710 2 0 56 3820 3947 1347 1 
+2,7705 7387 11 7410 118 713 3 3504 65 102 46 2211 35 241 1138 8 33 430 5675 3 739 115 386 2 0 7341 5936 1 
+3,17 7635 1164 58 6866 98 0 199 41 45 2519 1076 8 28 18 19 29 3291 2819 2177 1730 5 715 2 858 2169 20 3347 51 2674 9 49 5073 896 1 
+3,27 264 2009 25 5283 23 0 4288 2415 39 4418 52 10185 0 235 4 135 2615 18 0 1284 1 
+0,7675 8594 10 3 1431 38 1992 8 0 2204 1849 4 385 193 13 4886 42 8683 1 
+3,260 2208 2742 50 73 268 153 8 3275 1448 47 217 77 526 15 281 15 46 8685 51 9 4526 3 1026 296 8 28 1140 1 
+3,7821 5 0 179 4787 49 61 1346 3 65 8 1981 118 291 4006 2859 6 6083 7754 658 1714 437 13 362 2 34 2743 8 590 447 6 10789 0 
+0,130 97 22 7 649 2 247 3 216 1334 9 16 2059 307 163 5 0 398 162 10 888 4 350 4 2208 30 5277 2 2484 8396 290 0 
+0,11 3473 7239 697 703 5 370 20 530 113 2 1868 113 3 1261 4 135 697 30 4588 210 15 566 1 
+3,2882 3215 7180 6594 97 8 28 18 19 1518 3 169 47 5701 104 1169 768 2 672 10731 1221 8558 1842 563 2 1169 8682 8560 0 258 4485 5 
+0,11 365 240 5 0 56 13 1005 778 10923 5 16 581 83 94 19 123 32 4979 3270 365 12 3224 8191 514 3620 5 232 1684 1 
+1,24 27 150 2084 1309 28 10 0 2191 51 1291 284 34 3577 9 0 7187 66 237 5 6938 7 5721 6172 2886 225 4993 4258 7468 1349 4998 3935 20 5455 2 0 606 1936 1008 132 0 903 1700 1 
+3,11 370 4 67 25 4856 15 706 50 9588 2 1325 256 10341 60 4 100 772 3637 1 
+1,11 411 107 31 4405 16 1528 522 28 12 105 2 1226 888 4 38 887 5 0 642 683 4523 20 3 56 10795 4 669 2 38 2013 1508 1 
+0,6393 10 5030 5409 42 2802 50 185 5570 3033 123 19 25 1088 3 38 1727 4559 5097 3492 10 68 5090 138 4 185 458 503 1 
+0,1059 725 191 767 2 452 3629 449 187 2 0 258 2057 296 20 0 56 945 4 3905 1 
+0,3052 2952 1628 2610 48 208 1554 8 16 2952 41 45 156 729 8 28 15 368 4 3 1800 172 2750 2 878 8 399 10 1413 371 2 4641 12 716 1 
+3,1 37 451 36 37 451 36 37 451 36 129 1580 4152 37 451 36 37 451 36 412 17 2324 3340 67 96 772 9 3 63 1527 8 28 15 100 3584 655 23 505 16 1487 6 1145 2664 39 5051 4759 583 4872 443 5 0 484 179 4 253 37 451 36 1 
+3,5988 7 239 6244 8844 5100 303 123 19 29 10602 529 54 1426 8 12 21 1111 5417 1 
+1,74 205 11 406 550 3 5461 4 120 19 31 166 5 0 447 10015 276 847 993 901 21 2114 46 64 1209 2 355 3282 10587 2829 2264 2815 28 5 35 131 2136 137 4 0 59 146 7 316 630 1 
+1,17 11 238 790 906 229 188 3908 7 3188 259 802 60 4 3 558 2187 11157 107 8 28 8759 3 216 2458 2 0 127 7 502 3346 717 800 1 
+3,80 93 17 41 45 1991 697 64 391 0 5524 5080 8 28 15 3 1709 191 1252 23 523 1964 1922 2405 58 37 27 376 305 373 341 397 91 410 415 402 3072 98 286 223 423 418 36 3072 98 37 27 36 8746 3615 5143 0 8249 4 3283 67 96 1 
+0,2901 51 12 2222 4831 2307 1340 1682 2 563 8959 1229 9 2150 2 1326 193 210 1 
+0,1409 1363 13 549 3635 173 2226 2 3 338 2568 197 1243 156 175 1299 4971 6 6175 20 2280 4365 0 338 4971 22 7 5713 161 1790 1409 1363 1243 300 29 
+2,74 17 126 570 666 336 2335 7884 5434 219 747 15 0 59 112 5088 28 12 3 1292 356 4 5702 1566 909 1 
+3,6718 1779 866 10 0 10382 8168 4 345 23 0 424 319 33 430 818 83 640 267 261 251 452 160 30 141 9527 4 995 15 1690 20 0 669 1 
+0,1057 4300 1065 1055 7502 25 1010 3952 15 0 8458 2668 5 7597 38 264 800 5156 3840 483 2204 4435 522 11 759 9090 0 1520 197 3952 6 4300 2 672 3 192 4 173 1849 3444 395 6 175 5142 1547 14 0 3840 6 1234 1520 663 1881 9821 1 
+2,24 11 179 775 0 461 1175 0 222 274 2 521 8203 14 4392 1903 8 28 110 39 0 3653 2121 255 346 4 9561 4 3578 235 1711 8 16 8722 5540 4708 1 
+3,4216 267 6558 5860 680 10 189 127 13 5 0 3457 4 3 1728 869 27 218 20 2532 14 0 254 4 0 1195 18 0 127 1722 652 1231 
+2,151 205 693 468 1958 2891 554 3 4492 126 475 12 818 1499 6 321 10405 15 0 88 128 146 22 7 316 79 2490 0 2136 346 12 3 10516 2264 2815 224 44 11391 2829 1 
+2,151 205 693 468 126 1883 2380 0 288 22 7 99 263 336 14 0 1635 252 964 255 144 8 28 110 12 3 1142 5 0 3130 3141 438 1 
+0,2955 125 27 38 314 56 79 9 0 465 608 1633 2922 29 494 82 23 0 5914 201 1540 5 0 238 2403 10503 4 2343 5 838 314 201 134 18 11 77 1493 79 416 2 34 1664 8 86 29 2089 9 0 2922 8 1422 3 314 7636 7713 139 915 752 7 637 608 1347 7482 18 8 16 1341 1 
+2,17 151 204 821 126 834 2315 3587 23 1853 3712 2 108 0 1991 5300 59 9850 311 9866 282 109 5542 870 336 2969 4010 6 135 49 4980 726 1 
+1,24 2379 10407 16 1974 23 942 8 28 39 0 6206 107 1154 2 859 82 10470 5 3 8884 44 133 579 3987 2861 9087 9512 5 2379 1 
+0,129 2944 2980 273 24 4034 2037 6 4734 167 99 0 1667 42 4837 11324 708 0 1306 9 14 214 1012 5519 4 2472 597 532 12 0 3475 9026 1314 141 5360 4 11331 10821 148 46 794 2 478 1156 230 1727 1688 9256 2472 449 249 46 387 615 122 1 
+3,80 93 17 406 1015 35 1698 862 3150 8 72 14 3 159 173 661 5 688 1526 1698 862 8795 2 35 3437 661 5 133 103 7083 7 1373 755 18 8 28 1 
+3,1930 27 1002 303 10 123 19 68 154 5352 4488 526 14 710 301 8438 13 448 16 264 2226 2 21 4435 12 1203 2972 58 1 
+0,130 25 2656 51 1032 255 234 9 1227 2 2457 0 3127 527 4 178 609 2 0 357 338 755 1935 66 959 1334 1 
+2,1636 3743 10773 6 564 11320 11 5304 9140 3200 6 2499 22 69 1248 11 431 8646 2214 5403 8 6 701 22 69 5266 19 1 
+1,2379 10408 16 6634 1974 39 942 1154 2 265 1840 155 5025 579 2861 7999 30 5161 1 
+0,468 873 239 1173 1065 1063 2287 13 1679 1913 2675 5386 1389 2874 9 460 104 5 437 6 202 1389 2874 22 7 193 2592 5467 1682 1997 2661 6 4045 1766 
+1,27 238 692 142 1120 54 55 34 257 2 1472 715 335 524 2556 5 1599 39 21 411 5236 304 3 2563 9849 1276 28 12 101 7944 3822 134 18 1 
+3,111 152 610 1136 10 5321 122 2116 3290 9 1560 8538 529 209 3286 8 28 44 1510 10 46 85 5370 10129 14 3 72 4 652 2779 1 
+2,604 7869 0 110 4 0 5070 282 6 0 5349 3 110 4 8894 9 407 22 7 4208 4247 6 4712 9 324 22 7 834 2315 1 
+0,2392 25 9679 3 1868 104 338 2 154 0 56 7343 755 1490 16 785 5285 525 1577 2 1316 337 2 0 338 0 759 1263 3 1569 104 8640 4628 6 
+2,6642 236 22 7434 1008 39 1279 3 5591 520 487 2720 5 0 1013 1482 4 2116 28 2 777 0 88 128 3 66 73 224 44 379 535 336 1366 6 3 422 5 86 22 7 112 144 137 1 
+0,607 1043 953 97 25 265 96 8 16 220 697 20 15 281 15 1868 113 1027 967 8 28 18 0 1084 64 1505 11246 2 11006 9601 4 0 92 22 7 199 765 342 1 
+1,24 538 172 105 28 9 1034 38 1748 5533 5 0 642 683 39 21 1661 41 45 1791 1139 10 25 4998 0 1621 11 1621 10021 99 2336 1528 5 0 642 683 6 1599 2406 5038 225 46 619 2 9050 21 1469 150 1 
+3,638 17 11 702 304 8 2 0 1503 106 7 1733 8 81 5443 20 3 10539 5 67 96 6 2782 2010 8 0 41 45 221 23 637 1394 134 1 
+0,11 10246 175 84 4809 49 61 3 194 4 0 4601 1717 4 0 512 107 4594 2392 39 3 5114 47 9060 19 18 1 
+3,1636 10167 259 44 0 2559 4 3796 9339 8 3810 0 234 8 99 1816 19 13 2685 16 10263 8718 4 9664 11390 6 8829 10001 1 
+3,111 2116 1136 32 1148 3286 5 0 88 128 10 0 107 123 29 3666 529 22 1021 42 3484 148 29 2730 2703 529 22 556 1 
+0,842 5 0 88 128 32 4916 3297 5096 84 68 145 1844 15 657 15 5135 132 1672 4832 11 724 148 35 1456 55 231 2 990 71 3355 
+3,8271 1565 87 4268 5993 0 1387 927 4 7546 1213 4268 29 265 4198 5309 2 6542 6 1378 6538 5 0 43 1147 4 3 318 2 3291 169 5309 1 
+3,6388 14 0 3200 171 5885 750 1398 32 11319 2067 44 35 10120 2 1398 22 7 369 3 1820 18 81 1 
+2,1641 6278 10514 15 781 613 14 7182 401 389 39 295 0 7904 2 0 2966 5707 2465 316 403 1 
+3,638 17 11 702 9478 8 81 39 3352 0 1503 106 15 232 485 304 82 974 2134 274 4 652 301 23 0 88 128 1 
+3,533 428 4478 4 75 33 22 7 424 456 5 272 84 9138 1472 101 261 251 5 489 280 141 654 4 995 15 0 669 89 977 1445 134 18 1 
+2,11 575 1896 6111 993 1516 3 768 759 75 110 12 2151 7310 4262 4262 0 384 1784 3186 5 1582 22 7 2040 473 2 3 401 47 338 1160 50 531 104 360 
+2,11 1189 22 7 2593 2179 3478 13 8185 48 11138 1448 77 10245 188 0 2682 464 2342 635 22 7 11406 650 68 167 16 1411 925 597 0 6125 55 34 5 9 357 1329 1 
+0,27 5105 2764 788 2 5267 1727 9860 1663 2 2139 3475 6 5516 2621 5096 53 1477 10597 3 7525 4051 596 25 280 1 
+2,11 56 3640 79 2490 16 5324 2042 145 164 0 177 8 28 12 3 139 73 108 44 325 324 22 7 194 1324 112 144 1 
+3,80 93 747 855 747 13 5 1412 2111 2 1652 4012 57 4 0 199 56 1275 4 1090 2 2231 1256 6 8488 9484 187 2 1098 337 2 0 349 1 
+3,5942 314 704 239 3084 25 680 10 19 29 76 1127 15 281 67 15 302 40 47 165 8275 8200 54 30 8926 16 696 7986 2 556 16 2680 10591 2220 1988 1 
+2,74 11 8312 737 8 9 217 169 1008 310 414 948 8444 51 6 1131 11325 3146 0 9440 8222 75 110 87 7929 87 948 813 26 5518 4551 7086 87 78 928 795 1137 40 132 87 1 
+3,273 6680 4 529 1799 8 0 1825 4 0 2348 945 109 2 1017 0 3392 4 38 1136 46 148 29 646 15 232 15 115 104 406 2 2093 35 2116 556 211 0 152 610 134 54 3226 0 1696 2159 5005 49 529 29 2429 3554 9 876 556 243 0 318 84 965 1 
+1,27 5985 7575 996 543 4 10035 0 2729 4 7537 10194 76 2066 2 259 3228 647 1 
+2,4163 7 3534 72 59 4677 13 11043 6 5343 0 921 565 4 3 7416 7 6279 8597 1992 851 167 99 0 56 4961 0 357 306 3660 6 3 566 8963 3271 974 2 0 351 451 486 784 165 7 3 9237 10 7 90 10325 2 783 21 2446 6 1795 5252 11 357 8919 1068 32 2 2203 51 631 9877 1 
+0,469 923 10 22 7 1416 9612 426 54 651 22 69 4593 2 309 156 3970 13 1456 3 1341 10 3581 1372 4320 2489 1811 2 32 2582 99 0 49 10119 210 142 1697 30 3277 49 371 2 2493 1372 8059 1 
+0,7871 6053 2 1183 22 7 295 10444 622 32 90 4514 20 3574 9373 5 0 369 8377 4 183 5108 1 
+2,262 57 25 90 49 8906 83 212 467 9419 14 168 61 1175 3655 3886 3085 532 3085 1475 15 155 52 85 34 14 26 4880 28 110 14 4392 1903 1 
+2,2932 7774 1734 0 1648 132 28 110 22 7 137 14 4066 1903 6 9498 3 2731 3456 396 3119 0 3669 4 0 421 65 1 
+1,24 11 2497 44 0 1963 259 159 4 636 594 1998 149 329 25 5501 3 617 1118 8 57 4 10 1430 7 5052 788 9027 427 0 1287 4 324 7 1879 7888 4173 5 3586 8398 6 9801 10999 4 0 7033 6260 1 
+1,17 313 1115 3 2508 715 8 81 83 238 92 2478 336 1878 2309 54 13 1339 5 0 88 128 9 8789 1808 8 7953 21 9491 317 18 1 
+1,4272 17 413 689 9459 9599 14 101 152 9 3 63 106 8 81 975 180 3 718 9477 6 785 5507 6 18 401 622 349 8 7376 7 1488 3108 1071 3267 1 
+1,1391 5 6827 4755 6963 5233 5 0 359 54 30 1525 0 626 4 9707 4617 1760 1 
+0,27 79 5294 0 5307 4 3 229 4 4372 3825 2550 32 280 2669 4 35 5646 47 266 806 1 
+3,7716 523 3059 2386 3065 25 720 3 63 158 207 44 2778 0 386 19 181 44 0 398 1291 0 47 132 11 111 1605 206 229 84 13 8629 10369 0 3696 264 735 
+0,2950 29 672 973 2 190 4413 6 2049 1409 1363 873 1 
+0,470 379 2950 29 672 973 2 190 4413 6 2049 1409 1363 873 1 
+1,7395 7192 7049 1612 245 5169 7881 1011 5196 10463 2 11427 164 189 5178 5472 1 
+0,953 25 10750 96 290 0 1422 15 19 3530 2 167 291 38 1794 1473 416 40 8181 11 4062 910 518 25 1137 1084 4 44 742 1016 3175 310 96 9 953 22 7 1246 264 
+1,260 946 244 2930 4318 18 26 127 771 76 318 2 4722 444 2 298 5 681 315 42 29 672 2222 4472 1 
+1,56 6 269 444 1670 1006 5 253 8 81 6 269 212 6597 1412 2 601 460 3751 4 0 484 179 22 7 2330 1579 4299 2491 225 5009 5072 2 1300 480 1385 334 432 32 90 9422 51 9 528 1 
+2,6329 1078 1038 6826 22 7 105 9 3 314 2711 14 4007 1194 140 32 11095 0 2483 259 12 1440 414 6586 1 
+1,24 260 946 2397 2930 4318 11141 3 10850 4725 1681 1748 531 161 3265 94 23 0 1556 681 298 233 1030 0 576 107 2 277 49 2 200 19 3596 9 0 4909 2157 2 626 168 1 
+3,11 580 417 2765 1221 5 313 550 278 4509 195 15 4321 4153 1819 2 8887 3 524 7 1719 105 12 3 1570 153 9445 645 9 4361 1 
+0,2883 428 890 53 0 386 4 407 22 7 5931 66 915 75 6246 3680 0 646 140 32 90 1333 7002 1859 1 
+2,273 17 7740 6479 5793 324 7 182 11034 50 2738 670 25 3372 192 0 3420 4433 12 26 2735 5 26 43 283 1 
+3,4321 4153 1364 832 58 1759 60 3 1570 153 645 9 4361 2326 8 81 4583 3 524 488 20 7069 1206 1364 832 2 1253 0 92 22 7 580 696 1 
+1,125 117 40 10265 1942 3888 127 225 5230 13 1111 6 0 4596 184 5592 8627 3401 8 0 390 7 5108 19 13 3 4537 1493 54 1460 53 264 803 1529 3696 1 
+3,6666 58 5471 22 440 849 206 1259 62 4659 20 8875 2100 8225 6176 5968 555 21 1469 1440 81 102 3 238 996 5 1039 22 7 1417 444 5283 23 16 1422 1 
+1,428 935 228 31 879 3541 5213 5 16 814 4 77 56 146 543 4 5493 6 3415 6 0 88 128 85 9084 141 3401 8 119 65 22 7 5551 3 56 317 18 8 81 1 
+1,17 1961 5312 4 566 10809 9832 12 8347 8549 4679 30 1759 60 5 3 10916 9462 6 0 10774 5139 5 10984 9313 2577 21 9554 2 3 8725 2448 304 20 26 9929 1 
+1,2908 2837 2979 2836 942 2377 1989 2349 7 8462 31 1489 977 81 5 0 43 41 45 388 5510 248 379 1407 686 8108 14 3 5207 885 225 26 5046 4582 0 800 15 2780 1 
+0,80 93 1173 1065 13 4427 734 2 1226 16 2138 5 257 449 896 6 19 22 7 405 505 38 395 6 38 2428 2 277 19 1 
+1,4181 8937 7769 22 7 2246 6 5412 30 8893 10 160 13 1285 46 68 277 2 154 180 691 28 26 5412 6991 5959 31 1457 2 4661 0 306 4 26 719 769 5335 
+1,1956 4 8876 5545 64 10911 14 6643 1875 81 39 260 1165 3166 5325 4 1730 213 4 1539 3625 6 2222 9418 1 
+1,27 229 975 787 87 11 1374 1169 5 100 87 18 2928 793 3744 16 170 161 196 71 0 1521 4 3 1279 2333 2614 68 22 69 34 4935 1 
+2,74 17 1371 59 8870 336 2391 6422 29 2093 26 112 144 6 34 9092 23 0 252 39 1224 1442 1136 0 465 59 1359 2921 18 81 1 
+0,607 1915 745 1173 1065 58 18 19 25 473 2 1652 1389 2874 9 460 104 5 202 6 437 2 3295 175 2691 2 1459 982 6 1766 322 1412 156 7370 399 
+0,268 822 250 1173 1065 109 172 3 38 3920 7436 1920 1881 1888 6174 9 7052 1881 394 1 
+0,830 698 5 3 337 63 310 1178 1315 1112 596 123 1 
+3,2824 352 250 11 2394 4352 468 109 5287 1512 2 2373 1164 22 649 2 4985 3 4867 8 99 2038 5482 1115 164 483 1237 525 4882 122 12 16 621 1791 1230 6 8595 0 
+3,88 1164 123 19 362 29 157 2064 9 4807 5172 105 3 592 10 85 34 0 199 417 4710 20 3 56 62 6 55 231 2 3 11056 3350 8201 10565 0 10625 6 5064 9160 4 0 2252 1 
+0,4375 9657 1275 2950 6 7152 32 1135 60 381 6 809 264 10039 567 854 1499 12 2606 4373 44 1913 1406 563 1 
+0,27 1155 1835 25 1822 3 239 899 3187 3 4861 285 405 3 2223 10 55 520 3 38 2124 4 1789 4752 242 148 1 
+2,56 4954 7488 2909 555 26 59 403 70 39 256 3142 5 0 2723 131 2744 4 0 5635 6582 1942 11428 653 20 6544 7796 4 2343 1 
+0,17 751 712 8 81 720 2 16 434 729 21 11243 5194 23 16 620 12 116 7 882 84 1018 3571 4851 76 9 120 19 1804 42 9 16 11121 118 132 0 156 208 705 7 281 1215 586 183 394 1 
+2,24 7655 5261 1054 852 13 99 42 3177 76 2 521 136 110 102 176 1514 1399 829 2660 16 283 213 4 4657 1703 5 0 676 7 1216 800 2 0 4164 353 10 54 2087 3062 25 8112 2 0 4164 9 3 10304 926 8791 5571 6 10466 4 852 3290 1 
+1,24 149 329 1108 2 4611 4766 4211 6 1208 2 1549 817 52 55 1127 478 293 2671 588 23 0 1067 1912 61 101 152 25 1 
+1,24 606 1271 30 3553 9 3 4648 1669 8 1735 2638 5 0 3998 2675 714 3 38 1125 4 4655 197 2195 8626 6 10059 42 11315 2556 443 2 777 0 622 7 679 4 5210 3 49 2646 5487 1 
+0,24 469 4 1448 77 9676 5 0 88 128 6 217 57 158 4 0 390 7 3586 4652 614 1500 10 94 284 3429 71 1220 4791 2565 3171 160 1 
+2,80 93 2843 8562 6016 7568 18 28 52 85 1651 107 5019 8 10895 1152 155 0 797 22 8150 10452 2 1079 0 545 1136 84 145 9 48 49 103 1 
+0,17 11 3825 5923 4286 13 243 1028 23 10070 67 6 883 4845 6 2403 2544 5247 3737 0 41 98 6369 7362 18 8 81 1 
+3,2373 1164 18 19 85 504 8454 3 169 4867 9 9616 3 2227 14 16 8045 2477 5 8827 1 
+2,74 1905 103 380 5 1400 39 0 56 9353 89 2578 144 513 14 0 177 9 0 43 72 5 1162 103 9143 267 2853 6158 31 5367 14 3 5456 4843 10 0 11203 502 89 3144 11 288 89 2106 23 1533 2 384 5 0 92 5 57 47 0 146 23 1533 2 1453 1 
+0,1185 1163 3877 2990 23 2823 195 2601 46 5013 2 5238 365 5313 23 0 3488 4 6029 6976 21 3118 7 2575 6 2475 304 44 0 2348 1041 1745 391 7439 7122 11 11225 653 304 8088 248 5655 25 795 1015 4868 9 245 4 0 365 901 14 0 653 123 2823 851 5013 2 5238 99 762 8184 2 167 40 660 5284 18 1641 7903 2823 10842 1 
+3,3885 121 4951 180 210 1207 832 3889 2390 3867 18 5 951 956 10 26 62 25 8954 1866 104 9 3 1077 190 1755 2240 5 171 1 
+1,273 6713 2 3535 1267 20 677 41 45 388 1271 5 100 6 14 0 2381 31 3 257 4853 5 0 1968 1522 14 1871 2316 1504 42 160 13 141 654 46 2115 245 9904 21 1469 1787 3211 1 
+1,836 100 100 7 182 1295 540 480 1305 168 23 407 8 70 2 154 3163 21 157 2 217 77 528 4 443 5 253 6 13 975 8 26 4898 2 1116 180 5 3 5079 2 10416 0 484 179 26 10841 6 2800 18 1179 3836 1579 6684 334 4306 626 506 15 1743 443 10149 5 253 7 1191 635 1 
+1,5926 11383 5302 369 1249 3685 883 6 10584 51 2039 4 2034 5 1584 8 81 188 3 996 454 358 2 1017 3 736 4930 475 10 709 323 94 6 1861 888 14 21 669 718 1302 1 
+2,7615 1358 555 3 77 145 231 310 3937 6779 1260 6 10904 48 1013 340 2451 15 111 332 1025 0 952 2 35 1013 699 386 139 186 81 1 
+0,6738 58 21 385 193 62 5 3827 105 2 592 16 2588 2 6583 5 838 11 10482 29 1225 5644 588 2 4229 947 7 2292 1 
+0,1199 3341 178 609 527 68 34 2230 8 969 4013 30 2639 2 154 10298 922 6 5103 1 
+0,6580 1483 2 1519 5611 817 2625 3548 1554 3088 2 10065 1 
+0,11 4220 7339 13 549 2 32 57 1696 1794 320 0 452 9 446 6 2109 1 
+0,11 470 240 13 2524 3 478 891 40 47 4 10623 16 41 45 300 6 478 2222 563 6 4789 2 309 2586 1811 2 32 181 3 3255 187 2 0 126 3920 4282 2923 1 
+0,11 88 749 164 16 465 330 3037 468 6696 29 10036 0 43 578 844 6483 1041 8 136 5 21 773 2 3285 936 371 50 0 1420 4 1906 6 330 3037 844 6397 6 4810 16 3742 309 5 0 233 1 
+0,129 7243 6499 3843 3856 24 844 239 130 97 1187 6 0 2100 6 2048 10254 1402 1408 58 1 
+1,27 11081 1147 31 1148 109 5 0 1797 2 1456 3419 717 5 0 869 1841 6222 233 4 315 102 0 2825 717 349 11220 2431 3 133 1125 2040 3104 1 
+3,849 6298 683 2408 4334 1063 1938 840 18 8 70 10 19 13 5 349 12 56 206 5920 832 5899 98 1063 1938 840 50 3 719 338 5 518 12 0 327 696 22 7 2150 9 
+0,6764 22 7 107 8 81 172 3 56 139 104 6765 5692 104 318 2 672 513 156 854 5 902 4624 290 0 2083 1 
+2,165 209 241 2813 3771 909 9 7773 852 3928 2 4826 189 668 5 0 8081 4 5912 551 852 3928 3230 0 1248 518 
+1,1630 5 1182 114 399 239 4339 6179 337 819 293 8 35 143 1085 1301 73 5689 9 0 62 1 
+1,5754 204 352 250 8923 2388 1107 473 70 2 5164 5 717 349 12 0 576 107 256 304 5 0 2986 359 4 2825 39 1080 243 1126 2 4747 6 1651 8608 2 3165 5 0 127 22 7 
+0,117 21 857 2 1824 0 448 11114 4 367 5322 3389 11097 6 0 10307 4 5031 440 519 0 2934 945 6 150 893 2541 134 32 9553 120 3610 2 34 0 199 8925 1032 83 2197 249 358 10166 6 135 156 8588 4467 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 1855 91 37 326 36 37 572 36 1 
+0,11 649 13 3 1138 9 0 48 275 6 35 318 2 1652 2869 3 481 1021 663 303 1 
+2,6812 1876 4 5880 5596 0 5509 12 3 589 2212 70 5167 49 61 779 4 0 482 5 0 131 3423 6 4792 6936 6624 4 325 14 0 1248 518 1 
+0,2332 116 0 2985 2982 1640 2041 3 10104 197 2982 5620 116 5 35 615 83 130 19 13 3472 2 4828 0 1850 5363 197 0 48 3633 6 1024 155 0 5457 10 3746 564 29 235 210 1 
+3,7369 23 321 1521 1107 32 2145 537 1924 2 1795 3 921 758 10 1315 3 738 1676 3277 3 3914 1317 861 249 10414 10 19 1460 8041 134 4 0 4479 148 1 
+0,273 27 1155 1835 25 1822 3 239 899 3187 3 4861 285 405 3 2223 10 55 520 3 38 2124 4 1789 4752 1 
+0,149 2274 0 3125 537 4 0 56 13 302 2 694 8 86 2039 4 5045 83 1657 2197 879 3 363 257 581 83 0 10255 290 0 56 1 
+1,2878 6203 2294 654 3691 20 650 511 5 0 1050 937 1313 761 507 387 32 90 4654 3546 18 5 3 228 2564 296 70 118 48 276 132 2085 2181 13 2 504 3012 18 46 89 280 8645 5 2294 982 5320 2607 2 2047 1152 31 3091 1 
+3,1894 3 5577 194 158 383 218 10 1025 16 400 10208 822 113 81 10252 1506 4279 6417 58 1 
+2,74 17 7454 7536 6 6358 7433 355 1952 7 6774 5984 6 7263 6651 1569 819 1569 818 8 70 2 4491 1171 7 43 146 7 59 1671 2789 112 144 1 
+0,2981 2887 29 247 3 4944 8571 12 5134 466 2001 897 6 12 16 621 2256 194 1104 344 392 382 119 47 1 
+0,1406 2032 11269 7862 25 1512 8061 255 5628 104 23 2240 8408 6 13 590 2 6876 324 6 496 2 7978 21 502 11133 449 904 1 
+1,125 375 25 3 864 9 905 718 9717 102 817 222 2 0 2131 5 0 127 7 43 1294 7725 594 237 0 2419 7 4820 2 0 390 18 1 
+2,2878 17 3012 5 0 2697 507 83 41 45 316 285 1050 937 30 3555 0 5546 4 2294 654 4693 2 0 650 7 507 595 301 1071 2 32 90 9804 6 387 32 2 34 5479 60 1 
+0,2279 7626 40 65 21 1490 4 16 1872 8641 663 114 3095 2 0 671 518 3 63 565 12 49 10629 6 1412 9259 1 
+3,130 8 70 172 0 247 2 3445 4 130 7220 4138 2954 741 6 2954 741 7925 3958 3 38 2534 10 0 62 2140 358 2954 741 2899 1 
+2,6414 6610 2808 4 205 8504 3 9457 8997 5307 2 457 3 112 144 274 4 1453 422 9178 6767 7323 4 325 1 
+1,74 205 1614 539 181 189 151 1085 5 1853 4446 5243 8020 2 0 174 1269 131 1876 747 89 0 1667 4 1244 5 0 3653 9 99 0 1862 1517 1 
+3,2310 1617 408 18 8 70 19 29 3335 9794 3376 3604 5 1867 161 6464 2473 48 5476 4 84 64 3641 5 1039 1 
+0,1949 2189 6 2169 1751 5 0 63 158 12 363 439 805 206 8 500 71 0 178 307 163 448 877 61 35 2420 4672 187 2 308 303 2910 58 1 
+0,56 8032 32 10373 0 887 4 339 94 54 2436 2771 888 4 350 4 2196 2107 6 135 4665 3449 44 0 156 1583 537 149 2274 123 1 
+0,17 830 58 1893 236 0 92 7 199 470 342 55 694 21 4836 2181 4 16 381 1096 518 5 0 119 678 528 3 725 240 2438 18 8 70 1 
+3,7383 7684 0 194 199 56 10618 1689 18 8 70 19 1141 3 338 2 2185 7298 58 7249 98 1063 1938 840 8930 5 16 56 2664 2540 3 821 47 1520 12 6154 6156 408 1 
+1,1184 11 411 1977 13 192 2 3714 3 38 5580 1087 2 167 243 0 9992 4 1621 3 791 10899 1423 27 218 5 0 6962 3238 8 70 18 10 0 388 84 25 
+3,2373 1164 97 1247 2755 83 2394 4352 468 5 0 56 2886 1891 9 0 2886 4 835 1972 10 2394 9570 3452 10 29 200 19 49 3256 9 0 1076 2 425 5482 1 
+0,11 1036 378 129 3919 3815 
+1,80 1892 204 352 6848 25 1509 2 2089 141 3659 11254 2 2047 0 3596 626 4 0 77 952 54 64 1148 3382 5 100 1 
+0,2875 29 34 1209 2 871 2058 8512 3126 6 372 9579 9941 446 2109 365 5499 87 2150 2 2795 87 399 15 566 15 854 710 440 519 8157 1 
+1,80 93 1198 9005 21 11224 1792 83 1735 3711 70 5 3 9803 10 8657 12 7850 101 6277 1589 7 5445 2010 50 1735 1021 6 0 2646 279 0 622 29 452 14 119 65 7 420 1432 27 1787 181 51 5041 4 3213 1437 1664 2675 9685 10 860 9 3 4648 8066 8221 398 1529 2638 6 10022 894 10417 4 245 3647 9 1735 3188 3711 1 
+0,941 13 5538 545 563 6 175 522 2 2041 49 5077 300 1 
+0,11 2307 10660 1429 114 6 873 15 368 4 21 890 53 1111 2187 4 4665 2107 365 6 446 44 21 156 2123 2 2123 392 1583 537 149 2274 172 70 1 
+3,7872 1042 539 111 1649 22 7 5878 465 1875 690 5761 7126 86 720 546 3257 578 207 449 8374 20 3 2212 5 3500 2774 6 18 19 2051 2 1771 143 8621 9 0 741 1728 47 383 1 
+1,24 606 4258 2382 6661 4 2362 31 3510 70 2 781 0 1912 4054 1359 1214 3 3376 237 47 1701 44 230 2 1426 60 3 216 1490 4 0 390 7 1113 163 1 
+3,27 696 5 2847 25 4720 56 647 10 19 9693 450 9 238 269 294 3027 2920 6724 5 3 679 18 19 25 10922 3343 587 8035 2706 2 0 615 83 1111 11151 1 
+3,1398 1164 105 2 479 51 2 3770 2280 174 1303 5732 5690 2674 1160 50 289 1868 153 0 1076 18 70 1 
+0,7523 4 910 525 560 289 289 113 9308 5 0 63 158 2 351 3775 153 15 984 9 4048 7737 3575 347 39 3 77 47 8921 143 308 303 2910 18 5 3 679 1 
+3,353 3 162 174 1302 10 3579 2693 96 2 38 1109 0 439 4 3 1072 10764 9 0 384 106 2626 0 370 169 293 61 47 380 904 1 
+2,24 2942 1408 29 453 14 2151 9 0 111 332 1907 40 65 1658 26 438 12 5306 2894 2359 9 0 3570 283 891 2496 1 
+1,24 6880 3533 2289 7 3046 1891 8 70 2 11262 3 895 228 649 10927 3585 4003 3833 3009 4 3391 23 2688 595 0 238 4737 284 279 5032 9 421 785 1021 4403 1 
+1,11 3556 2707 3680 10 0 152 581 84 25 3579 18 19 25 141 1840 2 0 7697 8300 5555 229 13 4785 2 396 0 660 291 19 1 
+0,7095 44 10728 697 160 13 3 38 2724 9655 8 0 9439 4302 8414 22 7 7594 623 25 90 788 248 0 5652 22 7 42 0 2687 10 200 13 8166 436 200 19 658 2 235 12 1 
+2,461 1067 331 1785 4071 7435 554 3 79 4 9963 23 2961 467 2843 5 3 77 784 1276 70 12 0 1271 4 0 1405 2843 1954 3834 1 
+2,17 944 1620 5481 48 2596 6 989 5 133 675 2 616 0 828 471 331 2 21 351 139 108 44 0 849 827 1049 5 0 126 467 70 1 
+2,2409 17 944 1620 5481 48 2596 6 989 5 133 675 2 616 0 828 471 331 2 21 351 139 108 44 0 849 827 1049 5 0 126 467 70 1 
+1,125 171 25 680 1398 134 83 3727 1401 333 39 3 798 6 11240 2776 20 0 179 150 7 38 294 118 528 132 52 209 489 10909 1840 12 1355 1 
+1,744 17 100 7 176 1200 1268 480 1179 3836 1579 334 4306 1030 26 2756 8666 8 253 8 86 76 2 1242 0 8232 484 179 631 52 970 3 677 3106 18 1 
+1,125 7224 64 215 1984 8 0 1898 1859 2 3649 136 7 131 4 0 11171 4689 630 39 3243 83 8058 0 9190 2176 5 3 645 2 355 0 5249 10 25 5082 40 11334 51 653 9 119 162 7 6688 1885 7792 5 536 1 
+1,125 1061 1702 8123 726 605 170 226 292 2 2047 46 2200 335 0 176 401 5 0 59 146 7 482 3379 438 1 
+3,1398 1164 690 496 22 7 182 5220 1076 13 3147 38 2674 29 154 19 3439 5166 23 2895 6 6090 1193 1165 690 1 
+0,11 610 1115 3 218 3398 10 3293 4 4565 4740 6 135 1257 5506 2574 64 0 241 362 3311 9 483 1559 1 
+1,24 11 3560 4 1469 817 5 111 2361 32 2582 20 217 323 161 5 0 75 169 389 3 5211 3549 101 152 6 524 149 329 2 0 150 333 6 333 1 
+1,6947 1366 27 56 388 10221 543 5 0 1871 2316 1504 1968 1522 89 0 647 83 189 2156 109 15 3 192 4 5207 2590 5606 51 14 21 126 1419 5 1366 1 
+3,7619 25 215 0 974 3076 965 4148 58 0 786 239 10 4809 6114 1879 7 2523 5 1069 13 337 2 4626 3 338 2 425 16 846 541 840 6 1920 1754 4460 2 2849 58 9 50 570 104 6387 3410 29 34 478 82 9 19 1 
+3,6217 537 1924 97 25 5598 3 3914 1694 10 1315 3 738 1676 3277 11386 164 179 917 39 1521 8025 8579 0 62 18 109 1 
+3,6404 2662 5 833 9 2015 9809 3363 2700 0 580 1733 5 133 389 111 168 191 9836 10756 187 2 3 2119 4 723 
+0,2915 22 7 6815 2375 25 866 0 4835 4 3 2933 2734 899 10029 3 8889 285 0 43 57 1822 20 3 392 4 2444 405 5468 141 9688 61 0 2658 121 68 479 5 1323 1 
+0,1384 61 460 1539 14 0 945 4 7924 6 7296 32 90 8866 9 4779 10235 14 235 1 
+0,3904 1948 466 1096 239 3887 746 29 516 0 92 22 7 10768 481 1680 12 3 11446 3424 3887 22 7 10101 2734 3969 1680 13 281 2194 61 716 15 9 0 43 72 19 1337 3 4579 3424 1 
+1,836 100 17 27 5109 475 8 3 2666 3468 5 0 1029 4 6846 8 86 709 14 214 352 94 15 100 7 182 9544 1200 1268 480 1108 2 0 5124 484 179 4 253 2 495 2 157 3 4515 77 65 2238 1 
+1,74 205 429 302 0 41 45 146 7 316 79 89 16 1742 585 5 3 10360 137 83 1952 8 86 1 
+0,3964 1525 293 142 9 35 448 393 4 344 1090 140 34 5017 5 38 2045 193 10 4186 97 13 8810 5 16 3638 190 395 1 
+3,80 93 17 2849 58 37 27 376 305 373 341 397 91 410 415 402 2840 98 286 223 423 418 36 2840 98 37 27 36 57 4 0 199 41 45 2566 1468 1767 8 86 18 19 29 479 16 657 2194 8576 846 541 840 1920 408 3446 0 580 2765 5 0 3120 663 264 40 47 1 
+2,11 1343 47 266 238 536 613 31 1847 14 3 490 508 1994 2 21 157 7917 22 162 174 208 9 3 5436 2 6240 539 1 
+0,130 97 13 916 2 10378 51 1237 4 16 791 142 527 9 0 178 609 307 163 42 967 148 46 184 1099 0 62 2 592 14 3 10473 2737 2117 2 1220 2245 4922 1 
+2,407 7 1356 6306 6 4304 6658 108 1995 5 3 3690 131 5706 282 8 0 7541 4015 1 
+0,4366 1396 1901 7072 1640 1156 208 2046 6 208 705 1768 57 793 43 1156 0 208 705 1280 639 30 48 9261 1439 5519 4 208 705 122 9006 729 8857 6 10829 208 2046 3035 8858 8752 30 4555 20 3 1493 5078 9559 301 50 11359 2365 1 
+0,2893 4073 156 5874 7489 462 961 2332 0 2893 4073 475 9370 195 71 13 19 118 10901 51 5314 7946 4085 25 1018 1697 4 3 216 475 8 0 2609 195 4085 25 90 3740 2696 15 595 10 160 85 34 3 216 367 475 83 411 1 
+3,80 93 17 6478 2358 37 27 376 305 373 341 397 91 410 415 402 3974 98 286 223 423 418 36 3974 98 37 27 36 18 8 86 16 5111 3409 71 10535 5193 1106 14 21 8087 904 4 323 268 113 5 833 3206 12 3 820 186 113 708 5 688 1 
+0,17 936 2558 4781 667 4 10340 10018 11368 5 21 417 2505 208 9 599 30 1838 2 10952 0 3214 7 3710 10511 3279 3 727 712 680 8 70 1 
+0,953 22 7 911 4 0 4527 9195 143 1756 531 268 113 5 0 63 158 187 2 143 308 303 4972 97 1 
+2,2364 22 7 2918 1900 2914 124 5 151 81 21 59 403 5 0 5639 282 39 48 656 8160 5 1400 6 826 1 
+2,2973 2462 1201 7019 1987 957 2422 6 4075 3990 4033 55 9229 57 4 0 182 3239 454 10117 5 0 3851 4228 1 
+1,17 269 5035 4 321 1717 4 3 4097 62 18 5 3 372 679 8 86 46 85 247 0 8411 1014 35 3294 4941 1287 5 100 1348 5851 921 147 1 
+3,80 93 430 823 111 3016 2849 58 2840 7125 3 295 56 206 2566 1468 3120 663 62 25 2164 473 2 479 846 541 840 855 1920 941 3 927 4 4148 58 9 5698 104 5 3 437 6 202 
+0,111 3969 3463 1467 0 5475 565 1032 2562 3 38 4579 3424 1 
+0,24 577 5453 4 5068 1838 4691 5 3 393 4 633 3222 3 41 98 354 4452 9 21 3096 297 825 104 5 4472 86 2 2139 0 11270 23 1672 3 585 5321 5185 1 
+3,1190 1555 2661 148 11142 8399 667 4 987 4 8314 416 2 0 906 3155 4 2936 1 
+2,1921 88 772 53 0 1885 467 229 2124 12 3 3203 170 226 224 44 6280 6014 14 1191 3060 8 70 1 
+0,17 953 97 6694 236 8 86 18 19 25 3717 0 567 1283 1229 4 16 6099 518 4 5138 220 697 2 1413 3162 392 854 12 2075 142 1 
+1,1905 146 543 4 900 2 2618 942 7 3987 2861 32 90 5161 20 2379 7 267 1 
+1,11 3046 1891 5 2289 25 1309 10 0 238 4737 537 3009 284 32 26 3391 23 2688 5275 27 2088 31 1678 20 3573 4 1657 1851 4 0 388 10457 7219 6170 1 
+2,24 6559 2345 7893 1141 86 12 0 111 332 2387 54 1099 0 11247 513 3105 2 2487 9 0 262 73 891 12 1054 6314 1 
+3,80 93 17 6553 4222 37 27 376 305 373 341 397 91 410 415 402 3999 2347 286 223 423 418 36 3999 2347 37 27 36 25 473 2 247 99 8511 2209 4 16 990 2 3618 3 2088 10 543 19 4 11403 5125 385 50 0 4448 7287 0 111 332 1583 537 7 489 18 8 86 1 
+1,80 93 6005 4312 238 4488 8747 6 545 9158 2345 3980 68 1654 255 403 2 26 5292 1693 285 691 70 3 6786 7422 10335 866 3980 13 2561 5 4312 2783 4161 7354 372 192 2 1085 5 2642 1948 1 
+3,571 1880 2956 2885 523 6295 537 147 3 820 113 627 5 1728 63 158 383 1258 20 293 191 6 895 647 1 
+1,431 946 244 1349 1897 123 3 56 890 25 1802 8050 10 431 1551 2215 944 6655 31 7975 310 5 56 3234 5 375 6 942 1 
+1,125 7400 1374 480 1871 6616 334 7009 502 4730 5 1378 8 21 9108 2165 23 0 88 128 31 1978 243 2760 4 8564 71 2137 1836 8003 601 407 1 
+0,17 4327 7 112 144 457 173 9637 7665 6668 10415 52 68 3421 278 293 42 242 148 52 6 135 1219 64 502 337 2 0 3429 4 120 46 68 7993 1 
+1,17 313 7 5073 143 9 41 45 8240 10600 20 5075 4680 2037 458 13 362 2 1019 2007 9 0 1020 4 40 47 41 45 1479 240 134 18 8 86 1 
+3,11 2400 67 62 25 90 2067 3 73 139 153 1988 9 8960 3207 5 0 7168 8796 20 0 107 4 4184 1 
+0,17 429 368 4 3 988 2 4475 49 288 53 0 9792 4771 372 137 240 3 522 9 8145 137 1089 14 1399 7045 254 29 488 3 288 241 10636 3488 18 8 86 1 
+0,830 25 3717 16 4226 518 12 0 4415 4 0 4226 5714 3 822 2679 3444 9302 9700 66 2764 1 
+0,125 56 476 1978 14 214 5627 2217 6 1247 5634 3229 2489 248 688 73 5 3 4681 8 1850 9236 4 367 995 1583 537 149 2274 18 1 
+0,1057 11 88 6835 7 1070 2244 229 0 5795 7654 1037 25 280 10 2489 9754 83 3 130 1055 1187 712 758 10 10884 10 500 31 49 4839 61 178 64 11285 1 
+0,1057 574 1619 574 25 18 19 29 3550 7841 3 3536 304 303 10 10327 1232 5011 114 2 11251 301 1316 1 
+2,1375 2860 4 1354 1101 384 5 0 1455 4 0 570 364 9458 195 1279 60 8 3 144 6429 7531 4 0 3945 1393 124 0 112 144 1 
+3,11 439 4 3 1072 4 67 5 1378 25 2516 2454 531 15 704 96 32 867 2 10757 11 370 4 5999 509 5 1378 442 2 3 77 65 363 4 22 1 
+0,607 1043 1880 2956 4347 18 16 63 158 386 3741 23 3 47 991 8 293 1234 2427 1433 1906 1 
+2,74 6317 4218 4 0 88 128 2490 3 9378 11175 4 483 174 1269 2770 102 52 8724 0 482 12 26 714 3421 5 86 22 7 131 2 108 59 112 1 
+2,6319 3902 17 6797 2181 1673 5 0 1050 937 2697 507 8 95 102 888 4 623 2617 9165 60 3 5241 2 154 1438 155 46 68 5366 5 9633 5 3 814 2082 282 1529 6 4577 1 
+3,7039 855 408 1722 3 4783 1028 23 782 1266 6 2122 22 1761 39 3 56 596 3679 16 7859 8140 872 4568 3 4928 1306 61 524 9842 1 
+0,17 6423 4 3 38 1900 4185 3 3507 10 4545 3317 1859 3502 30 11234 583 1333 1780 2224 84 1782 9401 0 2514 1859 4445 0 379 7043 2992 18 86 1 
+1,6568 771 76 10510 245 7761 1852 5 2413 165 13 3 1852 49 23 6371 1884 5 7878 6286 379 1 
+2,429 0 56 146 22 7 316 79 9736 53 0 59 144 346 0 878 25 90 8 0 79 22 7 9765 796 2190 1 
+0,7423 8078 140 34 2167 9 3418 217 352 113 4 44 9184 5319 599 1537 3171 82 56 8530 3 596 296 8 86 3680 1 
+1,3996 2858 11 411 1977 8798 739 692 3384 188 21 9523 5 0 1145 1599 2406 1029 4 7403 8 86 692 142 1316 6 2800 18 1 
+3,6849 4184 22 7 1947 25 2121 3 2708 8143 2400 22 7 2986 927 2 556 73 139 153 5 3207 2 10014 4624 9 1500 3 1947 1820 18 1 
+2,7448 6572 3 48 72 56 330 336 3365 0 43 346 231 5 0 2841 330 5 7176 6349 1366 12 3 115 243 1788 2261 310 2868 2975 2075 26 603 753 535 3179 20 3332 57 2752 122 8 86 1 
+2,1597 4387 6 56 330 3743 2352 2323 2021 2 3 115 73 115 186 224 44 219 6922 8 86 2 3101 2 0 10362 4 0 2408 3075 535 1 
+1,11 88 128 355 1171 66 73 5 876 72 2 108 0 288 7 59 880 630 1 
+1,845 288 12 8337 2472 30 517 362 2 32 35 11194 3510 51 20 3599 148 511 1 
+1,2908 2837 2979 2836 942 204 823 250 117 3 1443 815 10 5361 650 1761 3 4395 5190 543 4 8698 1837 10312 8127 8 86 2 10499 2069 132 3 56 
+2,2883 7120 22 7 1201 6730 349 12 1404 961 1317 2009 6133 5998 50 0 258 306 5 1331 6 482 14 0 151 177 1 
+0,11 56 107 1847 105 8 86 9 3 10548 220 206 522 405 1292 385 2 3388 1213 5166 54 140 5195 3 1028 2 474 2774 1 
+1,24 2289 7 3046 1891 2206 4003 3833 3009 4 3391 23 2688 86 5 3 1807 10 10551 1746 4 26 4897 10 52 387 2200 814 8 647 4 785 1021 4403 188 26 561 1 
+1,6272 32 5401 3 10423 4 0 131 1825 4 9944 409 2753 5829 6256 1 
+1,80 93 101 152 6 1946 149 329 8325 2 0 11399 4 557 3450 149 2970 8 86 15 0 267 8989 0 606 5333 7 894 615 83 357 450 730 999 1107 4429 5125 1069 6 0 636 5133 10645 3 1694 10 2561 2970 1 
+1,1636 4343 5317 209 3 1443 11209 75 110 102 19 2538 3 257 5600 5 0 369 890 25 3259 348 12 180 3 1001 4 11354 9531 54 10978 8215 21 1657 3223 857 5 67 2171 3965 4014 1 
+3,1388 97 101 1886 4218 8 28 18 7297 58 300 32 477 49 3203 12 0 2689 4 3 9855 197 0 48 114 1250 278 15 0 2147 11150 8189 3 2017 1807 23 3 6257 228 1 
+2,74 324 22 7 112 144 2738 797 651 22 69 118 148 9322 46 148 9398 87 11 923 78 3436 87 10538 1075 6794 6472 18 87 13 10 7729 6 7880 2720 1 
+3,111 6258 430 823 11 1533 106 4 0 5514 454 8 86 1311 459 49 5514 256 1978 243 0 6377 1397 6971 1872 6325 5 0 6077 1 
+3,6307 7507 0 5428 10 2016 0 87 7593 22 22 4875 18 109 10 6646 6606 85 1147 198 15 612 307 1781 1 
+1,4004 1131 9 0 10289 12 0 6648 7788 730 306 5329 10301 99 0 512 150 6 1002 3167 5 3557 5 7798 237 1 
+2,11 111 332 6780 6 2151 2285 7294 30 590 2 9173 3 768 3316 20 119 70 1 
+2,17 3966 3949 7 77 145 997 5 0 63 340 1258 0 2374 2834 1019 176 4 0 126 467 1035 642 12 3 374 186 108 44 0 3568 1354 1934 86 1 
+2,80 93 17 3966 3949 7 77 145 997 5 0 63 340 1258 0 2374 2834 1019 176 4 0 126 467 1035 642 12 3 374 186 108 44 0 3568 1354 1934 86 1 
+2,1869 1054 6615 246 21 3013 1143 6 826 5424 16 231 5 0 1618 1899 20 457 16 384 5 3 2715 115 2 186 44 3903 1 
+1,125 27 229 975 787 0 7564 1374 1169 4273 29 247 321 4968 19 25 90 1259 9 49 61 3 162 15 1144 15 35 4097 62 123 19 29 141 1003 3483 5 100 0 4273 172 1 
+1,836 100 17 2388 269 480 1383 334 432 8 95 2115 26 146 1265 253 7 2330 1579 3468 2 9699 198 35 2797 6 1116 350 4 1200 1268 3513 796 0 628 1 
+2,171 22 7 6600 6785 3282 124 0 288 22 7 170 364 10853 75 110 6 7931 7064 181 19 3 73 66 1248 9 0 92 22 7 4767 5442 348 0 1142 1 
+1,4330 17 1401 908 529 1712 60 321 4519 23 9937 6 2173 5 3 2105 1852 10 31 246 20 3 2514 9684 5511 20 1962 5796 42 441 184 3575 3164 134 18 8 95 1 
+2,7511 7874 2997 204 823 2268 3884 22 7 43 997 4 0 728 506 14 3 3504 72 9 2361 3884 246 3 77 145 997 5 0 1863 340 8 86 2 10313 6595 2361 53 
+1,1184 411 444 4514 790 2172 5 1599 109 39 10571 64 1249 14 21 411 1029 6 196 5491 198 3384 5 3 5265 1681 8 0 6348 1223 9226 49 1846 39 538 22 7 172 1124 5236 119 47 
+3,607 1043 1880 2956 5952 6633 2241 1078 1886 4159 6 26 2246 7133 4159 8 86 8909 3772 139 104 2 3040 254 6 16 893 676 1 
+2,1636 92 10720 12 2160 102 4008 892 35 2144 2 1704 138 4 0 92 22 7 4880 1821 3722 5 72 9 0 151 177 1 
+0,470 379 1177 2410 38 2683 174 3419 5001 6 10306 30 1547 14 1272 1806 6 489 1 
+1,24 101 152 647 53 0 131 10599 2 0 606 212 1591 12 3 1743 581 2176 5 257 3657 52 1011 2 1426 5 1928 1 
+2,6856 7145 1883 4304 550 19 674 3 1104 380 5769 7711 6047 2993 7712 1950 1158 69 10733 50 2484 880 42 60 292 5 0 9391 4 324 15 6683 7139 3530 9 3 38 283 0 3614 4 4447 13 3300 1 
+2,1604 771 52 166 9750 5977 10889 8353 8202 9338 7155 5913 4367 10269 26 12 3 359 2347 6926 6 2722 203 9250 8 3 1327 228 3563 61 1759 60 8 3 1497 117 4367 7 2103 9776 3 11052 4753 9 0 1029 7 5390 1083 57 10 55 3597 0 339 47 1110 4 1327 9640 3 1 
+0,11 4700 7856 7363 3 849 229 900 2 516 3 9427 1476 915 5 838 13 780 2232 776 16 10102 82 0 4933 327 1801 30 9708 4 8121 0 516 532 564 160 7 0 1769 4 1456 786 129 6232 5997 1 
+3,7495 6219 3945 1393 435 7533 3852 10770 26 8210 2234 53 3 5508 5 1582 2626 180 12 48 9408 2726 9720 6 3 4550 4460 26 10044 10983 1025 26 2733 3624 1 
+0,7304 220 239 949 958 408 13 10893 2505 61 417 53 0 381 1096 4458 12 16 1245 671 5060 249 651 69 34 5448 155 121 4951 50 463 1637 966 12 4020 102 3624 9 185 119 921 1 
+0,2875 4 1632 97 6 4269 1888 58 29 34 1209 2 309 505 275 567 156 2012 12 517 9377 243 3 10412 338 172 95 1 
+3,80 93 17 41 45 202 3351 64 217 5524 8 95 15 485 3128 257 301 8 0 221 10 55 1438 0 143 7 382 2518 1 
+3,412 430 2253 111 3016 2848 8341 239 4056 1628 6702 3956 25 1518 0 4456 4 16 3621 9 16 2147 2427 4 0 2855 4542 3812 1 
+3,4342 1165 25 473 2 479 401 2267 5749 1672 0 5637 1213 2 715 0 38 87 10973 87 0 603 1656 342 18 95 1 
+2,6960 17 4327 7 6481 6762 181 26 645 9 3 75 9277 753 535 1538 12 3 2198 321 243 1788 1345 5 0 2841 465 330 63 346 8 95 1 
+2,7397 67 239 7596 195 1802 245 10959 4 3 1430 4 999 3309 197 6121 6 6062 1616 54 30 416 2 1771 5 0 1885 467 1 
+0,129 7497 6855 5902 4135 24 5870 2742 103 1842 132 46 68 5057 82 53 201 847 1117 2 3483 10716 10146 9095 5 11442 9332 6 2043 2470 155 46 793 1 
+0,24 4254 58 920 0 4775 91 1231 6 21 761 23 130 97 837 19 7 734 2 10483 787 53 3 1299 4 1015 156 1232 1 
+2,6491 20 9241 11204 6 3 9149 2114 0 88 128 8828 1475 9 154 2412 966 506 7589 7699 2 192 1329 674 1 
+2,7244 7385 25 2029 912 4374 145 5 3695 22 7 169 5615 282 14 0 151 177 961 22 7 1443 649 698 118 276 39 407 22 7 285 174 2520 5315 31 215 11362 14 0 10567 39 3549 51 5 0 59 1477 1 
+1,5844 56 652 449 442 2 66 268 5 0 63 158 4 250 3446 3 2193 23 0 170 3306 3 162 380 1 
+1,469 4 48 314 4430 10 4684 217 3631 31 1678 198 20 3 1837 1411 134 18 95 39 1456 5498 4 3312 5 0 3519 7 2250 27 434 729 4642 2 1374 1006 1690 0 1653 31 4642 2 752 7 615 83 6118 5338 1 
+1,125 11 1129 4 287 594 1998 149 329 31 3 8692 71 3 294 188 0 1963 1407 387 34 10372 5 0 88 128 42 8 0 917 4 4029 94 619 9 118 57 1134 23 0 126 237 0 9088 4 947 2424 152 1 
+0,2978 1440 7643 1610 25 124 26 8740 507 83 3 1195 206 1288 54 9905 0 3266 668 2 10431 10981 2 3 10236 434 3035 8122 1309 95 1 
+3,3946 69 34 597 9510 778 5296 651 69 1416 1478 281 1 
+1,836 100 497 214 352 8456 6 8285 4519 64 985 5 0 8228 4 3 3574 228 192 51 20 906 480 1383 432 5 253 22 7 1191 635 369 18 1 
+1,4174 501 244 25 1802 0 2419 488 4 9840 5 349 12 2963 907 211 7588 5910 6273 25 76 1309 60 21 4836 2174 9 605 2 10515 0 1430 5 0 4036 5036 1 
+2,78 5607 5 2166 2 185 1763 4 430 823 250 84 121 760 0 88 128 59 1359 2 2570 2 59 112 3453 414 948 4 0 88 128 4 324 1 
+1,24 3960 4225 6275 408 3 238 150 145 5544 18 95 19 13 900 21 586 183 394 8 0 1206 7672 2305 5 838 3 338 10 55 34 0 127 7 580 38 202 3432 5 401 103 1 
+3,273 17 11 41 45 221 1534 49 1530 5 0 63 158 61 43 924 15 67 96 560 6 0 733 774 2763 0 107 18 8 95 5 3 218 10 866 2648 3320 5 0 2199 1 
+3,11 56 221 1534 49 1530 5 0 63 158 61 43 924 15 67 96 560 6 0 733 774 2763 0 107 18 8 95 5 3 218 10 866 2648 3320 5 0 2199 1 
+3,273 17 11 41 45 221 1534 49 1530 5 0 63 158 61 43 924 15 67 96 560 6 0 733 774 2763 0 107 18 8 95 5 3 218 10 866 2648 3320 5 0 2199 1 
+0,577 959 60 0 4772 130 5516 10722 2 1613 84 124 69 711 2695 15 302 1 
+0,24 27 38 596 1315 2105 9317 30 348 783 5 7945 212 1194 42 3495 134 1665 69 1829 230 2 4943 0 2138 4 0 9986 3111 1 
+3,6328 1642 3906 2382 6978 6 26 56 3221 2391 7965 8039 9 3 2233 5 0 258 11152 2715 44 107 1212 9 4482 726 2280 6 2267 1 
+3,11 11156 1979 0 4721 1541 62 1004 166 3 10195 5134 1542 1 
+2,7408 13 1252 9 136 5 7637 1369 7 120 0 79 29 277 2 8676 1 
+3,129 1040 6318 6480 6228 129 2391 7613 117 0 5647 4385 7361 3 10907 6 8401 342 6 1375 6520 3 10787 342 3371 2 9822 48 10735 5 2290 7213 6 6357 5842 7178 1 
+1,2372 6747 22 7 10224 32 21 8096 3733 15 0 7449 687 7267 8769 287 2 9004 21 659 338 12 7605 6428 1 
+0,87 7843 22 7 8651 30 8782 87 4077 2328 7015 2804 5 0 819 1786 649 2186 3119 0 1975 582 83 6687 1 
+2,536 5425 8 26 234 4194 13 52 3078 19 34 1064 1375 22 1194 71 1191 3060 4194 29 52 1019 14 4007 1352 7891 7479 195 2067 5 3 5502 2165 6 192 5 2104 26 3609 2603 
+2,80 7956 2720 21 1845 59 546 5 0 146 22 7 5509 109 102 6614 2860 141 8487 9 9346 26 10581 355 26 8572 5268 92 336 5958 6291 20 289 5738 909 1565 6302 6548 1 
+0,270 18 86 19 89 4668 3336 10 1685 0 4728 4 0 201 1532 1590 5 6424 1161 6 10 3 9911 1532 85 34 1131 2 5292 1730 3645 119 7647 1 
+1,24 428 2551 3591 3 369 974 5 0 10526 6223 233 8011 2 2288 8 95 6 586 723 2601 48 94 64 709 0 4057 306 354 18 1 
+0,24 165 7 3 9254 4625 7 9046 4860 9747 5058 185 861 60 4 3 3193 9412 6 1535 164 0 2192 3418 118 5 72 2 9930 53 185 8932 1 
+3,11 220 2586 239 356 788 9 395 322 15 2995 6 2683 8 95 1759 60 16 318 2 477 3 1424 668 5 381 1096 395 322 15 1251 1023 2410 365 797 6 0 1090 10 592 1232 197 203 1 
+3,4342 1165 465 105 2 479 401 2267 5748 546 3244 1656 10 29 34 3250 5 3763 6 5666 11 1213 13 436 10031 48 3096 5747 1656 1 
+3,3843 7496 7245 17 7583 775 35 8277 8 95 9 3856 7 4056 37 27 376 305 373 341 397 91 410 415 402 4043 3841 286 223 423 418 36 4043 3841 37 27 36 2 479 1171 7 3812 37 27 376 305 373 341 397 91 410 415 402 3790 1628 286 223 423 418 36 3790 1628 37 27 36 37 5755 98 36 5 3 374 289 153 338 10 29 1082 0 92 7 199 4542 1 
+1,2273 1702 0 88 128 79 4 212 2844 1037 1321 2264 2818 292 95 5 0 59 2723 1455 8835 0 77 72 870 2476 1 
+0,11 56 201 354 270 2496 235 8 3395 0 1521 4 0 201 1532 132 0 9198 4 10028 10530 16 5564 2 0 587 201 1540 119 47 1 
+2,74 17 11 41 45 146 7 316 79 31 3142 20 2273 95 3251 19 21 59 112 144 9 0 43 72 248 2810 102 2370 797 1823 2488 1 
+1,24 7149 1198 30 3718 101 152 2 3714 3 3590 63 1151 3104 14 26 1432 119 65 2 1139 817 878 23 0 11242 259 5 100 6 135 1267 10 30 3 4760 2 26 491 237 988 845 8648 0 622 284 4763 0 3998 2572 2497 44 524 149 329 7 8552 159 5 1963 1 
+3,27 547 528 380 5 3 1542 8 2988 1619 7121 78 760 94 2 2210 3 327 9628 2 792 542 13 15 447 3 422 15 245 2 2710 0 3743 1 
+2,126 336 1958 6964 124 0 59 10219 11293 403 8 95 12 3 252 159 139 3787 666 39 21 9011 8941 12 2221 7768 7671 1 
+2,260 2464 3613 3814 4091 13 256 1030 2 10713 3 357 450 592 2 0 2143 3560 287 52 5596 71 3435 26 492 14 59 112 8 33 1 
+1,273 17 11 2307 1987 160 13 21 411 5397 14 0 930 2593 661 4 0 2381 1880 1056 147 8 95 1 
+0,574 29 396 3 8542 4 16 691 6261 2706 395 6 1834 2 40 922 119 65 14 0 41 45 330 1327 8447 4983 3 3367 206 1470 3168 4 320 1777 11412 360 48 76 2443 12 0 630 1 
+1,5901 25 680 315 10 19 184 25 3 662 4 235 2 277 2 3597 0 587 2486 10 19 13 11028 120 0 88 749 25 2509 15 0 2801 1564 3385 869 1 
+1,11 2744 13 5940 6 0 8001 30 4107 42 0 2446 2087 0 3271 10990 2181 4 0 8689 267 13 256 2142 5 6886 1295 9965 3048 1 
+0,11 62 10837 3 38 668 8 0 5001 2 34 404 5 16 3509 2683 1 
+2,4266 7958 747 2142 0 106 22 7 1755 356 2178 3 339 243 1788 1345 2 1116 2312 3995 14 0 176 4 0 5048 39 0 63 346 4 0 186 1 
+3,407 22 7 221 7979 2 0 1246 578 2117 5 217 133 103 5 0 63 158 15 3445 2538 23 3 5375 6 981 10390 51 629 0 107 18 95 1 
+0,1640 1771 16 11120 130 25 10630 122 16 11065 3108 9 0 671 544 1010 1613 1 
+3,4301 0 92 22 7 384 199 342 4 344 1123 18 95 10 3 114 4885 10 68 1082 3 10174 1805 5 16 3473 190 3463 387 1749 383 5 16 4944 1441 1 
+0,130 123 19 105 2 4549 247 0 174 3128 527 2 16 2059 178 307 163 8938 22 1613 22 5 1568 1 
+1,24 5778 7522 334 5862 54 5340 15 1966 7 1974 2 0 88 749 9 217 742 103 770 5 111 332 8 86 39 3 174 3390 1966 7 946 1923 6 16 41 98 4151 18 95 573 31 2261 1 
+2,2357 1944 1092 22 69 794 2 108 0 2848 1179 1624 8 33 2 9948 26 1845 1813 2906 469 2529 403 1 
+2,435 0 4047 146 22 7 316 79 2106 2 262 73 5 11 1036 378 6 6326 1404 961 2416 352 2131 8 28 1211 19 31 3 730 3464 9 0 522 6 0 797 1 
+2,149 3925 3547 3 1486 2759 9 1002 3362 1100 109 102 52 8563 2 521 5 119 65 22 7 6274 683 534 14 7717 4 828 5 7183 1 
+1,125 431 501 244 149 2328 31 443 2 3443 26 8705 39 317 11153 1418 51 2019 582 50 120 52 2620 4 3 1431 1567 1312 908 4 8301 1 
+2,74 205 462 1614 539 1400 793 1811 657 49 61 779 3 92 503 3055 177 64 138 8928 225 912 1309 1331 6 482 12 3 4923 2229 6 3 2762 5376 457 339 665 
+3,2319 2340 1042 539 951 956 22 7 3313 1719 1534 3 2229 5 833 15 302 42 184 867 14 546 3257 1130 2226 2 173 733 5566 12 9784 171 1 
+2,325 22 7 4170 3846 31 44 0 1281 39 457 0 1142 144 5 0 59 9643 6798 1034 653 195 6301 92 336 3846 1249 23 0 453 6 209 21 382 231 42 9117 
+1,2399 2276 348 44 15 1061 22 7 3768 3534 1120 8 136 13 3 238 798 8214 3227 12 9549 38 361 53 21 965 1985 221 1 
+2,2416 1514 6 870 1693 420 336 3062 2660 16 283 3695 83 1405 2411 3051 22 7 1054 852 0 356 880 670 76 5 0 1189 210 1489 0 356 2484 880 670 
+2,469 106 39 212 4037 467 875 2241 267 6 612 894 1781 1038 3925 543 0 7116 4221 1037 4 9010 87 5 3 8453 87 12 10456 2 9962 3 8544 8224 759 6 1987 0 453 4 0 250 5622 283 13 5 5028 213 0 2236 1340 2 553 5407 10370 6011 1785 2335 4066 18 160 7 1 
+1,4080 315 2831 12 9661 6373 8812 9413 3 3519 10 215 6902 1683 50 3781 3326 6374 6 1457 19 2 1760 109 5 0 576 359 132 10988 2 142 444 134 18 1 
+0,80 93 117 3 1827 1139 5 0 234 0 152 610 25 10239 0 727 4 1691 1079 3 38 218 2 1360 9213 8 512 308 3398 10 3293 4 4565 4740 6 135 1257 5506 2574 30 0 241 362 
+2,74 205 6932 14 19 40 234 497 214 0 56 316 79 124 22 69 34 760 2 777 122 16 112 144 691 3 106 10 31 59 5 10640 505 9 16 3623 1849 6 16 9592 11365 0 
+0,462 0 421 547 103 2242 32 1980 130 22 7 307 163 434 2467 71 440 519 1507 3609 8 3 5582 3136 1 
+2,11 409 330 5315 51 54 89 10305 2 0 75 133 12 1715 31 8580 60 4 10803 15 1813 2721 2281 5606 51 224 5 241 5718 1008 1 
+1,27 4395 228 5025 339 334 1626 2756 9 169 103 136 9 0 2459 4 0 409 10975 6909 6 5336 2 769 255 3458 54 10202 2 2618 0 56 1974 2 0 743 150 1 
+1,5897 7260 6967 7351 3893 244 7819 5962 25 90 1457 60 4 489 39 133 276 4 718 3271 15 0 660 4 5458 3460 506 122 2 9381 0 5960 6768 1936 1 
+2,74 205 1631 135 61 185 8074 1453 422 1248 5 0 174 1269 6 0 713 4495 1492 5 0 570 364 625 9 3 357 1727 87 2884 1058 6446 22 22 230 510 185 106 222 1614 539 1 
+0,428 8152 0 493 4 3 710 2749 2188 506 601 186 311 1480 4 498 40 2199 1005 19 0 4613 391 1774 417 2700 56 2444 18 40 65 1 
+2,1614 539 0 10364 4 1400 54 1101 426 3762 177 12 3 159 339 1331 6 482 665 1097 189 119 177 281 4739 95 60 4 665 6 5 5464 1 
+1,27 228 5 1966 25 5336 57 343 2 769 6 740 716 2 1504 1550 9 3 728 4 644 6 1837 10201 5 2811 360 0 2459 4 3 409 67 11038 1 
+1,4044 7158 8973 788 15 4877 11405 501 244 2399 2276 25 3606 456 4 766 1253 0 420 1666 1 
+2,74 204 2253 325 124 0 59 146 22 7 3379 630 9 0 43 72 5 551 8 95 972 0 1927 66 73 12 3 4923 3360 1 
+3,80 93 17 3977 185 5328 11 1198 30 1080 2 1029 464 1329 222 5377 14 0 606 212 1591 0 202 143 55 167 3 4543 858 119 65 511 148 1 
+0,5930 3828 3650 3 142 2719 14 6929 1875 5 828 40 65 102 912 1843 2 1422 3 6074 1874 1729 12 3 10212 5 189 9364 11 6563 6550 285 210 2818 31 9192 20 3 4351 1199 1578 3105 54 985 0 9660 3 10921 141 141 879 374 351 1950 1823 5609 10 19 1158 69 9405 6 18 11 812 396 19 160 3 3326 3500 18 1950 892 5609 50 0 812 0 812 0 812 353 0 8272 31 8609 3828 209 3 9645 23 189 4491 6 813 0 3105 19 8244 2 0 812 132 3693 19 14 203 1557 10163 91 
+1,24 27 2471 2234 6 11055 8546 5 3 9933 233 4 681 2364 136 515 822 94 6 3403 50 323 716 0 317 6948 306 354 147 1 
+0,464 366 861 770 3694 78 3237 32 3 662 517 8079 10172 16 5436 61 78 85 155 366 1637 8617 60 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 3020 3005 37 326 36 37 572 36 1 
+1,125 576 907 3735 503 23 633 1644 717 349 2 783 3 822 784 4533 5 1017 14 1657 107 644 8 3189 5 0 259 1841 681 1511 4 298 1 
+2,74 17 11 41 45 288 7 316 79 1140 35 146 230 2 108 112 136 15 263 825 161 5387 9206 2 0 59 3653 9 3 9392 2445 522 8 0 10141 110 4 0 151 252 1 
+0,3946 22 69 32 178 609 6916 51 8098 213 160 22 7 3 3425 2 1117 76 2 9852 10766 396 8918 121 68 309 2 8085 185 178 609 405 2062 249 46 22 1273 993 859 198 6 8017 
+1,6410 6 4249 6 281 4 5836 4239 30 210 4659 20 3459 12 41 45 196 8607 2 796 2451 1 
+3,11 204 821 1458 1786 971 8 0 1960 4231 3731 0 4038 6142 9504 10 0 790 1132 94 10266 0 4231 31 16 87 9279 8471 711 87 6 0 707 8384 1558 1 
+1,56 444 6 1300 540 480 1385 334 432 22 7 1007 1670 136 5 1170 278 15 0 2233 10 1097 0 4515 443 197 56 269 196 6 0 1007 444 5 253 304 9 3 63 106 1 
+2,24 7683 20 3 5373 9366 5 0 2102 4 0 282 0 88 128 555 2 2318 407 20 3 9455 4 3 63 136 110 5 0 570 364 625 427 3 282 0 126 146 2784 8905 1 
+1,125 11 88 128 2218 1730 4 6043 5715 388 9397 5 313 8326 2 1510 44 3 2503 5 21 7211 11238 8395 1 
+2,3961 2281 4 752 5154 4169 3932 4 684 115 66 66 115 289 139 6 124 0 4219 4210 1327 630 136 2281 3608 1813 124 189 194 3073 403 1 
+2,6161 4168 693 468 3030 3608 2352 2323 1516 33 22 7 131 14 0 5700 161 2408 3075 535 3 131 56 330 11197 51 1 
+1,6057 45 938 27 1464 2795 31 1115 9 0 227 939 3198 136 15 2567 5203 2417 3043 4002 85 200 9680 391 6113 8 33 110 1 
+1,24 636 1946 149 329 18 136 52 7 5 443 9921 12 48 389 2 222 2 0 594 15 26 8055 3247 180 23 2694 50 26 11287 5 1963 1 
+3,353 57 4 0 182 8986 528 5 260 1165 22 551 0 10428 504 3694 7467 6342 0 1213 22 7 9336 431 612 875 123 52 29 9932 3 585 890 
+0,9627 1565 87 6980 855 6867 32 118 1000 87 6556 87 3 3708 8568 531 6498 8318 7815 658 988 1296 2632 12 6979 169 1 
+3,27 38 872 10 9163 3 1253 4 1970 2706 9777 8790 1575 3 295 646 4 8280 5 0 4796 124 3115 155 76 4456 23 3 1787 4 8024 2 0 56 1904 6 2890 1578 1 
+2,2352 2323 13 57 664 503 23 457 26 63 699 3795 403 12 0 431 2699 0 131 4 0 2408 3075 535 14 4116 4061 1 
+1,6436 1417 146 5 1991 11236 4936 0 9861 10872 295 2 0 576 228 5830 0 94 9823 14 0 4529 241 426 54 711 317 6 10646 9764 
+2,7580 3608 314 3961 2281 124 189 43 403 5 48 103 20 972 684 22 7 4169 3932 115 66 66 115 289 139 5 0 131 4 0 4219 4210 630 1 
+1,327 378 1035 2900 315 24 1199 25 2075 1265 3165 5 315 7 598 1841 298 233 42 3265 5560 184 279 644 6 1968 102 2089 0 3165 3 88 749 79 18 136 11423 51 3 591 10 55 1438 287 315 13 246 12 587 1808 1 
+0,101 152 5059 2 1128 0 11002 707 42 40 65 19 31 0 3342 44 14 0 2934 945 54 1732 0 10244 2 222 39 0 9067 0 2658 8 0 156 1 
+0,10091 487 1056 6250 5 4301 1345 728 344 1123 55 646 8758 123 0 62 87 117 3317 1684 40 5566 55 231 2 885 1434 1 
+1,2999 501 244 8972 2399 2276 136 3606 456 4 766 5 0 212 5864 7108 0 1295 895 1262 4 0 2120 3 10505 243 0 127 22 7 8635 1 
+2,2318 407 22 7 3814 4091 54 1475 249 4987 5 457 0 5636 5200 2004 14 0 4078 465 2927 2854 6105 292 75 919 25 3 864 9 21 59 112 144 5 0 9733 1441 195 1 
+1,17 6136 30 4494 23 9420 1484 1988 8299 6 10704 30 4425 2 986 5198 2562 288 5 0 659 1487 2999 179 4 7306 1 
+3,117 3 592 965 5531 4 5 16 240 1387 3918 6911 25 8104 9 175 799 188 0 7357 4 5817 22 7 43 48 389 4 10608 263 2915 6 13 10449 3 5194 4 0 175 2005 2 1244 54 25 2021 8 
+2,412 204 1162 2268 7618 2720 3 589 9031 9 1921 88 195 15 0 808 1839 73 73 14 5969 4305 2827 6433 808 1475 1108 9 35 63 2384 467 2506 4 0 
+1,81 27 540 3407 1071 2 34 11350 15 269 444 2106 2 601 460 3751 4 0 2330 1579 4299 70 100 22 7 176 540 480 1305 168 12 3 717 3402 4716 21 157 2 0 443 5 253 1 
+1,165 209 66 276 9 752 22 7 142 175 2 694 120 2788 1244 4798 3145 23 0 3464 48 2038 3500 4431 10207 2 2042 3631 
+1,431 501 244 149 2328 8 33 172 10 0 119 512 237 29 34 304 8 838 374 573 813 3 490 508 292 10 817 29 1435 
+2,74 11 252 4 0 7934 7214 0 778 1091 10 1158 22 69 979 2 21 4806 157 40 1211 6 722 0 92 10086 151 21 8105 1 
+2,117 3 659 139 161 364 282 2918 1900 2914 4 2364 54 124 112 14 73 1034 666 75 65 10054 6821 5941 4 6380 5 
+2,6182 59 112 3453 414 948 13 122 5 0 88 128 6 1131 2 592 8 948 5 7922 9 0 7465 855 7473 
+3,11 1213 18 19 29 504 2671 241 139 4 0 459 682 439 4 8313 3 2373 2227 164 3 483 1237 163 322 15 2394 71 6519 916 70 1 
+2,74 6065 4279 701 69 711 597 3203 8 0 144 2200 75 110 429 0 420 8093 31 901 52 737 23 348 0 2247 11426 82 26 781 2 1796 19 122 8 2 348 19 82 333 6 1259 19 290 26 4590 1 
+1,7957 752 17 6119 1157 33 9 3 38 267 5 3 3690 237 42 232 8915 0 1616 1418 369 1781 54 31 192 2 108 85 1766 2 10874 60 10402 5 0 11200 233 1 
+2,6894 61 48 552 132 0 59 146 22 7 570 364 625 3612 8 95 0 88 128 3897 947 852 802 149 6075 23 0 282 39 256 813 20 
+1,1 11 409 2120 2121 0 893 5 2360 2 1669 87 8633 11019 87 4 9123 23 16 150 676 163 4013 9 4454 0 893 9473 1925 4949 2336 10742 6 710 
+1,24 7516 8278 51 3 3514 5 1145 100 8 33 5 0 258 475 2766 0 127 7 4693 67 240 3 677 67 317 18 1 
+1,4245 754 1168 1176 15 754 1168 3669 9584 12 1031 4850 6 9107 12 16 707 15 0 127 10 2142 182 4 0 8158 4 2398 351 1567 0 4725 5036 7 10593 30 1485 8 1777 9251 2 9909 6 2646 35 390 1 
+1,2994 17 684 7 107 6 1271 4 16 1925 2644 1030 269 1006 33 2 513 48 409 2615 46 64 1259 3382 5 3 645 2 584 3002 2 10552 16 1669 8 1925 4949 5 3601 1 
+1,5892 100 17 750 6 883 10831 188 624 3595 644 8 100 7 1145 67 3515 9480 33 6 9180 1670 2 4777 0 4884 1 
+1,412 27 3585 6173 5179 25 477 0 2657 1493 2 1832 0 2303 2286 947 6012 825 215 6303 536 136 1009 2587 9 0 409 1887 1 
+2,24 219 1944 8508 21 1845 1813 2906 469 2529 403 14 0 2848 1179 1624 8 33 583 76 457 9 118 0 63 72 5 740 3557 40 283 1 
+3,11 77 199 56 1002 3686 4667 181 3 10835 40 919 12 105 2 425 372 399 8 35 3728 6 301 1473 5 3 547 103 1 
+3,353 75 65 22 7 1528 12 575 1044 1754 8023 6531 3067 7875 121 32 2 11408 230 1812 0 1945 6 2305 1174 13 50 5231 1282 2063 3621 1 
+2,532 0 48 6 3 779 122 10801 12 57 6 3 779 11210 5 3 10175 1792 735 60 2 34 26 2227 2 3 1142 144 1 
+1,6045 45 938 2417 3043 4002 3154 0 227 939 3198 12 5249 6 391 1464 3665 1562 382 33 4892 2172 6 5037 60 558 2 14 214 1573 161 887 1 
+0,7709 612 994 1781 149 7656 25 4417 7743 22 7 9816 5 16 186 139 153 3 47 344 190 264 29 5352 40 47 5 0 279 4 1751 370 438 6 0 448 439 4 4409 38 300 1 
+2,7557 2476 6097 3606 3 159 1813 3676 108 44 4006 726 2387 8 33 12 3 73 226 224 8691 4 9876 2268 4345 22 7 11298 589 454 1 
+1,27 2999 946 1931 1820 33 18 4060 771 76 4864 21 520 1097 8132 282 5 227 496 187 2 0 317 1036 378 4 1061 5771 1 
+2,603 753 535 1075 5954 6874 1010 6007 2868 2975 6 6942 2296 15 26 5593 2125 8 33 9 119 162 22 7 664 83 0 88 128 1 
+2,4109 7186 2311 2096 3869 6151 18 75 65 10 1274 2311 22 7 2257 137 5522 159 2 2832 3671 5 0 8703 87 213 5132 5059 203 87 42 5447 10 13 76 2777 1 
+1,80 93 7745 4 350 4 2034 9814 421 0 2959 3039 6529 729 4 0 606 212 1591 8 33 4586 4517 11375 6 1683 579 259 8220 15 1437 1734 2 9982 101 152 9 3 63 1151 691 0 4827 4 0 1432 0 2034 2666 0 3664 23 5356 2 5356 9 323 1675 15 46 10761 1247 421 1 
+0,464 121 1416 3345 2 5323 9 2242 527 931 2148 145 6285 6252 71 145 245 135 163 1330 711 2 0 4338 7547 9 154 1 
+1,7288 349 197 298 907 6 0 576 1910 32 3584 39 3 822 784 4533 20 907 54 543 4092 4 11309 3 1687 20 515 1573 3189 5 401 11308 1 
+1,1597 3838 612 2913 3950 25 3577 0 2715 44 1641 2850 22 7 649 2 222 2 259 12 100 3950 31 1457 2 3583 23 26 1294 966 12 238 3838 1078 6540 6241 75 6769 39 4117 
+1,24 27 229 4 2646 1198 232 174 60 4 489 358 8 101 152 6 0 606 622 2 979 122 2 0 2633 8 0 4827 4 0 606 212 1591 1 
+2,6673 179 151 645 3 131 9129 2 0 1219 6 4937 4 0 250 1635 252 12 3 2198 622 243 3 585 1281 1 
+0,171 1000 21 2549 1022 53 1783 33 1979 3 4116 2360 5688 1076 1308 147 7936 171 22 7 107 145 306 354 1 
+2,111 332 1648 674 4871 2314 4295 713 33 22 7 83 0 849 827 1049 12 3 5392 215 4442 4295 13 2630 15 106 2 106 1 
+1,80 1892 5776 1570 501 244 4140 756 8 33 8616 0 242 6 2542 9 0 5435 516 4 0 3805 686 1278 1 
+2,2333 7666 5917 25 124 0 146 22 7 1477 5 3 72 4 66 169 2816 7142 126 7028 6818 31 3 2759 10598 12 1171 22 7 7838 6908 3332 194 1 
+2,2894 2359 2161 0 111 332 1907 2151 4 0 618 462 210 0 891 8245 2 2942 1408 1 
+2,429 4008 167 3 858 19 2161 3709 155 1828 29 1478 293 10886 5 499 129 2382 2310 7648 5608 4 11 6132 2396 7078 1 
+2,1952 22 7 7054 5838 6783 124 0 2841 330 26 384 403 8 0 603 1555 40 283 6 2868 2975 31 57 4 401 9321 2 1427 753 535 8250 33 1 
+2,3911 2 557 4502 0 558 795 722 10783 122 14 7611 8 33 11 3428 506 8 39 21 784 2033 42 19 209 138 876 72 9 0 4496 2662 2 672 138 876 11346 1 
+2,24 2352 2323 6 2948 2881 55 4787 51 2 297 1034 161 876 14 0 41 45 330 213 46 1101 1979 0 9514 56 330 7577 3654 1 
+0,130 13 10681 51 16 105 9 0 119 565 4 178 2 167 0 114 82 0 8927 1422 6 53 2995 20 0 157 4 1568 1 
+1,497 214 321 94 32 90 709 5 3 1423 2455 5 764 2341 0 63 2026 2551 5 375 44 0 736 1 
+3,1039 6 0 88 128 32 656 2 673 21 759 8 10531 56 2599 4 327 2631 4572 1002 490 147 33 1 
+2,24 4356 6552 31 4411 23 0 1344 106 4744 1001 33 20 0 2828 3817 6 31 122 5 0 5060 83 0 835 1403 1 
+2,1384 10 77 103 248 52 1552 26 75 3516 9 0 1064 1379 1587 5841 775 51 57 11230 145 6 57 246 5 401 889 33 9 4355 27 7037 5 120 55 34 26 131 453 5 0 9894 1 
+3,80 93 17 3977 185 5328 11 1198 30 5 1029 464 1329 222 5377 14 0 606 212 1591 0 202 143 55 167 3 4543 858 40 65 511 148 1 
+3,638 17 466 223 64 1251 5 2642 1009 733 8 28 12 766 5 0 2038 652 1289 3319 2 2112 207 348 10 246 624 9273 322 15 1266 6 905 3305 223 1 
+0,24 171 8 33 1000 3 1022 10 29 1426 60 1760 10992 6 135 1523 1508 9 678 276 6 626 2 498 107 2100 147 1 
+3,637 4155 97 0 1985 56 3298 62 140 425 16 850 206 4358 855 4179 4222 39 0 850 1441 22 7 1469 5172 11190 1802 3 2806 104 437 488 
+1,11 126 388 1277 9 339 552 8 33 12 3580 4 0 3565 480 7076 334 432 5 0 5565 1170 540 5127 4 432 
+1,100 7 1752 3534 1120 18 10 52 89 304 798 9843 12 3580 4 9575 1107 23 6411 4249 6 7528 2 10153 203 2 1651 3 107 3109 488 1 
+2,407 22 7 4090 6669 6 2364 22 7 2918 1900 2914 1714 35 1293 335 59 8155 9712 292 8 136 15 46 124 35 63 9320 4 0 252 1 
+2,1352 0 1355 177 13 184 133 103 503 0 634 359 13 502 4915 51 2 1261 0 653 11 179 4 44 459 104 13 10450 1660 9682 5 
+2,11 43 1494 5 0 1189 2040 75 1582 29 34 0 43 7378 82 0 8248 9 0 1907 15 2894 2359 555 0 438 9 0 916 891 2 3723 2942 1408 1 
+2,464 121 2154 0 624 9075 4 7131 5 693 6710 6 64 396 82 20 0 1661 878 8 0 9471 6008 2932 7 1292 361 34 8557 5 0 3416 10 957 6843 7 5703 1786 8260 1263 2129 4 880 597 11 1992 13 9085 10508 6 1263 3688 8073 23 5935 6409 2997 2 7801 3808 2 111 332 1 
+2,6609 263 0 92 10877 2788 5 0 2102 4 0 1193 2376 0 9519 7765 6108 13 118 276 503 23 3693 26 43 1492 4 0 283 3994 366 11053 10794 78 29 34 1856 180 78 1558 121 2 2795 180 597 1 
+2,3021 3979 123 52 22 7 1131 2 4821 0 707 15 256 597 3639 2 108 5 111 332 11 92 22 7 262 73 670 23 4328 25 901 77 56 4193 6 555 5 0 384 346 1094 72 1 
+3,80 93 204 742 250 6092 2558 6 9380 993 590 9 3 3556 157 2 3 818 162 1669 8 2631 4572 655 2 0 56 30 60 4 9770 39 327 5803 244 1581 4152 
+3,117 838 182 216 8434 1682 318 2 453 8578 3 5235 3088 3432 4 567 190 1121 1 
+2,11 144 3176 89 1097 3067 6106 89 502 2188 3 547 5464 42 52 89 304 26 8994 925 164 99 0 9453 6 8748 278 164 0 224 
+1,4060 27 38 3882 5 1061 13 362 2 34 11015 5 8 28 48 276 39 2903 244 2399 2276 31 181 0 127 22 7 3768 501 244 1 
+0,542 971 3261 468 3797 4242 3912 5 178 1949 1161 84 1415 3098 2 5240 344 371 6 5552 35 142 5199 132 884 203 585 854 2 0 392 7276 66 4 66 1 
+3,469 4 0 146 54 920 95 22 7 4863 10628 1231 13 5 3368 8598 14 1594 7420 6671 6 0 48 135 10995 64 296 8 33 3 2071 10843 18 1 
+1,125 2991 2131 777 325 7 669 2348 1936 3 357 231 44 501 244 149 2328 7 3213 107 15 8394 1673 9 838 374 2535 42 967 148 0 905 282 13 184 597 337 2 1228 1 
+3,7534 123 19 1971 2 1081 16 2765 4 850 5111 9721 3796 57 162 10802 61 5153 1124 1 
+2,205 1843 2 1492 0 177 4495 82 2 1355 8 33 110 42 19 31 3 2230 891 11 634 30 234 274 4 0 8733 502 1 
+2,435 19 698 2 884 51 174 8208 471 331 3517 1958 7876 25 3 807 2101 1606 77 528 39 52 5449 3 4615 159 401 168 
+2,2415 166 455 200 2582 146 1326 2 3687 87 938 9917 1739 167 0 8751 60 87 3004 1615 10709 14 3 357 1023 5 0 471 331 8524 109 15 52 5576 0 827 1049 495 2 1248 82 0 1648 12 48 10063 6 0 623 457 145 14 0 1497 5 0 1013 340 5 849 1 
+3,607 6951 745 11 2144 4 156 206 372 25 174 90 9363 20 3219 6 2673 458 2737 4735 51 2012 2222 3178 6 381 8872 9 1856 8289 3724 8 35 168 1429 1 
+2,11 3613 4 0 59 2464 630 3422 109 10 160 22 7 141 5432 9 2052 497 214 76 5 0 1805 1 
+3,11 329 7265 1418 6104 1364 1397 832 25 147 16 43 487 386 248 9521 9495 20 3 791 11431 4 3364 1 
+3,2883 27 4895 4 2014 8216 5 0 421 547 103 2625 3 183 354 10914 9 437 0 7295 5947 6596 3913 1 
+2,3030 3134 957 5944 5406 306 1115 0 1692 679 5 0 471 331 5406 2176 22 133 137 1831 4 0 1594 3056 109 14 3983 1194 1 
+3,953 4685 3 49 1295 2101 765 568 1264 16 493 8612 0 303 7 2019 1 
+1,80 1892 6720 1797 2 1707 3 5379 1278 3246 209 3 1147 2570 195 102 19 3675 922 1249 0 1830 2 1830 3805 686 1278 84 68 2501 3684 5 0 3762 5682 9659 1303 23 0 6733 3053 7413 6703 14 
+1,125 27 5641 347 62 4 2986 9536 3337 60 4 2825 2587 9 0 259 1841 681 576 233 4 298 2 1116 21 633 1644 584 5231 1687 2649 1 
+3,7250 1042 539 4309 4 4017 690 7506 5775 325 22 7 63 580 112 1506 25 4856 53 610 44 7968 5695 104 4955 1992 5053 1 
+1,429 0 7185 6656 6086 4816 16 131 106 369 148 46 30 10196 12 230 19 25 2578 249 657 1870 5681 161 94 209 2 0 917 8 33 49 61 546 0 43 106 75 47 2 8433 531 103 4 0 5586 1378 653 1 
+1,7350 1962 6102 25 10204 53 1145 313 2219 14 214 133 94 2 35 3241 6 3403 49 61 742 15 19 1758 60 558 2 350 1 
+1,6602 17 1963 25 296 217 374 161 721 360 169 5002 2798 1684 19 123 89 3272 587 1667 15 368 4 2772 10108 1738 274 4 3397 3174 8 1948 66 1 
+3,3026 4309 4 4017 0 92 22 7 295 10976 4 11040 4455 796 1767 8 28 39 3319 2 673 759 12 8707 1 
+3,6265 501 244 6 2903 3957 244 1610 6679 9620 18 28 0 390 22 7 545 652 9724 13 416 2 11252 2779 22 22 50 0 618 84 5 815 11351 0 766 4 143 797 1 
+2,6263 2984 2462 4305 1878 7464 25 10840 4 26 5266 14 76 256 1209 2 1081 26 591 39 256 10479 4 26 4784 195 1 
+3,3047 3298 62 5873 6186 18 28 19 29 425 16 8974 1330 264 2 951 956 206 7738 6723 408 1 
+1,27 176 3106 2 100 22 7 906 1200 22 1268 294 1385 334 432 28 358 8 0 7032 1169 1007 2 1686 616 290 100 6 18 432 31 2137 2 694 105 9 3 216 718 522 1 
+0,7471 7319 1565 87 27 809 4280 6143 206 62 4196 4158 195 11257 3 38 8590 5 3209 22 2008 5604 1 
+2,27 8786 6748 10287 54 1980 0 294 188 11438 59 1477 31 1105 3 57 47 2218 5335 5 151 195 1 
+3,2319 2340 1042 539 171 7309 855 6122 97 7510 0 127 22 7 63 199 67 6 883 1506 28 147 3 2814 1269 5 43 779 383 6 18 19 105 2 858 16 10438 4563 20 50 57 1453 44 77 103 1 
+3,56 381 629 5259 5 833 3 2191 0 221 140 34 1446 23 21 382 919 3245 3910 629 560 226 268 75 162 2460 20 861 6 725 191 1 
+1,24 27 677 411 4742 5 433 25 1277 12 3 2381 2438 256 5022 20 0 2307 8 2760 52 2121 2481 385 2 538 411 134 866 28 1 
+3,273 17 41 45 1748 240 2532 8 28 4574 10 1304 1262 370 2573 5 138 1217 4 0 127 140 76 34 5452 1 
+0,10180 1565 87 953 29 694 10 19 25 8700 1034 104 11155 8 2 3 1143 2101 765 10710 203 5 493 20 742 1 
+3,56 1165 22 7806 1055 306 308 400 1756 965 323 8 306 10 663 6 5180 64 122 14 0 5456 734 2 9359 60 21 759 8 235 8593 2 1809 0 62 1 
+0,24 2370 285 1050 937 970 14 26 1313 761 814 28 15 8165 5 0 507 54 2743 0 736 10234 44 10366 3529 2 1129 623 2617 9530 1 
+2,536 5425 219 7237 18 26 2622 4 2404 6 7478 22 7 2622 4 2303 510 76 4963 3174 4 0 2855 22 7 9829 9 1197 2960 5 33 22 7 73 226 108 14 6975 1 
+3,80 93 17 41 45 67 96 442 49 61 73 8 28 8 867 207 348 15 1506 229 1929 4852 3396 5 0 1080 389 5 16 1553 3649 4563 8688 458 44 5429 269 67 655 1 
+1,744 100 2388 540 480 1385 334 432 358 9 26 4898 290 100 2 157 443 83 41 45 6 269 444 6 13 900 2 1116 0 718 800 5 0 1080 276 21 334 432 3106 18 28 1 
+3,273 17 41 45 1748 240 2532 8 28 4574 10 1304 1262 370 2573 5 138 1217 4 0 127 140 76 34 5452 1 
+2,1921 536 693 468 1921 88 22 7 5427 867 8 28 102 46 656 2 2178 5 3 226 226 2769 12 3967 14 1191 3060 1 
+1,80 93 24 1946 149 2970 18 28 19 31 9119 137 2 8715 2295 149 329 7 579 259 1510 77 2028 380 4882 21 714 10614 15 1198 14 35 420 1432 1819 2 10238 101 152 15 3 347 11338 294 1 
+1,1377 7007 855 1056 5802 497 214 8980 94 360 48 56 2480 64 709 102 3 2234 1423 2053 5 8920 2341 5 0 63 2026 2455 2 454 375 44 0 736 1 
+0,11 448 1028 4 1741 6 2242 25 2145 130 2 3592 60 3 153 702 1334 4 16 178 220 307 163 2 3666 142 1 
+0,6796 4177 1063 2287 25 1000 3 38 3402 1087 2 3394 9585 4 557 194 622 3114 6 3334 3452 12 16 621 4289 7624 6881 2032 7860 4232 468 2032 8111 1 
+3,24 428 11249 218 8 381 4992 192 82 3 10815 4 207 348 8 641 541 28 15 485 2802 10 3 11084 221 85 4822 275 194 158 383 3818 703 5 67 96 656 2 10679 0 9310 23 0 143 1 
+3,571 17 41 45 3607 1801 30 2611 88 7440 58 37 27 376 305 373 341 397 91 410 415 402 4362 98 286 223 423 418 36 4362 98 37 27 36 6 32 10942 138 3090 2701 0 62 18 8 28 1526 16 400 198 1569 139 113 1 
+1,6511 752 17 2288 7 38 294 3730 8 28 2 3566 0 2726 233 6 8723 4850 39 457 21 237 4633 20 1021 1107 15 3 2744 3444 452 6 20 433 15 5339 4886 1 
+1,11 3117 4 4141 4343 7 2739 75 65 13 0 258 2547 4 3 4681 8 10083 1 
+1,273 11 1871 2316 3535 1968 1522 55 3287 157 0 8416 4 133 1169 9282 54 30 1474 4995 2 0 9897 42 279 141 3229 647 11 133 30 5365 60 9 11292 8792 4 2018 3462 12 9567 4 3527 5 48 4630 9602 723 296 75 65 1 
+0,6052 766 13 51 5 194 158 187 2 2907 2130 1 
+0,6913 3878 1499 60 10 624 1637 1069 2185 15 3399 4 7980 5 2633 3523 1 
+2,5765 7386 31 5 2071 109 9 3 5323 8 26 2604 2619 39 2128 164 0 5159 8226 5 33 22 7 1191 2904 3191 1 
+1,7113 3238 644 8 3515 6 10342 2746 89 2693 198 100 22 7 67 655 2 35 3437 1125 5 217 3 47 407 22 7 1364 2415 1284 147 195 1 
+3,3026 1042 539 431 921 8349 7579 468 690 7505 18 81 487 207 1756 5719 2 27 3786 170 104 9 0 1728 47 1097 688 823 2460 20 207 10294 23 0 425 198 4 16 1319 5 326 3942 1 
+3,3082 14 0 2310 3520 5 6621 30 776 3 63 5529 106 82 188 0 2523 276 4 919 11 62 25 2748 1666 518 3537 14 0 3520 195 213 4 3 867 2669 5348 3 2310 317 18 1 
+0,4196 4158 3 38 62 2571 20 238 7782 1063 2287 1098 8765 3 1724 4 5604 28 10 732 6 1411 166 3 2008 4 805 1 
+0,3794 855 1204 13 4476 2 4373 44 1913 213 156 11074 13 4588 2 488 2 1227 6 981 6 3581 517 11261 1754 61 0 266 4664 11395 6 2772 5455 897 1 
+2,353 255 1709 1307 2776 0 1067 331 64 170 186 8 3 3682 164 1594 6 1358 3 106 82 1829 85 711 4817 1 
+3,6801 2858 4154 204 742 7802 342 6739 5916 97 8 28 3249 2564 16 578 218 9 0 63 72 3 592 10 8924 400 895 20 49 61 2257 113 8 3651 50 0 62 22 7 3481 11304 1 
+1,11 1923 4 2882 866 109 10 171 22 7 388 89 5598 182 4 16 196 23 6299 4061 225 19 31 2 783 21 578 259 137 42 85 76 148 155 0 1653 2601 1355 31 975 82 0 9800 10 10730 
+1,11 9466 5487 4 3926 7566 22 7 10192 2 0 119 56 610 2 1809 538 23 787 5341 141 1848 3551 10017 8962 204 823 1 
+2,24 1176 12 3 357 231 5 0 1618 1899 0 826 1585 1665 69 348 4451 9 1738 1 
+2,24 7366 2617 5 0 1050 937 2697 507 64 760 35 9147 8 10371 10271 9586 10471 9817 9540 6 5032 9 0 2171 6 1725 5 21 3782 9617 5241 296 28 1 
+0,2411 275 3734 9 3 357 10532 5 629 8 220 2586 114 897 6 1539 478 318 2 3733 50 133 49 103 2907 840 5224 109 1 
+1,11 466 385 399 240 8514 51 191 4 739 5728 374 153 2805 5 1728 1567 187 2 3 218 8 5332 175 2602 9 1567 296 20 0 1923 4 3957 1642 6 2925 6954 1 
+2,2932 5980 11420 0 1083 12 0 5070 997 5 0 740 47 551 4 41 45 6096 1903 15 0 1067 331 1239 0 3008 374 268 1 
+0,2962 2209 32 1474 0 4651 4 9984 10793 6 9253 9636 12 964 10957 2 3475 6 8163 2714 4 4734 1 
+2,3077 85 32 924 10 0 2880 6196 22 3478 85 34 0 214 4 613 1038 7271 799 353 3235 35 916 2151 5 1842 1681 3630 3 1431 
+0,17 431 369 1969 15 368 4 21 587 4697 8675 392 29 34 1209 2 5505 10135 54 309 0 156 2 9342 71 3439 978 9 1529 243 38 2624 2121 20 2120 8 81 1 
+2,11 1382 22 10006 2312 7907 18 75 11392 6 53 0 2199 31 2 521 9835 446 589 53 0 283 11 1785 31 8606 26 5297 79 55 2487 9 43 422 
+1,1597 4397 101 7614 7059 1286 26 174 3249 650 14 0 4397 259 3228 5510 81 4724 0 3140 4 26 4291 94 15 2182 650 83 9583 10403 6 4848 644 20 1374 11337 1 
+2,4356 1879 701 69 521 245 650 83 939 5 136 110 7 2050 137 7755 13 52 701 69 521 281 3478 71 730 1149 3289 1 
+3,2304 47 183 275 396 3 393 4 2694 132 35 10903 9 3 456 1373 11363 5 8 287 2 10434 62 3258 10401 8170 6 8120 71 2618 105 2 777 357 202 3485 10094 2 677 1098 1 
+0,851 5391 94 6 395 10 10157 1336 6 0 9404 10624 1336 23 8086 1 
+2,4163 6 16 1724 4 1997 897 4891 126 3383 12 217 9987 4677 4 0 151 177 6 0 2750 966 12 347 3505 20 0 56 1149 5 1544 6 1740 10576 76 241 3 5252 708 
+3,27 176 2333 9667 294 123 26 2236 55 2493 891 1084 14 5811 2 2139 0 1213 22 7 1231 15 529 14 0 9191 1076 8493 9 2513 4 16 439 3235 908 318 1 
+0,1643 91 11 512 107 13 4583 21 2440 228 1807 10 134 1451 85 10897 0 1719 4 1997 973 399 20 8376 0 2691 12 38 2159 1 
+3,125 7698 1984 229 7809 18 10 19 89 1015 5683 104 987 5670 104 9055 2 479 7631 0 359 896 1441 4 367 3163 1886 4283 2 3666 16 1792 8 0 56 1055 143 1 
+0,963 1089 30 9954 3232 44 5517 2 0 1613 1307 9812 232 604 3098 2449 494 9997 3963 604 300 5520 9674 14 214 
+0,7189 13 10494 16 264 2 878 8 48 257 1217 500 6 3389 663 11 1283 114 303 22 7 7190 6 500 1287 29 34 9215 53 3 7329 6 5848 1397 229 6061 723 1 
+1,11 260 107 25 172 105 2 804 3 176 946 1931 10500 2 413 689 5 1948 4078 244 9 1899 496 1038 7411 29 477 0 43 260 1120 2 816 
+2,24 3936 3872 320 122 1054 3815 29 1774 0 3301 283 213 4 3 9341 1471 10997 75 736 5 21 2050 137 83 2327 1 
+0,125 602 6167 1000 0 3473 565 4 16 3387 1280 84 14 48 4990 5473 13 0 92 7 5475 2510 220 0 62 18 1 
+1,17 6169 9 73 66 104 576 3265 5 298 767 2 11416 1214 1031 644 5394 2037 6 1743 5251 84 11424 9383 12 1212 8670 88 749 3103 18 8 81 1 
+1,327 378 6038 325 24 6281 5022 81 3 218 10 21 431 67 4813 89 90 7970 5 1966 20 1417 11174 42 3 8611 218 23 1966 18 160 31 141 3415 1 
+3,11 603 1644 313 6 1171 124 379 1642 2992 3129 2 4985 11047 8 56 2599 39 1360 656 2 157 1111 2014 3674 1160 5737 104 248 1567 1 
+3,80 93 1586 2974 11 6042 4 0 176 682 56 275 10 1025 175 588 3493 802 198 657 49 556 61 35 4672 14 135 710 275 75 47 3 596 18 81 1 
+3,80 93 17 6625 1364 1397 832 58 37 27 376 305 373 341 397 91 410 415 402 4022 98 286 223 423 418 36 4022 98 37 27 36 8 81 1073 0 258 2605 2 660 3 207 1854 1839 2 299 328 0 10930 731 2 246 312 5 3 1093 103 1 
+2,639 31 419 4576 102 4047 146 22 7 797 280 60 46 64 1514 262 73 5 0 390 109 1211 639 31 3 137 2 521 14 110 1 
+1,2913 2958 5183 0 461 1175 53 0 231 5 0 1618 1561 862 282 6 775 6316 3848 3 108 2 2162 2958 1552 321 2728 889 9 26 5691 1425 108 3848 550 26 73 5616 224 15 3 2096 6 461 355 0 2976 2898 139 66 8 28 110 1 
+0,2994 17 602 746 1847 39 3 48 162 2033 16 38 3387 2510 220 8 81 84 5010 3263 3276 6 2684 53 3 1251 986 517 61 48 4990 5473 1 
+0,2994 602 746 3297 255 3253 3223 8 81 11256 3 38 3387 292 10 9520 99 4 0 1292 220 22 7 9557 53 3 1251 1787 1023 10 4493 8 21 8060 2200 1 
+3,5808 58 5753 98 1063 1938 840 0 262 66 56 4932 8 81 147 3 10948 703 5 16 559 207 15 1743 5227 
+0,6946 4168 4217 2302 6 2395 2302 32 4815 53 3 338 10 29 3295 2395 2 4722 2001 1090 405 4217 22 391 482 2011 765 6 193 1 
+1,314 101 2423 2385 195 18 0 4847 10 1678 198 48 4430 5 752 3 65 380 64 0 235 4 812 1474 2 0 334 1626 1837 392 1 
+1,24 299 6473 5422 15 19 8486 391 9613 4 0 2655 1882 12 9152 2799 302 2 9333 1392 1395 8 81 132 0 731 10205 8 926 0 5911 6 0 2741 88 128 1 
+2,1 693 468 11 2342 635 7487 495 2 167 122 8 0 457 1331 40 1447 102 46 767 35 77 137 728 12 0 1594 3056 14 6814 1954 1 
+2,1 693 468 11 2976 2898 64 2067 3 1138 5 28 22 7 3482 14 59 1954 2913 2958 1552 321 2728 889 6 737 66 9 170 12 21 3013 14 0 1497 2 231 0 1175 2 3 139 66 224 1 
+0,6659 7775 6 7013 3960 32 1732 3 1755 2240 2 9808 710 9742 8726 4757 9 1251 1023 3687 9042 438 9 3 3512 4 0 481 1272 1806 1 
+0,130 18 95 10 19 13 8793 0 247 4 3 38 301 2749 193 1010 4388 23 0 119 565 4 178 544 1010 1613 5 715 2 1704 0 307 163 20 1568 1 
+1,907 5427 2 200 4950 5 349 8040 2 1715 0 1430 5 0 298 233 1808 8 315 20 595 1616 5147 1808 1 
+0,11 2237 1475 3 419 49 4858 6 9247 8 81 11 3592 1228 4 1496 4508 0 914 163 2763 2730 12 0 2439 4 3 5512 4 1778 985 2801 281 
+3,2293 2407 832 4328 22 7 63 199 696 18 81 19 29 8554 16 56 206 2293 2407 1177 828 1754 927 12 16 725 6 798 1984 264 601 48 103 1 
+0,11 193 4948 69 90 3740 2431 42 2197 30 348 19 51 14 3 877 904 61 9718 440 9782 1 
+0,7828 565 4 40 2045 8106 3097 2129 4 1554 9 5131 371 1 
+2,24 4265 6475 735 198 3 1458 489 891 12 0 939 6682 6 13 184 3243 287 52 1340 2 767 26 2128 1425 1 
+2,74 17 685 1318 1890 1608 6 1607 1639 32 90 3268 3691 20 2521 23 0 379 3820 3947 1347 21 317 18 81 1 
+0,1951 9 178 1951 9 7333 470 6 1951 9 500 1951 9 2358 4188 853 13 513 1951 371 68 982 35 367 2138 6 
+1,7368 5 2936 32 2552 8865 6 9255 39 5050 4 0 3241 4 459 4174 4968 5 100 1190 22 7 1974 2 7380 7628 7686 866 
+1,17 6903 294 7099 6513 3283 26 127 7 234 122 53 0 587 9214 8 81 1073 0 43 743 294 2 2144 3207 9 6781 54 64 1457 23 35 887 416 2 3574 11079 1 
+0,130 29 806 16 119 178 2482 544 1010 1613 5 1568 15 5153 1509 42 568 0 119 1104 992 163 788 15 4388 1 
+0,129 4409 6808 84 2095 440 519 8126 114 19 1971 2 4822 3953 22 7 231 6 3566 485 22 766 1 
+2,80 93 17 4074 6344 8247 48 5380 2596 2 231 0 1261 1064 1379 1587 2 21 1238 374 170 108 44 0 575 1896 2998 5 212 467 1653 14 3874 1954 81 1 
+3,80 93 1606 3 47 380 40 65 111 332 1583 537 6355 2347 4313 5346 0 994 399 240 6 341 766 20 10541 10 133 357 668 1282 2063 275 89 265 3605 3240 2437 3 111 1605 4955 2063 2 207 23 807 1151 1332 14 0 2548 4 10032 485 1 
+1,1616 369 30 4634 3 5162 9058 14 0 6847 5248 1540 5 764 1616 15 46 208 9 3312 39 3567 21 8089 190 1228 23 3 343 1838 
+1,4016 1286 616 70 8 3 3218 1683 238 269 6575 2871 1772 5804 2862 5 21 1661 4469 857 10 1861 48 4 26 4520 2862 22 7 1820 18 1 
+0,11 5975 2153 6284 1037 5900 25 5332 130 22 7 7835 374 372 8533 9 618 5895 7392 1232 0 10034 18 195 1 
+2,6705 7695 5365 77 526 75 110 2 4826 60 3 1538 5 551 6 2 804 0 849 827 1049 3 419 2505 53 10007 1 
+2,117 1253 2827 7898 181 21 4987 453 2 26 587 1425 40 1009 12 11379 5 26 43 48 10076 83 605 14 7786 6002 1 
+2,129 0 72 26 664 12 1581 1941 31 44 4070 89 9151 0 585 9267 4 1941 22 7 9623 2455 1941 89 9949 3 3757 2108 2185 14 180 0 1246 2185 5 330 551 6 57 
+0,1575 25 8800 0 62 22 7 43 4783 983 9872 6310 983 2684 488 2075 898 44 1143 983 697 9048 5 9943 969 1 
+1,744 4016 8065 0 3218 4 238 269 9326 8686 267 3806 2862 8 70 3749 48 4 26 4520 1655 18 1 

fluid/PaddleRec/tagspace/text2paddle.py

+import sys
+import six
+import collections
+import os
+import csv 
+import re
+
+def word_count(column_num, input_file, word_freq=None):
+    """
+    compute word count from corpus
+    """
+    if word_freq is None:
+        word_freq = collections.defaultdict(int)
+    data_file = csv.reader(input_file)
+    for row in data_file:
+        for w in re.split(r'\W+',row[column_num].strip()):
+            word_freq[w]+= 1
+    return word_freq
+
+def build_dict(column_num=2, min_word_freq=0, train_dir="", test_dir=""):
+    """
+    Build a word dictionary from the corpus,  Keys of the dictionary are words,
+    and values are zero-based IDs of these words.
+    """
+    word_freq = collections.defaultdict(int)
+    files = os.listdir(train_dir)
+    for fi in files:
+        with open(train_dir + '/' + fi, "r") as f:
+            word_freq = word_count(column_num, f, word_freq)
+    files = os.listdir(test_dir)
+    for fi in files:
+        with open(test_dir + '/' + fi, "r") as f:
+            word_freq = word_count(column_num, f, word_freq)
+
+    word_freq = [x for x in six.iteritems(word_freq) if x[1] > min_word_freq]
+    word_freq_sorted = sorted(word_freq, key=lambda x: (-x[1], x[0]))
+    words, _ = list(zip(*word_freq_sorted))
+    word_idx = dict(list(zip(words, six.moves.range(len(words)))))
+    return word_idx
+
+
+def write_paddle(text_idx, tag_idx, train_dir, test_dir, output_train_dir, output_test_dir):
+    files = os.listdir(train_dir)
+    if not os.path.exists(output_train_dir):
+        os.mkdir(output_train_dir)
+    for fi in files:
+        with open(train_dir + '/' + fi, "r") as f:
+            with open(output_train_dir + '/' + fi, "w") as wf:
+                data_file = csv.reader(f)
+                for row in data_file:
+                    tag_raw = re.split(r'\W+', row[0].strip())
+                    pos_index = tag_idx.get(tag_raw[0])
+                    wf.write(str(pos_index) + ",")
+                    text_raw = re.split(r'\W+', row[2].strip())
+                    l = [text_idx.get(w) for w in text_raw]
+                    for w in l:
+                        wf.write(str(w) + " ")
+                    wf.write("\n")
+
+    files = os.listdir(test_dir)
+    if not os.path.exists(output_test_dir):
+        os.mkdir(output_test_dir)
+    for fi in files:
+        with open(test_dir + '/' + fi, "r") as f:
+            with open(output_test_dir + '/' + fi, "w") as wf:
+                data_file = csv.reader(f)
+                for row in data_file:
+                    tag_raw = re.split(r'\W+', row[0].strip())
+                    pos_index = tag_idx.get(tag_raw[0])
+                    wf.write(str(pos_index) + ",")
+                    text_raw = re.split(r'\W+', row[2].strip())
+                    l = [text_idx.get(w) for w in text_raw]
+                    for w in l:
+                        wf.write(str(w) + " ")
+                    wf.write("\n")
+
+def text2paddle(train_dir, test_dir, output_train_dir, output_test_dir, output_vocab_text, output_vocab_tag):
+    print("start constuct word dict")
+    vocab_text = build_dict(2, 0, train_dir, test_dir)
+    with open(output_vocab_text, "w") as wf:
+        wf.write(str(len(vocab_text)) + "\n")
+
+    vocab_tag = build_dict(0, 0, train_dir, test_dir)
+    with open(output_vocab_tag, "w") as wf:
+        wf.write(str(len(vocab_tag)) + "\n")
+
+    print("construct word dict done\n")
+    write_paddle(vocab_text, vocab_tag, train_dir, test_dir, output_train_dir, output_test_dir)
+
+
+train_dir = sys.argv[1]
+test_dir = sys.argv[2]
+output_train_dir = sys.argv[3]
+output_test_dir = sys.argv[4]
+output_vocab_text = sys.argv[5]
+output_vocab_tag = sys.argv[6]
+text2paddle(train_dir, test_dir, output_train_dir, output_test_dir, output_vocab_text, output_vocab_tag)

fluid/PaddleRec/tagspace/train.py

@@ -7,90 +7,83 @@ import math
 import argparse
 import paddle
 import paddle.fluid as fluid
-import paddle.fluid.layers.nn as nn
-import paddle.fluid.layers.tensor as tensor
-import paddle.fluid.layers.control_flow as cf
-import paddle.fluid.layers.io as io
 import time
 import utils
+import net
 
 SEED = 102
 
 def parse_args():
     parser = argparse.ArgumentParser("TagSpace benchmark.")
-    parser.add_argument('train_file')
-    parser.add_argument('test_file')
-    parser.add_argument('--use_cuda', help='whether use gpu')
+    parser.add_argument(
+        '--neg_size', type=int, default=3, help='neg/pos ratio')
+    parser.add_argument(
+        '--train_dir', type=str, default='train_data', help='train file address')
+    parser.add_argument(
+        '--vocab_text_path', type=str, default='vocab_text.txt', help='vocab_text file address')
+    parser.add_argument(
+        '--vocab_tag_path', type=str, default='vocab_tag.txt', help='vocab_text file address')
+    parser.add_argument(
+        '--model_dir', type=str, default='model_', help='model dir')
+    parser.add_argument(
+        '--batch_size', type=int, default=5, help='num of batch size')
+    parser.add_argument(
+        '--print_batch', type=int, default=10, help='num of print batch')
+    parser.add_argument(
+        '--pass_num', type=int, default=10, help='num of epoch')
+    parser.add_argument(
+        '--use_cuda', type=int, default=0, help='whether use gpu')
+    parser.add_argument(
+        '--parallel', type=int, default=0, help='whether parallel')
+    parser.add_argument(
+        '--base_lr', type=float, default=0.01, help='learning rate')
+    parser.add_argument(
+        '--num_devices', type=int, default=1, help='Number of GPU devices')
     args = parser.parse_args()
     return args
 
-def network(vocab_text_size, vocab_tag_size, emb_dim=10, hid_dim=1000, win_size=5, margin=0.1, neg_size=5):
-    """ network definition """
-    text = io.data(name="text", shape=[1], lod_level=1, dtype='int64')
-    pos_tag = io.data(name="pos_tag", shape=[1], lod_level=1, dtype='int64')
-    neg_tag = io.data(name="neg_tag", shape=[1], lod_level=1, dtype='int64')
-    
-    text_emb = nn.embedding(
-            input=text, size=[vocab_text_size, emb_dim], param_attr="text_emb")
-    pos_tag_emb = nn.embedding(
-            input=pos_tag, size=[vocab_tag_size, emb_dim], param_attr="tag_emb")
-    neg_tag_emb = nn.embedding(
-            input=neg_tag, size=[vocab_tag_size, emb_dim], param_attr="tag_emb")
-
-    conv_1d = fluid.nets.sequence_conv_pool(
-            input=text_emb,
-            num_filters=hid_dim,
-            filter_size=win_size,
-            act="tanh",
-            pool_type="max",
-            param_attr="cnn")
-    text_hid = fluid.layers.fc(input=conv_1d, size=emb_dim, param_attr="text_hid")
-    cos_pos = nn.cos_sim(pos_tag_emb, text_hid)
-    mul_text_hid = fluid.layers.sequence_expand_as(x=text_hid, y=neg_tag_emb)
-    mul_cos_neg = nn.cos_sim(neg_tag_emb, mul_text_hid)
-    cos_neg_all = fluid.layers.sequence_reshape(input=mul_cos_neg, new_dim=neg_size)
-    #choose max negtive cosine
-    cos_neg = nn.reduce_max(cos_neg_all, dim=1, keep_dim=True)
-    #calculate hinge loss 
-    loss_part1 = nn.elementwise_sub(
-            tensor.fill_constant_batch_size_like(
-                input=cos_pos,
-                shape=[-1, 1],
-                value=margin,
-                dtype='float32'),
-            cos_pos)
-    loss_part2 = nn.elementwise_add(loss_part1, cos_neg)
-    loss_part3 = nn.elementwise_max(
-            tensor.fill_constant_batch_size_like(
-                input=loss_part2, shape=[-1, 1], value=0.0, dtype='float32'),
-            loss_part2)
-    avg_cost = nn.mean(loss_part3)
-    less = tensor.cast(cf.less_than(cos_neg, cos_pos), dtype='float32')
-    correct = nn.reduce_sum(less)
-    return text, pos_tag, neg_tag, avg_cost, correct, cos_pos
+def get_cards(args):
+    return args.num_devices
 
-def train(train_reader, vocab_text, vocab_tag, base_lr, batch_size, neg_size,
-          pass_num, use_cuda, model_dir):
-    """ train network """
+def train():
+    """ do training """
     args = parse_args()
-    vocab_text_size = len(vocab_text)
-    vocab_tag_size = len(vocab_tag)
-
-    place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
+    train_dir = args.train_dir
+    vocab_text_path = args.vocab_text_path
+    vocab_tag_path = args.vocab_tag_path
+    use_cuda = True if args.use_cuda else False
+    parallel = True if args.parallel else False
+    batch_size = args.batch_size
+    neg_size = args.neg_size
+    print("use_cuda: {}, parallel: {}, batch_size: {}, neg_size: {} "
+            .format(use_cuda, parallel, batch_size, neg_size))
+    vocab_text_size, vocab_tag_size, train_reader = utils.prepare_data(
+        file_dir=train_dir, vocab_text_path=vocab_text_path, 
+        vocab_tag_path=vocab_tag_path, neg_size=neg_size, 
+        batch_size=batch_size * get_cards(args), 
+        buffer_size=batch_size*100, is_train=True)
+    """ train network """
     # Train program
-    text, pos_tag, neg_tag, avg_cost, correct, cos_pos = network(vocab_text_size, vocab_tag_size, neg_size=neg_size)
+    avg_cost, correct, cos_pos = net.network(vocab_text_size, vocab_tag_size, neg_size=neg_size)
 
     # Optimization to minimize lost
-    sgd_optimizer = fluid.optimizer.Adagrad(learning_rate=base_lr)
+    sgd_optimizer = fluid.optimizer.Adagrad(learning_rate=args.base_lr)
     sgd_optimizer.minimize(avg_cost)
 
     # Initialize executor
-    startup_program = fluid.default_startup_program()
-    loop_program = fluid.default_main_program()
-
+    place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
     exe = fluid.Executor(place)
-    exe.run(startup_program)
-
+    exe.run(fluid.default_startup_program())
+    if parallel:
+        train_exe = fluid.ParallelExecutor(
+            use_cuda=use_cuda,
+            loss_name=avg_cost.name)
+    else:
+        train_exe = exe
+    
+    pass_num = args.pass_num
+    model_dir = args.model_dir
+    fetch_list = [avg_cost.name]
     total_time = 0.0
     for pass_idx in range(pass_num):
         epoch_idx = pass_idx + 1
@@ -100,17 +93,16 @@ def train(train_reader, vocab_text, vocab_tag, base_lr, batch_size, neg_size,
             lod_text_seq = utils.to_lodtensor([dat[0] for dat in data], place)
             lod_pos_tag = utils.to_lodtensor([dat[1] for dat in data], place)
             lod_neg_tag = utils.to_lodtensor([dat[2] for dat in data], place)
-            loss_val, correct_val = exe.run(
-                    loop_program,
+            loss_val, correct_val = train_exe.run(
                     feed={
                         "text": lod_text_seq,
                         "pos_tag": lod_pos_tag,
                         "neg_tag": lod_neg_tag},
-                    fetch_list=[avg_cost, correct])
-            if batch_id % 10 == 0:
-                print("TRAIN --> pass: {} batch_id: {} avg_cost: {}, acc: {}"
-                        .format(pass_idx, batch_id, loss_val,
-                                float(correct_val) / batch_size))
+                    fetch_list=[avg_cost.name, correct.name])
+            if batch_id % args.print_batch == 0:
+                print("TRAIN --> pass: {} batch_num: {} avg_cost: {}, acc: {}"
+                        .format(pass_idx, (batch_id+10) * batch_size, np.mean(loss_val),
+                                float(np.sum(correct_val)) / (args.num_devices*batch_size)))
         t1 = time.time()
         total_time += t1 - t0
         print("epoch:%d num_steps:%d time_cost(s):%f" %
@@ -121,27 +113,5 @@ def train(train_reader, vocab_text, vocab_tag, base_lr, batch_size, neg_size,
         fluid.io.save_inference_model(save_dir, feed_var_names, fetch_vars, exe)
     print("finish training")
 
-def train_net():
-    """ do training """
-    args = parse_args()
-    train_file = args.train_file
-    test_file = args.test_file
-    use_cuda = True if args.use_cuda else False
-    batch_size = 100
-    neg_size = 3
-    vocab_text, vocab_tag, train_reader, test_reader = utils.prepare_data(
-        train_file, test_file, neg_size=neg_size, batch_size=batch_size, buffer_size=batch_size*100, word_freq_threshold=0)
-    train(
-        train_reader=train_reader,
-        vocab_text=vocab_text,
-        vocab_tag=vocab_tag,
-        base_lr=0.01,
-        batch_size=batch_size,
-        neg_size=neg_size,
-        pass_num=10,
-        use_cuda=use_cuda,
-        model_dir="model")
-
-
 if __name__ == "__main__":
-    train_net()
+    train()

fluid/PaddleRec/tagspace/train_data/small_train.csv

+3,17 3034 3611 641 541 7 2531 1983 4 3708 4698 30 1524 2580 333 1 
+3,17 4232 1754 303 6085 832 84 25 3 5280 9 1005 566 11118 6 10010 1431 5187 5 0 650 240 25 5243 2127 16 8252 8 255 368 4 0 143 1 
+3,17 2402 509 96 1790 458 50 0 221 6 0 1289 9 383 30 302 2 2583 44 0 202 143 119 65 188 0 2035 4 0 919 2525 1 
+3,17 1877 32 4941 67 3313 3338 23 0 790 3514 5 1145 100 39 1113 1140 3 906 1007 55 454 1470 21 67 317 18 8 136 1 
+3,125 7734 92 67 96 11131 2701 6 10910 2793 3532 3 38 652 9848 2449 77 389 132 0 56 594 2535 1 
+3,17 3041 1097 3637 293 8 95 42 2747 391 3438 9 0 47 15 67 96 2757 421 297 2812 3 1072 5143 3 5196 1289 23 220 342 830 58 1893 236 1 
+3,24 5866 4 0 390 7 725 450 143 1282 1254 442 20 297 73 632 153 5 0 258 65 2 297 5735 3788 11178 0 6743 941 4051 18 86 1 
+3,1643 91 7445 191 4530 122 3 1074 5 833 6 38 582 9 5029 1420 442 75 65 0 107 18 86 3398 0 221 13 3395 23 3 9879 5375 1 
+3,6468 91 353 8955 3 7251 1361 5 7623 3927 5928 1940 1823 2 235 15 0 996 2096 14 3 1694 905 3305 303 14 21 578 1419 5318 4 297 825 161 7630 39 3 994 5186 2748 20 26 4726 2 703 82 8352 50 786 1420 762 164 26 3294 211 14 5694 1679 786 31 0 9266 923 23 366 2103 123 1940 1 
+3,80 93 17 3034 3611 641 541 7 2531 1983 4 3708 4698 30 1524 2580 333 1 
+3,80 93 17 2402 509 96 1790 458 50 0 221 6 0 1289 9 383 30 302 2 2583 44 0 202 143 119 65 188 0 2035 4 0 919 2525 1 
+3,3050 17 1929 68 277 1285 2 4777 10641 67 96 102 896 30 502 10081 20 66 268 104 2450 1016 106 8330 4 509 1183 7 1929 1737 18 136 1854 10 96 55 1245 1530 1 
+3,4064 17 7174 1929 67 2054 284 2493 1264 10061 2 4663 159 509 96 1929 101 7375 7954 18 8 33 1 
+3,273 80 93 17 11 695 9 116 58 7 1461 1215 586 183 394 550 82 2 3 10573 453 8 95 39 0 434 208 62 10719 3 8371 23 41 45 3607 1801 1 
+3,80 93 17 11 702 11192 4549 8 95 39 301 1314 3 159 41 45 733 2031 5 688 8424 617 2527 8 0 221 7 2702 6 16 1650 2 2041 881 359 2 2063 0 448 774 1 
+3,464 121 732 121 140 794 2 154 185 4796 3573 12 35 4878 651 69 34 10717 50 780 0 450 1326 1144 1 
+3,11 10346 558 4 2086 13 3 357 368 4 933 0 122 2 676 283 25 477 322 3 514 1768 3507 1 
+3,639 13 419 646 9 2475 5 0 202 143 455 276 42 485 5 1849 3341 1282 1254 32 1132 2 3325 3 1074 10778 155 241 213 46 928 555 517 61 0 3342 54 3671 12 449 1 
+3,11 56 733 2031 25 2053 820 2 3 159 3775 5739 15 67 1433 989 2599 293 187 2 3 258 2057 1 
+3,750 239 2400 55 34 4534 787 9 3 2765 857 2135 23 409 524 7776 3 490 218 582 1 
+3,11 1674 1736 2235 9 116 7 183 394 583 75 1482 458 44 21 620 12 16 4528 5 751 712 1 
+3,7207 2057 452 0 459 390 9056 221 2496 2 4934 42 160 30 2794 19 140 2737 198 789 5 0 47 1 
+3,6339 449 5 313 10762 198 15 0 127 4840 3 703 5 2038 6 2014 629 1 
+3,6735 1130 30 11179 2 289 139 20 0 227 633 764 696 42 0 2622 4 1854 1462 0 10381 6 10986 896 1 
+3,11 439 4 1679 505 38 6 63 1255 2473 442 1530 44 0 421 339 103 3 38 1026 25 280 1 
+3,227 689 7 764 696 1084 999 1130 20 3 158 5173 1125 2 170 139 5 3 645 2 988 449 5 0 221 1 
+3,428 695 4 400 5 116 0 932 208 705 84 55 34 9204 9 15 281 15 5697 1545 422 8 95 1 
+3,949 958 400 1245 39 1709 194 158 2686 310 0 303 5573 0 131 158 29 436 1245 807 4 1449 1 
+3,469 4 0 2657 11091 2661 8 0 1182 2376 2083 4 7016 75 65 3628 321 2557 6 265 3785 588 1 
+3,2285 2441 0 6693 9 3 5064 2 556 9 590 39 49 61 311 161 2157 23 1430 1841 298 5 681 315 1 
+3,11 62 320 0 466 1488 3520 246 20 3 4863 7983 13 2 337 16 10395 9 1521 3183 1 
+3,3061 5927 0 3723 9600 1725 9 1233 0 6484 8515 1798 13 900 2 9203 26 62 8 0 202 143 1 
+3,754 1168 123 19 13 1131 2 2150 21 876 73 170 104 2450 3 106 4 67 53 0 143 2 154 5299 2213 96 1 
+3,27 229 554 20 0 7805 7 6381 105 2 2742 5662 5716 8 1719 39 457 48 344 190 9728 5 754 1168 1 
+3,468 7407 4889 5 5386 1182 2542 2 235 8 0 642 1887 6970 213 4 3 850 2735 5348 1 
+3,5896 842 14 7171 58 619 35 2549 990 29 8730 1256 2037 10799 211 9690 35 878 25 90 8 49 5115 5085 1 
+3,78 928 10939 366 645 2 479 400 4 116 58 5 0 220 208 62 7 239 695 3672 586 183 394 757 55 815 60 2 34 0 447 306 71 0 1982 306 1 
+3,2967 1986 4969 1140 51 5 8935 6 10705 658 8 109 7 191 2220 1745 1679 2546 23 11167 6 10786 8291 2 8402 6 8448 6 8706 0 139 113 2220 859 12 1994 203 53 0 1323 1 
+3,1362 519 13 3 5557 4 16 621 1828 757 7 0 4632 4 574 97 724 5 2859 54 32 2743 217 3 2027 4634 482 1153 14 574 6 135 275 50 230 1717 235 6 309 2536 519 165 7 2006 2 203 10 440 519 25 477 0 156 7 9650 968 1 
+3,1176 3 9288 68 5092 51 1038 6535 31 3 3161 11066 102 52 8535 130 42 15 52 4938 19 2 9334 5 505 493 6 4966 8623 52 8296 573 11087 11226 12 26 11077 6 10107 60 3527 166 7561 775 2 8455 4192 0 9689 5007 0 3472 8034 2 355 0 438 52 924 1 
+3,7728 0 7976 4 894 2797 851 68 99 911 0 10062 2552 20 2009 1201 5915 1635 7609 87 204 351 14 0 9651 5 0 179 7 902 5128 211 160 7 141 794 2 911 26 4978 573 3116 9 10491 4 0 249 358 761 5580 1669 3190 10 10046 7818 6 5764 5704 55 333 34 4892 0 917 87 1602 1 
+3,101 152 25 90 595 10 0 56 221 25 735 0 8679 87 636 594 1998 3031 149 2306 329 5 0 2791 4 40 162 7 902 588 218 10367 10 19 31 49 166 3 41 815 1 
+3,27 6998 206 193 62 13 5438 339 238 1717 360 77 677 1767 9 2436 4647 83 35 3294 310 934 8 714 3 2488 264 1 
+3,940 3039 5 2354 284 34 590 3 1074 4544 111 5357 4695 4869 3428 5039 6 1307 4989 30 1124 9 0 1178 2833 3862 3 257 643 4 0 5399 12 3784 161 5 150 3366 450 1105 2 2354 75 65 1 
+3,940 3039 5 2354 284 34 590 3 1074 4544 111 5357 4695 4869 3428 5039 6 1307 4989 30 1124 9 0 1178 2833 3862 3 257 643 4 0 5399 12 3784 161 5 150 3366 450 1105 2 2354 75 65 1 
+3,27 2088 83 2314 4389 0 238 612 4 2317 6202 10442 1667 8 120 2368 4389 2620 50 26 62 7 4878 15 19 9505 1 
+3,752 13 333 1446 15 3 5442 42 0 1132 25 517 2 277 12 1488 2797 61 12 67 1 
+3,2345 326 7475 0 38 612 875 4 830 349 50 830 7 11158 2736 5 1234 175 6 933 52 2722 3 3162 193 2702 348 422 1 
+3,577 0 127 5 794 4 437 6 2171 94 3280 2 452 82 35 5579 2368 7667 5228 3 2781 5381 780 0 107 425 2770 4 9979 1 
+3,7381 6288 13 9360 39 5277 3 710 386 9 0 63 158 1 
+3,464 299 328 8293 185 1262 198 230 68 121 200 185 786 62 556 1 
+3,11 62 7 345 452 10 19 7 76 2581 99 44 1150 92 1606 99 4 19 10 1114 69 830 1 
+3,353 3 2230 47 1594 7 1556 8418 13 122 2226 2 3 3450 8822 6 3 861 10 13 4705 0 4965 9436 1083 
+3,406 30 629 49 8 11176 60 0 1293 225 46 1832 1369 7 3 711 14 0 38 2796 4 7982 
+3,117 5658 401 9280 89 3 481 9975 10 865 0 2504 165 7 50 2 167 657 8709 
+3,6095 9132 13 4527 3054 11358 30 1994 19 2 790 3664 
+3,262 135 5262 3118 68 4584 126 2102 1428 11051 2695 166 6128 1040 6 26 356 803 1983 
+3,6574 3 2781 1298 1058 2 11414 121 68 548 3 994 5186 9 1448 10822 794 
+3,961 7 300 30 2600 4716 5 496 15 4799 5 0 92 2917 5868 4723 0 1696 3564 1450 20 195 7 9820 84 1263 281 293 1449 61 600 32 90 404 2 961 7 300 1558 898 6 46 1558 19 210 1 
+3,3076 19 7 237 72 5 0 1393 4 0 1195 6 10 2099 0 3465 30 10578 263 5 426 718 2450 780 15 281 4910 15 46 68 310 5043 35 2589 82 14 0 3662 9319 8 10 8451 3 4718 800 9761 206 8 0 126 1280 42 10 4706 9114 1004 49 166 3 6430 1484 3230 12 3 7770 6 4071 8421 3201 1263 3 1848 237 47 9312 1 
+3,4383 537 1924 1444 0 7212 11 133 1424 10103 6503 89 2 279 0 668 0 671 707 491 10243 6 0 381 99 964 51 2 3 1424 10 89 419 619 4 10406 1 
+3,1352 3 10771 4 634 6442 466 952 7752 6 716 140 8719 50 1524 35 1757 5367 5 5225 2134 5 4667 6 5152 5 0 1404 455 7985 477 5091 8728 173 9187 6 10564 4590 11110 9315 6 9501 10125 20 2633 5378 0 9882 6 9290 1151 9 0 1560 4771 3523 1 
+3,11 2852 13 3107 6 566 5 0 1195 15 9064 20 40 8115 4 0 7318 2852 1983 6871 5813 10140 20 2009 6144 6647 15 3 4673 2 26 2239 8961 264 6 2504 4862 1 
+3,6555 771 3662 1329 2 94 27 106 5 0 361 4 3 4139 1195 206 264 4650 5234 10 121 722 6036 1326 50 7100 7879 8294 6 235 99 8 0 398 1786 1 
+3,11 38 3792 1026 4 0 356 936 5494 2512 25 118 3414 82 42 19 7 2020 14 0 176 12 6124 6969 5 1955 118 295 23 10144 4866 951 956 3849 6 7314 5 1955 6 5918 5 2924 5929 57 4 0 43 2 456 6 2090 1336 1270 185 1517 1 
+3,469 4 593 38 4571 5 0 3792 7573 4 7326 1026 13 9 356 936 264 179 353 3 4676 4 276 1398 5049 0 5157 3343 20 3849 1955 6 951 956 6367 185 456 6 2010 6 200 185 5559 2500 1384 38 4571 1467 356 179 9 9748 6 356 5494 2512 1 
+3,78 1106 51 5 111 332 225 239 4436 5382 49 10042 2509 15 8046 4436 64 10596 2 10568 0 8492 10677 525 1577 2 9711 5545 3510 51 0 1155 8717 5 312 1678 203 168 2 111 332 6 2049 9212 0 419 8367 102 46 1106 597 357 9 0 1002 8596 3413 1 
+3,3943 182 604 1508 3318 845 1125 2 0 393 4 239 8638 322 15 574 2317 1397 6076 6 3028 54 3324 82 1982 4840 9000 20 9021 3028 13 3 62 2571 8 0 10398 10 604 1508 30 76 5435 123 3028 6063 2277 6539 1 
+3,2838 1065 58 25 3377 0 746 5867 465 58 875 2167 9 3877 7 7822 518 4 1243 663 114 2 781 2838 7 671 700 229 1 
+3,2835 40 65 11260 21 3715 565 4 16 9444 1128 8536 175 1547 14 0 2013 6 3445 2602 2835 6032 2627 1777 4756 1128 5163 360 974 10088 1089 8125 2013 1149 6 9113 1767 911 6 1766 301 11108 0 361 4 3 1128 187 2 2835 134 1 
+3,412 11 41 2339 7 212 2916 755 7117 25 11044 604 3582 2910 58 2 672 143 1113 399 15 0 782 5151 9230 274 12 3 9797 139 153 374 66 153 1128 2 1334 16 385 193 1470 1 
+3,1176 1027 116 7 48 882 775 21 620 2 751 712 5 0 3457 4 16 377 2564 0 3023 1414 0 62 7 394 2 222 274 11 2465 1247 0 620 12 0 3023 6 4668 1775 5 19 1 
+3,428 620 12 116 7 1693 882 416 60 5 0 545 660 4 751 140 2033 0 62 7 377 1945 2159 10529 120 1098 68 148 310 2137 2 425 202 9 0 43 72 1 
+3,27 365 523 23 0 1927 2452 602 20 9692 3 871 175 5 499 7 258 143 3139 1351 2978 240 11429 759 23 8788 62 130 3178 7313 2068 6 49 1 
+3,949 958 29 556 5687 104 9 0 260 62 1351 4347 1460 99 60 2 4475 300 2068 1632 1122 175 9344 9 264 567 3673 2068 6 49 1 
+3,4371 105 2 488 3 175 10 85 1411 15 3 1853 11014 1288 9725 300 2203 5 2229 14 99 526 11 522 85 804 190 860 11316 6 440 9783 11373 300 8820 129 6356 3852 1 
+3,27 38 156 1070 1252 1315 3 2193 5 1015 208 5063 5 0 119 339 103 3078 0 3540 8027 116 7 5229 102 19 1460 183 1 
+3,5973 30 0 9448 923 8 0 2371 42 30 46 9859 12 2772 3548 5210 469 4 0 182 557 6416 91 13 2436 803 5062 2332 19 0 2796 4 0 618 129 3926 7746 1 
+0,17 4378 4402 3 419 173 8 81 335 0 1739 14 0 489 3951 693 85 166 2 732 19 31 213 4133 1189 741 506 60 10 106 6 138 1100 4 0 880 5364 30 5244 614 2 494 3 913 106 2 521 19 1 
+0,17 27 229 4 193 275 360 960 2926 58 4335 98 4278 4277 2993 6 3871 97 3842 236 8 86 18 46 29 3543 3 38 567 1283 1320 51 2 169 526 0 1146 4 0 545 1104 1 
+0,17 324 1059 8 86 18 19 105 2 425 3 363 2141 470 2766 363 1468 6 2644 3383 54 2434 2 2191 51 9 3 47 4 4736 156 175 1 
+0,17 27 229 4 381 1096 1275 18 8 70 46 1664 0 4900 9 3 38 1104 4 3260 10 68 1541 339 526 0 301 4 2879 14 0 398 439 614 2 396 3 585 283 4 11 4311 8 57 4748 1 
+0,17 11 5119 4 120 737 1862 9 0 114 240 5 589 688 102 191 5404 14 49 61 323 1424 668 275 13 76 278 337 2 256 3644 1658 0 194 158 13 217 4940 44 1 
+0,24 11 1387 3370 3321 9 1236 943 2045 3225 114 25 1662 918 333 427 40 72 1224 0 5067 8 602 746 58 7 567 365 3663 193 1 
+0,17 11 1650 2 871 1081 5499 4743 44 1077 190 897 2 344 1123 13 1672 3 3564 5 499 1 
+0,4340 1056 430 739 250 
+0,17 1388 97 191 2522 8831 1461 4635 385 322 15 84 275 5260 1710 8 1388 7 264 114 395 6 0 493 4 0 1710 30 362 2 34 181 183 3 512 2084 18 8 95 1 
+0,470 379 3938 167 382 544 9 38 307 163 5368 184 256 3226 1 
+0,2957 602 746 58 29 520 16 63 466 725 1541 789 40 162 5 0 681 466 179 4 7228 19 18 86 1 
+0,24 299 328 7 1033 2108 584 209 2567 20 2759 6 1140 118 230 2725 3 727 19 184 13 2 3528 3 731 7 5016 427 278 12 99 0 258 1022 6 5245 193 1 
+0,24 577 49 61 186 161 242 360 1571 2987 1625 2247 780 1141 3 679 5148 0 152 610 7 309 4 1523 2432 40 237 47 13 1524 3 38 700 5 0 5528 1520 197 727 6 2676 1 
+0,24 27 238 4773 4660 2 4603 14 0 4212 4172 1874 1633 31 2030 95 2 48 1590 1218 54 64 335 0 321 54 770 5 0 1532 1091 3978 73 1161 1 
+0,24 27 314 2000 806 4768 12 0 587 201 1540 136 1994 1102 599 1103 6 135 5024 2 0 48 343 314 126 4688 3 201 317 18 1 
+0,24 3810 0 8218 4 0 8397 288 5 10589 9337 2565 6 425 617 8211 6413 2043 8 9205 10128 3591 9291 20 1858 23 5167 8302 1 
+0,17 2418 7452 1106 9522 1726 188 3 11070 449 3652 10 1311 0 2515 3310 16 8370 20 401 526 42 0 9142 8237 3433 1726 6 770 738 724 18 8 70 1 
+0,17 7360 3465 64 735 53 5603 405 3 3355 3702 2 2457 3 257 4535 2491 41 45 724 147 8 70 1 
+0,17 845 94 30 2461 2 34 0 361 6 10804 4 0 622 6 249 19 3610 30 138 4770 1 
+0,17 429 7582 18 3 560 20 245 135 668 85 10772 15 2762 3019 1 
+0,17 260 242 18 8 70 46 89 1518 5175 2 4609 785 4804 9 1770 308 5 120 46 763 2 34 0 43 322 1764 2 34 1738 5 499 1 
+0,472 91 428 4838 4 314 724 582 2 32 280 654 10 21 8049 10811 89 562 2 277 12 3 514 2551 44 7595 5 5649 2897 5 4473 6 2485 32 174 18 0 791 2455 31 1685 20 3 201 2173 1 
+0,472 91 27 5124 285 924 2 4946 2485 6 4473 210 1663 2 34 168 2 1496 597 11 10280 2801 30 2194 61 2933 6 55 34 15 1155 15 7332 38 10008 3678 1 
+0,472 91 27 3329 986 4 2190 1321 13 2235 6 10133 1491 70 53 382 86 1009 5871 1099 0 250 1390 1773 1815 2 34 57 4 0 356 11301 4 0 578 653 5 678 103 1 
+0,472 91 963 0 4845 4 0 7084 6 135 1496 5 593 914 163 13 3471 9080 0 43 1833 9 1218 6 5305 4434 13 2 2429 3471 167 2 426 2512 1 
+0,24 4251 68 477 872 4414 757 7 3393 2 1270 242 148 6 25 1148 3 174 72 2 2690 837 48 2209 20 409 6 260 724 452 0 3111 4833 0 398 8585 988 9 8531 15 94 277 1014 46 491 3707 9435 1 
+0,472 91 2271 2779 44 0 2559 4 0 1181 608 1205 6 12 3 257 5008 76 934 270 3800 4287 236 4088 775 0 222 274 28 9 900 3 3589 10670 591 1 
+0,472 91 5857 3804 4 7041 960 25 147 3 2503 75 736 4 35 10323 853 1625 1308 1 
+0,24 11 1222 11031 10256 2523 25 90 1444 23 3 1001 4 3167 9 0 512 9003 1537 1001 213 242 32 3211 0 10575 30 141 1003 1154 1 
+0,472 91 469 4 0 133 8153 2080 8 0 1181 608 1205 3628 198 991 40 65 6 2542 30 734 2 10177 198 0 1298 11 135 77 2080 767 2 3483 5136 1 
+0,24 847 4558 19 9936 19 10851 19 6 4891 19 211 1285 709 0 10347 9762 11361 84 25 477 3 2704 5185 631 134 14 0 7273 4261 212 7906 7428 192 3 603 8241 9760 8 19 1 
+0,472 91 270 724 32 5462 51 12 3 2119 4 4302 4368 1250 2 1226 3 10969 10 3560 8057 0 92 7 199 500 206 525 1 
+0,24 7177 8495 6 8037 5856 4449 32 477 249 1759 122 248 8136 5 325 2028 380 10 46 141 1003 615 5129 5130 6 32 1732 3 10968 292 10 10813 1344 1480 242 148 1 
+0,472 91 2304 430 102 232 94 30 11284 5 0 127 225 2192 30 4703 0 356 788 1773 1815 2095 16 4453 11 578 1390 1773 1815 15 19 13 358 1509 2 396 8 21 1210 1981 986 40 47 1 
+0,24 7442 5434 5 3347 3 9892 11371 10 5575 51 8 3 1671 5 2741 1171 0 616 4721 18 86 1 
+0,24 27 2155 9219 9124 1822 9 0 43 72 5 413 324 13 4817 8267 2 7001 7 7857 1 
+0,472 91 6522 2153 8378 30 0 182 1295 2545 5 0 2237 3689 10064 678 4912 6 362 10723 0 8268 4 3 1222 3381 1 
+0,24 11 2689 10 3 5513 731 6 3 1464 427 71 2135 48 3386 427 55 454 312 8 0 398 106 13 562 5094 148 46 32 795 1137 1 
+0,24 270 7 2911 915 3675 8013 16 868 40 65 15 19 2589 122 926 498 12 3 2717 4 914 1562 3496 0 201 354 18 70 1 
+0,472 91 6460 3 3199 6637 498 13 3 1695 3647 4 899 7724 6 10047 13 15 1333 15 3 3199 3418 8 1239 757 7 50 0 150 4 3416 9 242 54 10229 0 1129 4 593 899 7 3562 1 
+0,24 1602 1255 8 3 5486 2764 6 26 2554 8 3 220 1023 1933 846 254 4925 2753 3819 4094 1337 21 4462 4 1523 2080 2 1297 0 5556 1857 4 0 1193 573 13 3603 9 0 4401 4 361 1 
+0,472 91 2878 6056 3055 5518 5305 658 14 235 8 1381 32 11160 11063 5559 10 10461 0 2689 10 9870 361 140 4834 8 0 2155 899 1 
+0,24 605 13 5293 16 318 2 804 3 343 2 0 1281 20 3766 15 0 591 85 439 3 662 4 450 6 5610 930 419 5 626 0 420 201 354 18 86 1 
+0,472 91 435 3068 3230 0 1203 688 268 1314 51 15 3 2006 1222 4775 2 0 4712 4 667 4 10745 263 0 92 2444 9996 562 517 10009 11 3110 4 5440 2699 498 2516 20 226 73 113 9 3 547 552 1 
+0,472 91 462 426 54 1558 2 4846 1381 42 8406 3733 9 3 915 2 494 203 160 270 242 32 10446 3 1341 10 2627 0 996 183 208 301 6 1466 23 1503 3461 1 
+0,24 11 41 45 2311 755 8 70 1802 4819 1216 4 3 318 2 10217 3 10917 3648 4 2350 4336 2 1212 120 5594 134 1228 0 10389 11186 5 324 1 
+0,472 91 469 4 0 10197 4 10880 13 9998 6 9014 0 1850 8551 10 1321 986 5 4703 2192 3054 9452 488 2517 2243 654 4 230 5411 2224 11291 1 
+0,24 407 1738 16 43 1764 9 785 4610 70 2613 227 689 8 0 295 1239 4 1824 1077 308 84 13 5288 20 0 152 610 6 84 232 242 763 140 231 2 38 2775 9 3 1303 4 3262 1 
+0,472 91 27 10176 1308 806 282 2 1427 3 297 169 104 2272 853 1625 10349 9 3536 9175 1729 2 0 1239 4 201 13 4952 51 1 
+0,24 1369 7 3 2515 1456 10 1290 68 8114 11 11069 2418 2712 5520 737 164 21 9102 449 3652 3352 217 139 2681 3 106 1 
+0,472 91 27 295 201 3602 25 358 2 1129 0 5546 4 785 5385 10958 10 1526 1218 10066 23 498 13 10051 597 2015 6 0 727 1305 13 11182 1 
+0,24 634 1218 30 5 0 131 5402 4 2137 9 3 1476 201 591 10 29 1783 0 3359 740 526 132 2168 2 498 3 150 145 1284 147 86 1 
+0,472 91 27 10966 229 2435 10861 14 498 25 2582 2 323 526 0 493 4 593 899 6 25 0 623 2 11239 3 216 914 731 1 
+0,24 27 5074 5150 10 313 89 1124 2 516 40 47 55 279 1459 1703 2135 631 119 47 71 789 213 4 3 2064 5350 6 799 1236 0 1297 7 385 2575 1229 313 7 201 354 18 70 1 
+0,24 842 590 53 474 10358 6 1691 1079 32 280 10226 23 15 657 15 496 44 111 536 6 0 3830 1 
+0,24 2897 23 499 6 496 10991 73 570 47 266 10478 4 21 1660 8490 5 4029 81 6 18 46 64 3210 0 10284 8095 64 14 1306 23 9104 2 0 2537 1 
+0,24 117 3 224 9 4819 3 512 2084 1309 81 10 0 152 610 8404 1079 0 2201 1694 9185 793 1771 132 0 11195 46 2474 68 1426 0 2526 3596 9665 1 
+0,24 270 18 81 19 13 1485 274 12 105 2 804 3 3588 2 0 908 4 0 4428 1181 608 1205 1 
+0,24 6094 2550 985 3 5182 1265 3 2105 1303 5 764 6200 3145 2 32 0 92 7 4709 10949 11302 703 14 217 73 2817 1452 3 1523 9569 147 28 1 
+0,24 11 612 3602 14 0 212 5849 5 1354 25 1000 3 2170 4 0 2526 8340 522 39 0 769 4 3 186 162 266 2526 1 
+0,24 7044 14 413 939 846 254 30 934 8 3 234 2 49 7988 3451 5250 5 809 1217 1 
+0,24 27 79 348 3 363 1996 10865 14 0 297 169 104 2272 853 1625 9 798 1476 5385 2211 3 1138 33 102 35 1308 9793 6 2053 39 2190 517 61 73 161 1452 5 0 474 1 
+0,24 7355 20 3 3559 5333 23 4178 6 0 6370 7367 1347 1909 6823 6598 123 52 387 10422 26 669 2 3 512 7691 8502 8771 9 3 514 7969 9889 4654 12 5496 3737 6 11273 1 
+0,24 1877 1014 89 141 3186 42 2 703 2697 1684 155 46 11369 69 1209 2 2474 3 2215 132 0 10887 4 9735 9097 211 3546 290 0 127 2600 30 1679 1328 72 1472 2483 1684 3107 20 9529 5532 10387 405 35 2294 10300 1 
+0,24 1870 779 0 2239 393 4 9755 11205 32 9967 197 413 939 6 312 40 283 6 242 32 141 3311 9 0 703 1 
+0,24 27 62 10 1847 0 92 7 43 3194 3170 217 77 103 380 210 123 19 13 1131 2 453 9168 2662 9 3194 10159 3 1284 147 86 1 
+0,24 27 38 1235 1312 308 5554 29 34 1209 2 1426 94 2 2265 113 4 0 1780 4893 4767 2505 61 0 3321 5826 10 5181 0 596 4 10648 11299 1497 11067 6 1235 1780 4687 44 0 421 531 103 1 
+0,24 11 1399 8481 9249 25 90 280 5 2741 1405 657 796 16 2140 788 1303 11 4441 89 795 132 90 147 9989 4 5934 2292 98 938 50 1864 1480 2 0 3648 1 
+0,24 842 29 309 1412 193 795 132 3252 3146 0 1312 15 46 495 2 8869 38 4687 4501 6 10162 5 3 169 106 4838 2 0 4015 4 2361 7 4709 10393 1 
+0,24 11 2311 755 9070 0 3286 4 11385 8 829 1822 10087 5 9641 3 1124 708 5 5069 5 0 7598 4178 187 2 3 5071 354 4841 54 3746 8 0 318 1 
+0,24 3047 9296 140 32 280 3 234 2 3528 4788 528 132 46 2584 20 1776 0 3110 4 9862 166 11443 6 4664 5 10945 599 391 8957 1817 242 18 70 1 
+0,24 1387 242 54 9101 0 1235 1857 4 0 3830 2376 18 86 35 3328 427 360 120 1976 2 34 38 1537 4 2565 6 10862 427 55 34 404 2 801 2637 2533 3747 1 
+0,24 11 393 4 276 10 8235 2007 71 720 2794 213 4 1500 560 1530 5 1161 416 2 49 5250 1751 1776 6 11139 2201 21 2543 229 18 8 86 1 
+0,24 7886 599 2224 5 0 764 9032 1193 75 162 140 9525 0 453 4 3 38 1900 4185 1 
+0,24 11 2283 915 7 591 2 1629 25 1804 3 38 5246 8246 263 0 5301 899 6 280 10 9732 5 16 3123 13 10012 5 1439 3502 61 19 510 102 270 7 7866 3337 20 5 0 382 2252 242 18 1 
+0,24 6523 4 5189 5 3908 32 90 10751 0 92 7 9895 2133 4 1487 1560 10554 10432 0 2133 20 50 57 779 5 118 49 61 3 47 3 257 8624 5151 18 95 1 
+0,24 1957 10684 644 82 0 960 3198 5 0 421 48 389 30 1333 42 651 69 1478 160 30 49 10685 61 5135 966 0 1671 71 10 46 30 776 8306 2637 8263 6 135 511 148 1 
+0,24 165 7 3 3425 232 419 9454 55 1117 23 2448 3152 7629 256 1486 2 135 11440 10130 82 1879 1108 8694 166 678 135 8265 1537 2089 35 8968 5 0 5130 4 135 3152 54 564 3324 6 2152 0 8693 8472 1 
+0,24 11 297 169 104 438 2 804 3 798 1476 1308 53 201 1823 590 49 166 3 282 86 102 3 327 79 172 105 2 516 16 1308 77 276 39 21 126 229 5014 2 504 903 9 0 2272 853 5211 1 
+0,24 11 254 4 960 6 960 27 855 2355 254 30 900 2 8637 120 46 1228 0 92 7 199 1835 12 0 154 4 3 2327 8379 7 297 73 352 104 9299 1 
+0,24 3017 4558 11124 6 5507 5 0 917 86 734 2 988 82 5453 4 5068 8818 8 0 233 5 120 3 41 98 354 358 0 1564 5431 7519 9598 5 49 61 3 2027 1 
+0,27 142 2433 720 8 6023 8801 230 1741 68 4631 300 4 1204 344 567 6 4371 9681 3729 4532 11 2433 349 50 0 309 4 6067 2329 10845 0 300 393 2437 3 4561 4 0 4201 544 248 0 3729 11101 10 0 1234 13 975 2 2477 35 621 3729 1577 2 4289 2396 3913 160 25 90 141 2166 23 0 3721 6162 32 90 720 10 1204 2363 25 10026 4201 544 2148 82 20 4710 5957 815 19 8 1 
+0,165 7 9611 10 130 55 154 1459 8435 16 8903 5 0 2467 143 20 2613 16 199 524 5 21 773 2 563 932 2201 3471 497 0 398 72 10 1411 85 222 3 174 234 5495 3395 0 932 5 996 9 1089 6 371 4434 1 
+0,11 7431 1513 3 1542 8 430 5727 358 7312 1536 9899 9759 10885 165 3610 10 160 7 9715 9952 2235 5 433 6215 2 4824 9351 0 3309 5031 9141 893 6 10497 19 12 3 9819 7 11370 3273 1 
+0,6900 8253 5553 674 1159 4646 13 3 11409 4760 639 13 654 614 8 505 1816 1604 68 10023 582 34 2777 6171 3 10170 12 654 5444 8656 9464 6 912 29 222 590 9 3 49 3135 646 4110 7 495 10 1 
+0,27 162 380 78 31 2610 2 0 1860 4 7050 6566 117 807 78 31 8063 6 11039 20 26 807 4875 6 365 3724 1176 155 121 3373 69 3375 4 180 121 928 1505 1137 26 232 1860 5 6526 4622 1850 365 3724 6 0 624 1484 6379 7690 4 0 7646 7061 2962 4 26 1860 9099 12 2243 2537 10 102 1148 1973 30 2731 6 5115 4040 243 26 9826 9347 455 2537 979 925 2 1253 3 1461 9857 2243 2052 573 795 8432 2 2150 2243 1834 6 2003 593 4976 50 0 2243 2640 1 
+0,11 431 3933 2905 793 8779 6 1128 290 21 474 1312 774 542 3581 174 1303 4480 454 11355 11 6788 7676 6439 771 76 3643 40 660 6 8837 1293 964 1126 8 325 7 1765 584 4 8026 5264 2442 11 92 7 3246 2098 30 76 1233 454 10190 10 3643 431 1011 462 40 1132 325 1011 2 711 2 135 5121 12 5362 3246 1011 117 40 507 313 227 689 6 1401 99 279 870 21 474 1312 774 325 284 1242 3 3499 12 455 5121 2 1082 3 454 9162 10 10622 0 2204 1011 4 870 21 474 1312 774 11 1420 4 322 3 3499 29 34 232 1 
+0,11 111 1393 7 1341 13 3233 1683 21 2081 3512 84 3706 2 4842 0 4445 10 1658 0 56 221 13 448 3 662 4 16 2480 30 184 9146 2803 82 11 971 1720 0 345 4 3 531 47 1787 596 84 25 3626 10 1658 1478 4992 32 1751 1468 11289 25 1751 9825 4573 232 406 2 3325 517 566 82 583 0 448 221 1 
+0,7382 10937 256 10632 78 3678 600 99 2210 9734 9 4644 7071 51 185 9139 78 486 141 5190 249 78 1273 2210 3 4865 4 366 4592 639 31 3 738 1563 1010 3870 3077 11164 281 877 61 1118 1950 192 60 57 106 117 3 10472 234 532 1305 8 0 1503 110 78 486 4425 78 651 69 2162 0 2447 1 
+0,1 1048 1089 1782 3434 163 2687 2 8735 0 5316 4500 4 35 969 462 2547 3 8838 55 2032 3 5086 163 3542 2 2032 0 9833 393 4 7708 2012 10 35 968 910 68 563 963 1486 40 10281 99 135 969 23 405 40 5547 668 11 43 10394 2 40 85 34 2 9951 185 5547 2652 4050 4 405 5086 121 85 210 309 366 968 1949 4126 15 185 3542 668 542 1860 4 868 42 210 121 32 2 3434 40 3542 8 10882 183 2202 9576 4126 6732 10114 3049 9298 366 968 7562 1 
+0,1 78 118 1339 2 148 2226 2 574 9 330 7632 6150 1068 1739 722 32 185 1411 925 4089 51 0 447 235 1 
+0,1 961 2337 7085 6 78 1108 44 2 6914 39 9774 2 1024 120 1757 4 4663 2586 6 500 9355 46 89 8 0 9103 4893 4192 818 2 167 5 0 4772 249 78 9164 933 76 78 486 3 1074 4 3 10293 9242 435 121 491 1332 138 142 9 2629 160 30 3633 225 19 387 34 1160 19 117 3633 4 3317 598 780 0 369 20 185 808 387 34 3 447 2073 78 732 600 68 99 2434 10 1332 2629 9 1285 1114 69 1160 19 542 127 31 2571 8 2629 39 0 260 64 4404 35 4617 3526 2304 2777 126 284 8469 35 2629 6 732 174 6 658 50 118 2778 203 44 9 141 1132 1 
+0,1 78 486 916 2 5258 10 366 1990 13 3 1074 4 3 72 3441 462 2547 78 68 1082 3 1990 3302 2 3736 5117 50 562 10 4374 2584 5 0 618 4161 545 1926 1990 3302 13 3 447 2547 78 2804 40 9 5117 15 281 15 9 0 94 54 2154 366 1990 71 548 792 1557 116 78 1270 9 1829 10 1926 8 3930 13 549 2 8269 792 3645 5 0 618 6 78 486 549 2 3345 7873 2412 78 486 549 2 4 868 222 2 116 71 366 8033 6 208 9 1926 6 564 548 366 1990 1294 78 928 1711 40 50 3 1093 526 210 6 16 3598 792 3 7715 4 72 1 
+0,1 11 468 1402 3924 1926 3924 3392 601 3930 2625 3 662 2 34 8826 1177 82 46 651 69 1467 19 12 3 934 4636 1068 32 2 1082 185 621 4825 10000 8602 1640 200 5085 2803 0 3212 1092 69 1651 6218 2652 249 121 32 2 5078 1242 6699 6494 78 68 1156 10 46 387 76 1558 2 672 7197 3212 6 10084 57 71 48 1926 805 42 46 68 672 3 8951 4636 10 4493 3434 335 99 762 805 639 13 4 868 0 183 5341 1001 42 121 32 2 116 9 19 1350 121 722 794 2 277 292 13 1242 3 1001 4 805 910 72 9976 9324 910 5650 3772 5678 530 1 
+0,1 78 928 90 3 357 4859 4 4115 9 3 310 210 42 3373 69 5099 245 544 44 9 57 764 1132 11 879 518 4 544 131 2202 1186 2092 1186 3356 138 668 13 4439 11214 6 3256 2 235 12 2365 94 309 4115 12 3 2092 206 8 0 8500 1631 600 85 1505 1024 2202 1186 2092 1186 3356 3487 8100 8566 9144 9751 3981 4381 13 4439 11300 27 662 4 1089 10706 40 2 2202 1186 2092 1186 3356 3981 1428 211 40 184 2625 1336 12 265 6 10121 9037 1646 155 600 55 118 2703 19 2 2202 1186 2092 1186 9271 1 
+0,1 1586 27 5071 8149 4 334 1626 294 2377 1989 2349 5449 2 754 5850 134 81 3 754 6736 1923 317 18 211 19 13 3709 120 2174 155 245 6833 334 4030 140 32 89 5 245 1551 644 213 141 183 647 32 90 1247 83 180 11 754 107 5 3 679 296 20 16 1241 5 433 358 334 4030 7 2214 0 258 2517 1134 4 16 1765 57 162 488 4 9723 2 1551 2217 542 13 778 78 619 40 722 1322 2 556 82 6198 3642 2 1836 10 651 69 9607 598 1604 10447 3826 121 2671 1667 152 610 1 
+0,1 462 138 1132 78 795 296 40 544 78 1088 19 310 934 8 7160 6 118 9231 78 732 165 7 21 3316 9 2367 10 1415 121 2 1024 99 5316 9074 12 585 10868 5498 10 30 5530 20 2367 1606 3405 19 6 155 185 544 9285 21 5530 9073 19 29 452 51 5 0 9757 6 8 0 4645 7846 4945 9 2524 245 9372 2367 700 1 
+0,1 837 10 0 127 13 8193 5 7418 9993 138 1779 30 5204 10 121 29 1242 2676 4391 121 166 2 34 267 4 0 88 128 78 85 34 11231 78 486 21 8154 429 600 99 1270 10 13 562 94 124 69 1651 1646 85 121 277 155 3031 329 760 121 2 34 26 2241 267 78 85 1129 26 10620 3943 121 417 222 2 4596 262 78 816 366 2246 7 10619 1631 121 5382 1793 6019 78 651 69 1427 4451 211 366 11020 85 34 5 10 2518 78 8016 0 1723 10 0 764 983 4 6020 11060 9609 9887 6 9767 9653 6 8570 1631 2627 167 40 1324 573 7 21 3 1 
+0,1 4245 754 1168 1586 469 4 754 1168 7 182 1339 1006 25 735 2069 53 0 476 0 43 677 2215 2 2214 243 3 57 162 107 3109 172 75 65 7232 1348 4191 393 820 8 754 1168 7 182 1339 1001 4 823 4406 2937 6405 7 488 4 3109 84 31 181 75 65 187 2 754 1316 28 1348 4191 54 735 2069 5 8 33 110 31 3 264 2668 4 7581 1348 7590 0 43 334 1626 1772 2 1651 0 488 102 52 735 2069 5 95 435 78 43 1311 40 78 31 722 1335 78 924 10 19 85 4580 231 2 49 598 155 46 118 2090 812 82 0 1 
+0,1 744 100 1586 7036 4 100 7 1752 107 209 21 10004 4 489 28 118 552 39 0 88 128 1305 0 390 7 10808 48 276 274 4 2176 4108 20 100 7 1752 501 244 5888 2270 1094 1772 4 0 38 107 2127 3 1255 8 0 6839 6 1509 2 2185 12 10732 6 9499 269 9193 3431 0 5572 291 203 100 837 12 311 517 3899 7372 3834 1872 210 6 167 3 513 7892 4 1188 6270 488 241 762 5 3048 6790 754 1168 1061 3048 6 1183 1351 2477 60 593 38 7374 1910 6817 0 2086 5141 9 552 1 
+0,1 1943 25 172 10 6772 29 34 296 5 3 1439 9806 61 245 135 1943 1128 5 4835 851 694 0 4481 4 0 811 544 9 0 1048 1342 2822 8 5027 487 851 30 2082 1089 5 0 3307 4 0 1048 1342 2822 3904 1116 0 1048 1342 1508 8 5027 487 6 871 0 811 544 9 0 983 1048 1342 2822 5549 0 1048 1342 983 11277 6 0 1048 1342 522 4829 11 257 923 292 13 10 46 32 3547 0 983 11278 243 3 5903 1764 6 278 32 0 3881 762 6570 891 1943 1068 491 1108 5 0 674 2518 1369 7 0 241 1298 851 30 10475 0 811 544 9 0 9618 983 6 5549 1508 5521 1 
+0,1 78 722 619 1943 1092 69 330 3037 1048 14 4068 40 47 165 85 34 3 11085 649 6 5339 1749 0 1150 240 1351 19 85 1749 1943 6 78 486 1829 35 2167 614 2 5258 40 6896 619 10 1860 1 
+0,1 27 547 276 380 329 1025 263 40 758 4396 0 152 1589 581 1274 245 75 3614 4 8766 2127 3 8877 758 358 11 6403 4 149 329 7 636 1936 15 0 790 2560 8 16 1341 5967 6 568 3311 0 758 1293 3801 2919 335 426 1722 11 152 1589 581 793 1795 40 758 82 4 16 1341 11 309 4 3801 2919 20 245 581 10223 71 622 13 2732 1862 4381 4 868 13 3 5102 78 737 6 3270 0 372 84 13 118 5430 8362 1058 3 5084 20 245 2099 6 160 30 2727 4 2919 5 160 42 241 23 0 7096 2495 23 3 547 389 122 1369 7 0 1298 11 1 
+0,1 691 81 1589 5342 5 26 2174 15 267 4 0 1947 1071 5 0 8446 9 3 3509 1527 27 864 1276 12 1946 3003 1047 6884 7867 0 3559 2295 8 0 6792 1359 1073 21 8128 50 1589 7 1840 2 6611 408 21 587 704 399 8681 6 101 152 7 9634 9983 11 9077 1097 102 1589 2656 138 8702 2432 6492 5612 18 0 343 54 13 3 9390 23 0 5206 4390 120 0 9923 2634 25 2 148 50 40 165 1114 69 10297 6752 3 4906 3527 1 
+0,1 435 600 406 43 611 593 580 4702 31 4604 5 3725 600 2620 23 0 3153 2052 12 2937 947 4042 10 0 182 1812 1028 2 1086 13 2784 0 7987 4 597 281 558 5 0 1742 4 21 3968 287 52 34 3 2937 71 3 267 7234 9550 126 8896 4 8589 558 25 930 419 2 277 12 0 8450 71 10151 4 0 998 54 5592 10 558 165 13 0 558 787 10 793 34 8636 3181 4756 6 4566 4492 5 715 2 2047 0 3681 4 4906 117 4415 593 882 11054 1336 10 183 1451 13 0 182 3239 9008 4 1086 213 243 0 674 8489 19 1 
+0,1 78 43 3375 50 40 23 57 4 0 1089 4 0 246 137 2401 54 813 792 10 160 31 3 2017 4554 5 26 968 19 404 2101 674 39 3347 19 3 216 141 141 10 3371 2 235 4187 8 3923 42 85 76 235 243 178 225 2101 10 13 4905 13 362 2 34 10784 51 20 255 320 968 674 503 11 11088 8 0 178 79 64 549 164 1850 557 969 1152 203 2 200 1829 46 3746 4187 42 2401 892 8701 847 147 40 2 0 178 1089 54 8863 2401 3658 164 19 5 3 8764 280 0 4554 6 964 730 544 10 3181 155 2401 31 320 6 155 19 510 1 
+0,1 7961 11105 82 4 7555 8 0 1372 660 78 32 2 148 102 78 43 2154 10 78 217 442 82 366 4581 5043 78 31 5474 230 5430 5773 455 1372 1739 211 600 99 1270 10 130 13 585 4 3638 94 54 863 50 120 46 491 2524 1631 562 722 1092 69 8586 292 4105 72 78 3181 1372 1158 69 278 337 2 2560 10110 12 2367 7 2468 262 10233 8286 141 11024 8359 4825 2888 0 1372 79 722 76 1270 120 35 671 13 1279 7302 165 7 1461 362 10 46 118 651 69 1270 11 4509 660 292 13 10 130 395 68 69 3318 6 46 68 69 5433 130 25 531 682 153 5 0 696 639 1 
+0,11 6021 1931 6 0 507 4 0 4745 183 999 
+0,1 37 347 36 2991 37 347 36 2392 3372 2 4513 
+0,1 37 347 36 4111 37 347 36 7417 988 1319 164 1290 229 
+0,1 37 347 36 7335 345 37 347 36 3870 306 9 10513 37 4800 36 7430 37 4800 36 9673 
+0,1 37 347 36 6164 37 347 36 4004 426 10960 5 
+0,3079 220 897 55 1144 34 320 169 526 877 61 46 277 210 1 
+0,11 850 7 758 2244 1136 44 84 487 2012 68 34 2509 15 585 1146 973 1 
+0,27 192 4 6521 9869 446 13 1010 15 0 182 9558 137 4 0 47 14 3 4870 5 4284 1 
+0,130 13 5418 94 776 783 4 3 257 142 527 1557 487 206 992 2187 525 1 
+0,6300 170 1460 8 1310 5 0 850 14 3 72 4 2163 1697 44 598 5 372 446 1 
+0,3006 5 0 850 30 629 49 72 6 450 549 481 123 2011 2244 7206 1 
+0,11 4220 446 4645 55 34 1088 53 3 5050 1330 9 978 123 3 7181 781 11058 1 
+0,4005 0 193 5 422 9 151 250 13 21 59 1833 5 787 1 
+0,428 9012 1022 5065 163 1971 2 154 59 142 9345 5253 49 1512 2 4805 1 
+0,27 1342 9425 2223 55 11154 230 511 1538 2568 10725 6 135 9367 2522 1 
+0,746 446 68 5226 1298 10798 6 79 974 5 978 148 446 240 511 1 
+0,7948 94 30 2600 2585 2 479 10181 3363 71 1111 871 1232 23 0 487 3 1026 1315 1 
+0,11 6093 137 13 3 216 4746 10 9148 166 3 10215 2769 5 12 0 2058 1 
+0,7258 2386 78 17 4378 4402 3 419 173 8 81 335 0 1739 14 0 489 3951 693 85 166 2 732 19 31 213 4133 1189 741 506 60 10 106 6 138 1100 4 0 880 5364 30 5244 614 2 494 3 913 106 2 521 19 1 
+0,80 93 17 324 1059 8 86 18 19 105 2 425 3 363 2141 470 2766 363 1468 6 2644 3383 54 2434 2 2191 51 9 3 47 4 4736 156 175 1 
+0,2393 17 130 97 37 1187 236 36 18 19 29 504 803 3 2206 198 363 439 565 4 16 178 609 307 163 5 0 1446 896 4 2924 1380 6 1955 5 715 2 5393 0 309 4 3209 6 1707 193 896 1 
+0,1185 1163 17 27 229 4 381 1096 1275 18 8 70 46 1664 0 4900 9 3 38 1104 4 3260 10 68 1541 339 526 0 301 4 2879 14 0 398 439 614 2 396 3 585 283 4 11 4311 8 57 4748 1 
+0,80 93 17 11 5119 4 120 737 1862 9 0 114 240 5 589 688 102 191 5404 14 49 61 323 1424 668 275 13 76 278 337 2 256 3644 1658 0 194 158 13 217 4940 44 1 
+0,2393 17 130 97 37 1187 236 36 296 8 81 0 258 565 4 16 114 9 3284 6 10037 481 10168 6 1466 2 11041 53 2245 984 9 481 8393 6 10169 1 
+0,2393 17 116 58 37 1598 236 36 8 28 333 2460 0 393 4 400 19 105 2 425 5 16 586 183 394 595 19 29 660 66 289 104 400 2 1965 58 37 4393 236 36 2 3618 3 2088 44 193 404 2 986 1069 1 
+0,3064 9 371 220 873 2098 6 10509 651 69 8010 9547 178 371 4 0 2714 3410 5 7981 0 156 568 3541 142 3452 37 1677 36 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 11 433 1623 37 1668 36 37 1731 36 1 
+0,755 1935 66 959 9 807 1971 2 1826 2242 2249 2467 9414 6 2803 20 360 142 2056 10 94 32 2 1654 6 4416 8 35 621 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 3020 3005 37 1668 36 37 1731 36 1 
+0,6435 0 7652 13 292 3076 52 55 34 99 19 1545 13 9 57 2523 1785 2 10782 3 481 3509 4 26 4504 10230 2210 19 2 0 7653 91 434 729 6 715 10152 5197 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 6893 4377 37 1668 36 37 1731 36 1 
+0,497 6027 5964 5 6110 8 70 50 3 1093 1143 1290 23 2372 1405 1734 14 2585 784 2 9890 2043 6 3275 1350 64 8900 71 11381 1119 4 0 6407 8449 4 7272 7914 7278 1 
+0,464 19 7 719 2 3436 562 118 213 19 55 816 11132 5539 0 433 233 9200 9356 23 0 2192 6 11005 5421 53 3586 564 2334 7529 13 10775 1 
+0,941 1000 0 580 2536 695 4 202 5 641 541 551 95 42 680 10 19 55 279 894 5053 23 3 751 712 620 5 84 138 8144 4 0 156 208 705 7 898 64 10079 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 11 433 1623 37 1668 36 37 1731 36 1 
+0,7820 7282 3 297 47 266 114 4813 23 829 8760 631 0 75 1482 1 
+0,429 3931 1389 6972 1545 44 0 637 1642 1174 8 28 912 7 302 2 1226 8 0 4547 3104 192 20 189 10262 7766 1047 7103 1 
+0,11 1169 7 38 2640 3738 4461 3720 0 7679 5583 249 281 10 19 8732 0 1303 4 938 2806 388 2000 1656 10 1426 19 6 243 3177 980 200 19 4986 9 0 2674 2 494 82 3 38 218 9 1360 280 1 
+0,5966 97 172 109 10 19 3619 48 5045 83 6761 98 7916 16 612 875 7 2739 5590 0 62 89 543 4 629 667 4 987 8 2473 6 905 3305 2 3595 3 1755 2240 52 1513 249 52 55 3550 19 9 3 1986 370 1 
+0,1373 5 0 62 10 7 734 2 979 82 15 4696 15 3 4557 55 548 1328 1259 3 4883 4 5561 155 0 882 4 116 58 767 2 277 1329 166 452 51 5 751 712 263 0 398 72 35 62 13 549 183 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 1855 91 37 326 36 37 572 36 1 
+0,156 4477 1713 58 18 95 19 3093 3 352 113 1319 5 2502 21 367 2486 4 2481 1069 6 2569 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 1036 378 37 326 36 37 572 36 1 
+0,11 2079 3456 8282 8 0 220 14 6778 6210 7 4726 5 0 7492 1947 1931 6024 7210 366 6560 121 491 1725 1 
+0,3006 29 794 2 2043 49 11294 6 517 1479 213 4 2531 599 1325 242 148 1 
+0,2390 3867 123 0 1207 2317 6463 13 590 447 9 16 5380 2776 263 0 92 568 5264 1 
+0,27 79 4 260 2550 54 30 5294 0 1825 4 3 4372 2672 32 1516 7757 2846 1 
+0,11 201 1532 7 4848 1103 5460 4984 5 0 1161 1590 1091 13 2 167 16 621 1680 1 
+0,11 1181 608 1205 25 1475 53 3 4553 4 883 6 2530 256 9541 20 3 738 285 7 3496 1 
+0,7222 2498 1695 2 0 1882 25 4489 2 11093 9 16 9451 148 56 2637 242 1 
+0,27 7923 5116 13 1457 2 337 9 678 528 310 3 11001 4 9430 4499 13 5275 1 
+0,11 578 1390 1773 1815 55 672 3 2198 452 583 3 1709 453 1 
+0,27 10322 84 2592 708 9730 5 2249 1122 619 9 38 7274 7 6 3811 7 2775 1 
+0,850 11125 3298 511 30 11061 51 12 270 2 548 2775 9 3262 122 8 498 1 
+0,27 1304 6 3091 922 9 7707 55 265 9538 1130 263 0 92 148 511 1 
+0,11 43 2165 20 260 242 2 4609 785 4804 25 90 1738 20 511 1 
+0,27 8897 513 5613 9 9968 3152 13 396 14 1306 20 11288 54 265 198 0 8322 2563 1 
+0,4035 1858 5 0 3767 6100 29 34 49 2608 49 3348 6 1003 56 511 148 1 
+0,842 5 0 88 128 548 3 234 2 815 9703 3465 53 5603 405 3355 5472 1 
+0,11 3307 4 5587 23 133 9713 1760 8950 53 10918 11009 25 90 11143 5 9239 8956 7143 723 1 
+0,270 29 2553 2113 4 0 2417 7409 7027 4151 7719 164 0 157 4 250 5 1118 4 3 624 2165 23 0 212 7204 6 3831 1578 7118 1 
+0,117 0 119 311 103 2269 29 2052 3 791 386 4 16 659 11196 15 483 1559 1415 455 3719 10458 4 2483 3279 5038 2 563 2061 6 135 11295 10 29 4780 4437 5122 2533 1933 846 254 490 247 1 
+0,7443 32 3626 10 0 1203 68 34 2167 9 14 182 241 3 809 368 4 0 1559 44 0 75 323 742 103 7017 7327 4307 490 247 1 
+0,435 19 698 2 2061 474 1500 13 76 21 9028 5146 3374 963 8859 474 10852 290 710 1217 4 0 111 536 233 19 7 49 1810 5 0 2741 368 4 0 88 128 254 4 1208 2959 490 247 1 
+0,963 3 3561 3180 1691 140 4437 0 3208 4 1487 1208 7 2061 8898 322 15 5069 436 29 521 3 2017 2174 5 230 455 11017 2533 1079 44 72 254 4 1208 2959 490 247 1 
+0,117 3 596 12 9503 9 230 413 126 3703 387 5286 2 3 3180 1691 5105 385 8541 20 2299 254 724 10454 3 3740 4406 11099 10 232 4 0 3214 7 5508 1537 5099 3561 23 0 1235 227 15 4921 5295 14 0 157 4 0 75 6704 5801 2299 254 490 247 1 
+0,6385 166 0 778 6315 3865 4 0 5651 9493 5 11 6579 4 7930 6 5273 15 3 5503 653 9 667 4 406 64 3570 2669 4 281 991 8742 4 8934 3343 20 10426 5 0 233 187 2 38 2209 20 3 5114 9571 308 79 554 20 2299 254 2299 254 490 247 1 
+0,842 30 8985 8 3 1128 84 29 4846 230 483 1559 13 2514 57 4 0 92 7 182 4765 2533 2498 10433 254 4 2984 490 247 1 
+0,11 6958 5008 4400 270 7 3052 1022 1999 40 2777 648 707 4 1962 3015 8 430 459 14 66 531 7817 1 
+0,1612 2470 2751 8550 198 0 808 5595 4 3 1485 861 6335 1390 9864 55 396 8 3 3521 452 39 5441 8 430 3754 1 
+0,469 106 1218 55 11444 290 0 914 163 5 10832 10812 1604 29 35 3158 8006 2 361 3480 3 11208 168 225 3662 87 6187 444 87 561 1 
+0,27 3235 1239 3821 351 727 9094 215 291 5 0 4286 4 7781 13 184 320 195 1 
+0,497 0 3491 3430 4 593 914 163 3 914 3623 2796 13 50 2 10082 270 7 7865 73 915 1 
+0,435 2283 1516 1629 691 688 5693 19 4704 164 3 774 5 1629 7 10560 1844 469 4 3480 727 2080 2700 3 4895 4 1805 2530 9375 1827 0 915 1 
+0,111 1659 4 4901 4525 9526 10 4419 37 1750 36 2418 2712 37 1750 36 8854 1106 4849 3561 1512 2699 4919 493 1 
+0,7897 9240 280 12 1155 8953 4525 2710 9 0 43 72 230 0 8072 4 5587 6 4770 1088 35 8005 5526 885 1 
+0,117 0 38 1484 37 1750 36 3927 6247 37 1750 36 3 2234 4782 1840 622 8205 2 3 4581 6 4889 503 165 7 358 2008 8204 6 763 19 71 76 19 7 3 905 1821 1 
+0,7772 7 578 1390 1773 1815 13 362 2 34 3 2198 452 4 2190 1321 11445 290 0 110 3635 187 2 2444 1 
+0,4264 1858 427 441 71 49 3660 71 293 427 30 9243 4 0 1780 1014 924 2 34 5459 11033 813 20 10610 211 210 242 32 1022 654 10 0 791 11272 30 76 241 905 42 436 9120 1695 1 
+0,1646 510 0 43 7215 2043 1904 9417 30 10642 1660 11092 2 548 60 1 
+0,2831 5189 23 315 32 90 10374 6 3254 5129 940 633 1393 7 5594 5285 9 103 837 12 0 154 4 3 3458 41 45 8625 0 6039 3523 13 8130 6 1842 787 2 615 122 1 
+0,7191 1219 898 4814 9749 1479 6 4441 10605 30 118 3 547 4 0 1329 10 3616 0 1660 177 23 0 5104 446 123 0 2447 4 37 1750 36 11 7130 177 37 1750 36 1 
+0,2317 1559 13 1838 2774 2512 3747 1646 7 49 11165 1328 30 8659 2 19 6444 60 120 121 68 277 50 19 1 
+0,4383 510 3 2350 4336 643 5171 592 4780 589 75 47 27 38 218 123 9778 1235 2454 2757 11271 3344 0 2105 2 627 1 
+0,577 21 9545 4 8475 2140 404 15 2621 3111 312 7 1884 9 6080 6127 6084 13 5503 53 0 199 8474 10616 5 0 92 427 965 1491 1 
+0,6928 30 8565 2443 12 10187 1102 3625 6 769 211 46 30 436 9562 120 140 34 0 119 2796 5 9729 8547 8186 525 1 
+0,270 428 11241 314 2000 806 1474 51 0 465 608 1633 40 1009 2 1704 965 77 5488 4 1102 1103 10090 599 6 1325 2 0 1133 3480 11 2922 4234 818 4683 8180 4768 2 0 8030 2679 4 0 7966 755 7073 14 459 5618 3 486 1 
+0,129 5898 6799 7247 831 24 27 238 4773 4660 2 4603 14 0 4212 4172 1874 1633 31 2030 95 2 48 1590 1218 54 64 335 0 321 54 770 5 0 1532 1091 3978 1 
+0,270 11 276 4 2521 1005 1262 860 140 1811 166 1660 551 9 182 2122 5 413 324 42 5 838 77 1218 6 3 327 8901 29 922 0 258 8591 5 174 2520 1262 860 405 3 2781 5526 2621 11 1650 2 8600 174 2520 782 863 322 15 11072 6 11075 2758 55 34 257 2 9786 0 11367 4 618 10810 6 10521 2 1770 4805 8 0 465 608 1633 0 1281 71 1381 1 
+0,11 1725 7227 7895 1798 2466 4 11 1407 4 0 7927 13 50 2 246 0 357 1023 15 2058 9914 7642 6 3918 1225 0 4018 2313 7896 4602 122 53 0 557 4981 3826 600 249 8506 195 76 34 8946 1 
+0,1588 853 4252 2375 27 270 1588 853 2153 2375 707 25 1804 3 1696 4 678 9582 2072 883 3195 5 0 800 4 3455 11 10967 1588 10816 2708 181 19 719 2 8890 998 4912 23 0 791 3195 4 2072 883 1 
+0,1181 608 1205 117 40 1333 707 270 7 1181 608 1205 8413 3 1515 3725 4 0 4578 9033 4 3 9293 3 883 8431 8422 20 0 5411 1562 6 2608 11217 5246 23 3 2072 738 285 1197 9294 30 3133 2734 9424 10572 10 453 60 15 8363 5 11318 71 10657 2173 1 
+0,20 4032 3863 117 40 63 6 131 368 4 0 620 2353 1720 230 4041 140 154 242 1156 0 10040 4 361 8 498 278 155 19 1092 69 4731 361 8 3057 5869 4134 3057 13 0 241 1281 5 593 914 163 12 21 3123 6 19 13 0 10033 4589 10 25 90 8832 5 10 3123 10 25 3650 0 4981 4 1789 242 166 2353 1 
+0,6953 6803 2915 11 10686 707 417 1148 4 3 2530 3263 263 255 285 25 1804 5426 5 0 3263 84 30 5360 4 11248 1496 2889 1 
+0,7000 608 6459 1884 691 430 459 3761 270 7 1588 853 2153 2375 1286 16 10965 8911 9 0 43 72 2437 4578 853 2153 1118 2 673 0 3476 7 9896 542 57 809 1147 9 0 3476 10324 2 34 3 239 3421 9 727 15 1588 611 16 591 2 2188 38 1118 8 3 1031 5118 2237 9603 2 0 785 1721 1 
+0,6557 285 8526 30 166 10828 8427 4 1118 8543 4 667 4 1321 9772 53 3 201 241 3 547 1093 1118 103 290 464 0 498 10512 601 3 9309 2008 593 110 3635 85 34 4433 12 350 4 1321 49 3161 61 7606 1 
+0,129 45 7512 5891 605 24 605 13 5293 16 318 2 804 3 343 2 0 1281 20 3766 15 0 591 85 439 3 662 4 450 6 5610 930 419 5 626 0 420 201 354 18 86 1 
+0,603 608 1347 3955 10363 4 201 1859 11343 6153 25 985 11321 4 4794 914 3449 173 1210 0 498 11 10970 2574 11163 5 455 5426 30 1505 11199 35 234 53 0 2300 9779 4553 0 9780 1 
+0,20 2917 3863 2337 2353 3538 4 1789 727 6 10171 14 0 254 4 1353 7 4121 6 7328 4101 5 7797 1353 25 174 90 9131 20 1629 7 199 1281 3057 117 40 43 368 4 0 620 2353 1720 120 242 619 2 1117 23 4041 1 
+0,11 7057 7890 13 3 3719 4765 5127 464 121 491 1463 2 548 6336 1496 10 140 4946 361 121 1273 794 2 3470 0 208 7659 30 3 447 422 2 453 213 0 8947 3260 263 1321 10817 738 1496 1 
+0,129 6321 6807 3025 171 24 634 1218 30 5 0 131 5402 4 2137 9 3 1476 201 591 10 29 1783 0 3359 740 526 132 2168 2 498 3 150 145 1284 147 86 11 516 302 3645 40 162 29 2078 804 3 1476 4683 0 7587 139 53 21 10067 1783 10 14 16 4613 29 34 1864 1480 23 498 0 4112 6224 147 8491 10476 4585 353 3187 0 2042 678 526 0 806 29 1242 21 1783 14 50 5672 1480 23 2042 0 218 18 1 
+0,4072 4238 4100 7756 276 132 16 659 626 2 498 12 270 7 43 5320 23 201 248 0 3821 3461 0 2911 915 3675 2490 16 11207 11149 9799 497 459 1566 7825 1402 139 1566 3 486 1 
+0,638 24 27 5074 5150 10 313 89 1124 2 516 40 47 55 279 1459 1703 2135 631 119 47 71 789 213 4 3 2064 5350 6 799 1236 0 1297 7 385 2575 1229 313 7 201 354 18 70 11 313 3804 6390 1347 71 6755 296 3 218 2 3 107 145 8561 4843 302 1703 2 0 516 4 0 3756 104 4121 27 1297 1 
+0,2955 24 27 314 2000 806 2632 12 1325 6 873 3154 82 23 0 3847 4670 5 2343 8 70 1108 9 0 587 201 1540 3 314 201 317 18 11 4234 2355 682 806 209 82 14 73 5619 3 486 1 
+0,11 578 1390 1773 1815 13 1080 6 2567 148 19 55 34 11255 447 1612 2470 2751 198 0 808 5595 4 3 1485 861 4618 1390 6334 55 396 8 3 3521 452 39 5441 40 65 1 
+0,1 24 270 7 612 13 3718 26 1181 608 1205 79 2 490 274 12 105 9 3 3589 3578 591 2 0 4428 3476 595 4110 7 222 1809 0 1181 3800 4287 236 4088 123 52 29 4468 1360 9 450 2 7984 0 891 573 874 19 29 494 50 73 153 2 73 115 153 2 1707 6 516 3 3588 2 200 0 2654 5537 2 553 0 557 1835 320 6 2 167 19 60 4 1783 1014 16 235 13 164 1 
+0,4360 1040 27 38 5095 9 590 14 0 3208 4 2485 405 4933 206 5468 25 90 1088 20 8468 14 4360 1040 7437 23 0 9234 4 593 914 163 0 9790 4 2485 2577 8525 50 230 0 498 6 135 1496 1732 1 
+0,27 38 707 23 270 7 4313 608 1205 1315 0 10695 8987 4 3 3280 285 6 5 35 3457 3 5118 8916 5343 1805 11 3280 285 13 368 4 3 1789 9959 358 7115 5679 435 3 285 166 593 621 1203 1673 2 145 60 4 1103 16 983 10711 6 2591 51 8305 82 0 285 7 3491 9702 1 
+0,3802 8 230 2 167 3 447 110 7 10763 6 1220 618 1256 2232 1601 6869 4 3636 68 646 49 61 8936 19 68 4658 2 3 393 4 782 799 3034 1151 2044 4 2622 4 3636 1467 3396 5 1421 1126 6 2628 4 10919 9441 187 2 21 971 5 0 430 660 4 0 6627 4034 6897 1 
+0,129 4086 6293 2322 2389 1601 6536 925 138 8852 54 30 734 2 2093 3738 564 5366 122 6 9743 2 0 1701 5816 4450 54 7 8 3 4738 71 14 214 57 10 7 934 13 3217 26 71 189 318 13 0 356 1 
+0,6149 154 5267 0 2300 1691 20 10440 5440 53 201 11111 8673 0 1830 435 4614 3502 1079 46 9913 0 2300 704 3130 5 815 6 2224 8 0 2300 1830 1 
+0,129 3813 2909 2322 2389 1601 4105 47 7 4996 2785 31 657 23 3 3504 664 83 0 931 10 10718 94 42 19 2656 49 2148 23 0 3390 61 511 89 2140 924 117 930 738 978 0 492 31 280 2 34 352 113 2 2813 113 8966 5 5208 4996 166 3390 84 13 3 2216 507 4 4894 10 1158 69 866 5 3 2621 1 
+0,429 0 250 1635 177 1489 167 2235 95 12 21 587 2466 4 0 714 8441 782 511 1099 0 151 446 2 5006 4671 10251 2 477 49 4412 211 511 8430 8062 3497 10656 2658 284 76 1269 53 3 38 1317 3094 568 10956 5205 1 
+0,129 6805 6852 7500 98 1047 24 963 0 1327 3224 14 6822 1194 30 8223 8 3 2072 1211 1290 553 3278 82 5465 14 0 10738 3495 168 2 99 0 106 7 1653 1 
+0,129 2944 2980 273 24 4251 68 477 872 4414 757 7 3393 2 1270 242 148 6 25 1148 3 174 72 2 2690 1 
+0,129 3813 2909 2322 2389 1601 3911 2 557 4502 76 99 2086 54 30 8052 2 3404 5414 10053 35 10665 845 94 2798 8053 215 203 5 1451 4 4499 11342 6 0 166 15 978 184 5253 2 35 5414 15 4420 42 3 38 596 1122 2160 5812 2727 1105 5 4592 68 801 203 9 51 2 323 103 1 
+0,129 3922 4021 273 24 6208 23 8461 1734 263 0 92 30 1301 1697 10 812 387 495 2 475 0 2038 1102 6 872 1543 3497 4979 9507 3531 990 1969 1904 6 2890 1578 3906 6895 2355 2872 123 1 
+0,1602 1255 8 3 5486 2764 6 26 2554 8 3 220 1023 1933 846 254 4925 2753 3819 4094 1337 21 4462 4 1523 2080 2 1297 0 5556 1857 4 0 1193 573 13 3603 9 0 4401 4 361 1 
+0,11 2689 10 3 5513 731 6 3 1464 71 2135 48 3386 55 454 312 8 0 398 106 13 562 5094 148 46 32 795 1137 165 7 965 5531 4 150 9863 2277 7147 18 1 
+0,129 6857 27 4065 24 428 10005 2967 1563 6 189 268 162 266 9154 770 4 4627 530 389 39 10905 10879 2758 21 1661 43 10 13 295 2 2794 50 0 2714 4 10270 1144 39 0 2758 1 
+0,1204 2712 31 57 4 0 199 1479 8958 417 2 2792 0 1760 102 19 770 60 138 1345 104 103 380 497 21 8978 166 115 5488 19 5424 50 531 1452 2 3773 1452 174 6 5087 50 739 1452 5459 14 0 4965 11 3100 10741 1973 31 139 1452 174 12 11071 51 2 3 2060 174 1 
+0,129 6322 7394 412 24 407 1738 16 43 1764 9 785 4610 70 49 61 77 103 39 1672 0 43 390 2 8176 0 2223 2 1127 1824 1688 9 1770 308 27 79 4 724 14 2984 254 619 2049 2 1082 9573 2685 1688 10 55 34 11162 53 8841 2122 1 
+0,129 6341 6343 2322 2389 1601 27 38 579 8520 872 9 94 780 8139 714 5213 9769 0 904 4 4627 1736 2122 60 4 0 2071 877 6 1505 2718 2091 3 596 9176 532 19 2718 450 2 8317 123 2889 1 
+0,11 11122 1421 11303 4 0 1721 68 672 3 3358 10 140 3736 4 618 173 1421 1126 431 724 218 757 1456 698 23 3 8587 1659 4 730 1680 1466 4 0 10537 21 2549 2223 1 
+0,129 6858 7706 571 24 842 148 46 30 1005 4950 5 1236 3 2785 83 3 1695 5423 4917 0 646 4 667 4 3646 11107 15 566 15 3 2026 42 11223 4888 4791 2037 27 922 4 21 2549 2785 5 118 1162 94 2145 21 9494 2166 12 141 1812 808 2044 42 19 7 184 76 788 155 0 492 85 553 94 23 4570 0 5423 4917 1 
+0,129 4123 7513 6955 2858 24 7047 476 30 2611 0 1310 4 4855 71 10947 5089 5 3 1223 1029 249 557 335 406 1525 8463 9841 10 19 25 49 10161 61 917 41 45 1 
+0,117 0 43 814 4 16 1757 5881 724 32 4407 1828 12 21 8054 2785 405 4916 3297 10225 11 3328 30 147 5 40 65 7 660 4 0 7359 4 0 212 5782 4 7554 1 
+0,129 7393 6399 6850 2997 24 1177 9672 4106 152 3247 189 2598 7 1791 8 4803 1824 1077 308 28 975 636 524 149 329 7 2018 10557 6 3092 10314 4 10080 0 623 9 1770 8336 1 
+0,129 2944 2980 273 24 1905 162 266 6229 7205 31 391 769 2461 12 21 9015 1256 10 31 1512 884 60 7203 3 11161 5 72 31 3 174 492 1 
+0,129 6700 6220 273 24 11 11212 11296 3441 4862 8630 4 8477 3132 6 10250 697 2626 547 3486 9 94 1525 363 8392 71 363 5319 10779 757 13 3180 210 15 275 1707 896 9 395 46 1099 2 3597 505 10002 6 981 1 
+0,7022 123 3 1294 8316 1150 240 68 207 9880 20 348 202 5 0 1287 4 232 10759 449 2602 1 
+0,6838 1460 1265 0 604 8661 2 4828 933 359 629 6 2014 766 2161 363 1 
+0,2175 1937 1575 1322 1531 1346 2110 697 2821 91 2 474 8 1197 7 4323 6889 2169 9343 42 799 10148 1 
+0,7544 3177 2522 84 1388 6 716 1558 2 553 23 2554 4 726 140 34 3472 2 1435 507 1 
+0,604 529 766 5 0 4808 143 892 448 5 833 42 11064 30 517 2585 12 8004 588 61 529 1784 1 
+0,5882 4 3 898 2170 1549 0 8243 150 107 19 55 1809 667 4 987 20 3099 520 811 114 6 156 206 190 975 1 
+0,757 7 0 308 303 7 5551 8 130 7 2206 198 7661 3958 9 1236 3222 1 
+0,941 1004 2 9635 16 968 910 264 12 3 565 11198 9 2194 5152 1 
+0,2282 1056 91 7 1886 6184 2441 120 19 13 50 130 10 10350 249 232 94 1324 44 0 1239 1 
+0,1937 830 9844 1449 4317 805 8360 463 383 6640 191 5057 1173 7 207 1 
+0,11 3185 1663 2 32 1277 16 182 624 1700 9 3537 4 0 2194 2561 697 1 
+0,171 7 199 470 342 2722 2212 5 2686 42 3435 143 911 2 9940 275 1 
+0,3878 123 303 1311 141 191 2193 6192 345 154 8391 1857 39 1982 306 23 463 6 1173 1 
+0,463 6 1173 10846 485 12 624 2010 50 191 42 967 148 0 1784 1150 143 13 184 9517 966 1 
+0,941 9217 51 8 2144 4 9487 663 1230 39 1709 383 218 2282 1056 91 25 3422 1 
+0,435 9396 1661 414 7231 1545 0 10465 120 29 0 3185 711 166 1 
+0,5963 827 7 2515 5077 264 1736 3 858 12 3 603 1234 8632 2392 969 405 500 1 
+0,1597 1178 875 2 231 381 1768 1095 14 949 958 1 
+0,178 2452 500 5 2228 439 4 10089 6664 8 10 7 76 3 447 923 1 
+0,1894 1970 266 59 2233 788 15 0 8970 71 1259 4 1742 142 511 1665 69 348 245 3179 1 
+0,11 1387 3370 1725 9 3225 943 2045 10542 0 183 257 9 5805 2899 1 
+0,7365 1150 1320 85 5552 10664 1913 8009 2 265 2635 3508 6 4443 7848 3518 1 
+0,7565 4902 7723 6353 123 3 10122 9851 25 3287 735 392 142 53 21 604 1996 1222 3381 1 
+0,27 47 39 0 4132 2248 0 114 239 5270 959 9 178 609 4391 19 32 2748 0 1726 5394 931 1 
+0,11 820 47 266 1722 51 2 2254 389 5 1504 39 4418 2 1233 0 4132 326 2786 1 
+0,11 2670 9189 3 4885 5 0 440 519 910 114 10 55 34 404 2 167 854 2 5 4532 6 385 1 
+0,27 5569 5 1576 7 2079 5093 55 1413 644 1351 27 38 5909 2786 10595 290 0 2371 1 
+0,5875 55 8310 2995 405 2079 5093 7 503 2560 1576 123 3 3335 13 9492 1 
+0,2961 1077 190 1470 1299 105 2 309 463 7 5467 114 5 2256 344 392 2576 1 
+0,1058 99 4 0 1997 62 7 973 3673 550 3 624 4962 5 8213 1 
+0,2988 5325 16 43 216 567 338 12 21 1182 175 1299 1 
+0,27 719 7790 9442 9407 5 21 1111 1077 190 137 1462 3638 1123 405 0 7702 307 163 1 
+0,1406 62 10955 1703 0 63 9307 5 48 528 3055 4424 30 813 2 10404 35 8007 1 
+0,7154 2494 10938 2146 9 877 1409 1363 15 1221 1473 30 3272 197 524 1107 1 
+0,845 2371 190 371 54 30 436 973 3673 1665 69 3567 9518 860 3794 2255 1204 105 2 247 2670 1 
+0,117 3 2230 5327 83 10533 5905 5945 0 973 2672 9202 520 0 1406 10191 1 
+0,6485 7570 1322 1531 16 7933 765 84 29 154 981 2 4641 1090 11396 1 
+0,1596 123 2001 1682 30 934 8 3315 0 11328 11035 2560 2 1409 1363 897 84 55 1478 2 1406 1123 1 
+0,11 5337 4 4203 6 4204 29 34 1081 12 2896 3857 7 2707 15 0 5042 7 1078 1 
+0,2985 1104 943 10824 1664 8503 234 9 2098 2 1127 2154 241 8878 1 
+0,7340 2798 765 3526 602 7 4131 670 4874 9 377 7270 2577 3358 4 623 3307 4 1465 1 
+0,2302 239 8839 38 921 193 2 475 983 264 23 3 38 8077 1 
+0,6617 342 123 19 29 337 198 16 7268 2371 567 440 519 6 434 175 1144 1 
+0,6026 4 2143 372 3284 969 1263 38 2366 730 2044 114 1 
+0,111 9485 114 13 1005 2554 14 426 10863 1680 190 2126 1 
+0,7883 2 167 53 0 143 9 943 797 6734 438 6 10876 1478 10 121 387 200 49 15 3 9885 1 
+0,2904 4545 1150 502 404 5 1123 6 9687 2 2036 7560 4685 1439 10183 1466 9 1094 1721 1 
+0,117 3 218 40 65 0 41 45 945 4 6365 18 5021 3216 14 0 1052 5807 212 4101 64 76 3590 614 6 138 1429 64 795 4815 53 0 3417 7 5021 2023 1 
+0,6139 635 40 65 8364 0 3705 4 2709 10025 9 178 206 1125 4 1310 525 20 8084 105 2 9881 2 500 206 574 437 10459 14 16 2259 1323 1 
+0,3959 7950 3 1693 4902 4 0 3921 6178 123 5 21 2783 1992 10 0 1028 4 5144 5156 1114 69 5288 2 41 45 114 700 588 42 2 99 9654 4 3416 235 1 
+0,844 62 1098 14 3 6049 379 6997 508 40 65 4599 3 5548 4 1517 9 933 688 191 1534 23 159 67 96 2 5410 114 8868 6 3571 8892 8 3090 8756 1 
+0,11 3728 44 1913 143 13 448 6 20 119 47 682 4 99 1473 9565 5 0 41 45 30 302 2 5272 8 1406 193 583 4657 1697 10 9564 68 34 2015 1 
+0,464 160 7 57 923 0 3832 4195 1628 79 5525 15 231 3215 9 0 59 604 1470 19 7 10 121 793 1117 23 185 1775 1 
+0,4346 2838 1065 1141 530 5055 3240 1160 73 104 6 964 5705 38 300 188 0 158 5055 2169 1444 2 5633 170 104 1 
+0,1199 511 18 10 310 0 38 755 1935 66 9 178 609 29 4522 0 307 163 7 142 1741 29 184 548 3 234 2 4844 245 3336 1 
+0,2990 14 1203 2972 6773 6 6725 40 65 1804 4999 10242 15 4421 5 3289 1048 700 71 434 399 4651 1 
+0,6883 4045 579 1536 3721 6 440 519 114 275 30 1485 1512 2 1654 563 9 0 130 1418 579 1536 193 1320 7572 2329 278 15 138 440 519 511 2152 2527 10 0 193 29 235 1 
+0,949 958 8281 967 12 306 10 16 194 6 384 158 383 85 8198 1774 874 10288 213 4 799 5 16 910 6 2749 1441 1 
+0,607 1043 1059 695 239 1713 58 25 3093 3 352 113 1319 5 575 1044 2481 4422 434 729 6197 0 275 18 95 1 
+0,5794 7051 3939 58 7 1575 58 7 1346 2110 5138 2684 30 8 35 234 2 300 187 2 3 308 5137 2149 20 6564 7517 408 86 37 451 36 5759 37 451 36 37 451 36 37 9486 10864 305 758 3269 487 758 9476 1336 9479 996 9469 11023 5663 10030 5664 11382 73 9394 73 1223 226 36 37 3 9449 305 758 3269 487 4608 5746 5745 5521 305 758 3269 487 4608 5743 5744 3237 305 373 9468 91 805 9047 10332 4562 4518 2076 9450 8387 10092 1277 4518 792 1362 7261 7262 36 6741 574 3281 3494 525 37 3 36 37 1677 36 7349 20 6686 182 1412 2260 1074 9871 7254 11126 3494 525 68 145 7812 6 500 3631 6885 49 50 3281 3494 525 37 451 36 1 
+0,27 561 5223 344 190 1536 2431 20 0 41 45 637 1888 1174 1596 991 40 162 1092 69 3539 190 2 190 3692 2641 42 1596 134 763 0 38 561 3202 12 3 739 47 266 893 284 801 41 45 344 190 300 83 5536 1694 440 519 1 
+0,2896 3857 29 3583 15 1078 4 4204 58 8 2989 1162 0 62 172 95 11 62 18 19 2051 52 29 767 5 26 2174 15 1078 4 2733 62 4203 58 1 
+0,116 58 13 714 0 695 9 16 281 1215 586 183 394 377 8 95 12 105 2 694 0 10285 4 16 202 119 65 1 
+0,74 464 160 7 57 923 0 3832 4195 1628 79 5525 15 231 3215 9 0 59 604 1470 19 7 10 121 793 1117 23 185 1775 1 
+0,5879 62 13 1662 3272 2 0 8499 1280 5181 20 2486 10518 729 1 
+0,845 148 7910 609 959 3299 646 4638 651 69 801 15 1690 1 
+0,6332 8019 4 3715 522 284 871 333 2 2047 585 142 1 
+0,2363 1123 30 286 4 931 3701 164 1111 565 4 0 137 7090 1 
+0,4020 4390 10780 513 173 4711 3699 9 1996 3726 1 
+0,130 3498 12 5972 1059 2863 6 7097 91 8 57 1826 2048 5192 1 
+0,6917 4857 147 23 175 5157 42 9834 19 7 213 1244 7 256 3172 1 
+0,1576 3224 9999 2404 10819 12 1996 470 10 1263 3 47 4 1576 175 1 
+0,4319 3253 344 301 854 163 9 782 863 529 1 
+0,7739 1722 1504 9 5535 2635 2786 4 748 2851 2248 75 47 1 
+0,4338 584 105 3113 581 23 4793 2 4692 1223 894 1095 1 
+0,7153 9939 1090 911 470 5568 2 2249 1741 1 
+0,7371 1346 2110 1798 765 10309 49 558 61 545 6055 193 1 
+0,1397 495 1439 8118 2 10663 8173 2 9357 1536 1 
+0,7638 1665 69 8939 3113 2624 20 307 5144 46 491 118 9482 19 1 
+0,3079 1090 2736 421 579 9272 10222 2 520 3 38 2011 4584 2 5044 1 
+0,4059 5517 1292 4877 1330 7 11275 42 527 1114 69 4823 155 121 491 320 3 624 565 1 
+0,111 5677 481 1280 1122 173 8348 6 2708 14 3 363 370 1 
+0,6927 3341 229 9867 8287 8577 4 2545 10441 10421 306 6 2489 30 1446 2 396 3 393 4 357 4600 5 9591 3664 661 2160 1 
+0,129 1917 6562 7242 98 1047 24 27 150 2084 2115 3059 2386 3065 6 1166 91 2 235 925 249 0 367 523 68 7971 20 189 715 10506 1166 2 553 3611 8 16 434 729 23 3432 395 2 84 3059 2386 3065 84 436 896 164 1166 25 3308 1021 7692 1891 4077 4141 2355 1 
+0,116 58 1598 2568 274 12 16 377 695 95 278 15 0 367 208 705 294 4408 3 1778 2149 712 620 12 16 882 4653 9901 385 1 
+0,129 4127 4376 607 1915 745 24 1059 4477 1713 58 1 
+0,7682 562 5 3 183 5066 2784 1114 69 3 9846 2052 1068 703 3 3199 494 0 257 6 592 8 211 14 0 7663 4 6901 1800 3577 39 3 48 47 8517 10883 3 2118 1122 3 3358 53 0 618 1640 10492 337 6 10493 9753 11312 2229 21 2536 5255 10 3599 3331 1 
+0,129 4125 7508 6213 24 4189 3 10881 4673 2 524 949 958 408 830 58 1893 147 3 217 742 113 1269 5 487 1468 15 347 191 4 2683 805 6 9994 546 3257 2573 5 3493 896 1 
+0,129 2357 4098 607 1043 24 116 58 1598 7 1461 1215 377 1450 3 719 5429 2457 86 12 0 247 4 3 751 620 10 0 367 208 705 294 7 1693 882 775 118 132 0 62 1247 16 105 2152 170 153 12 16 202 394 1 
+0,607 1915 745 24 11 1387 3370 3321 9 1236 943 2045 3225 114 25 1662 918 333 40 72 1224 0 5067 8 602 746 58 1 
+0,129 4125 6959 74 205 24 2271 0 5303 4 0 1083 0 10726 4 0 1123 68 34 8757 429 350 4 1219 5387 2615 134 6 49 8816 8 0 685 359 274 4 0 204 1 
+0,11 1805 10736 5 0 5391 4 3 3690 2666 8169 6 0 9626 1454 8187 0 4503 14 35 2155 6 1991 8680 1612 245 615 160 7 1416 0 4702 4 3 10344 11265 71 215 4967 211 455 8328 32 278 49 458 166 8233 9118 6 114 2470 1 
+0,129 6500 4065 2393 24 27 835 173 676 677 5188 2066 70 5 512 228 2 5535 3 2786 4 0 2851 156 2248 84 8712 49 61 3 104 1429 75 919 1917 1610 4206 820 4 4039 7062 13 362 2 279 530 389 2 77 103 291 3132 39 10193 2066 2 57 2500 4 9578 4573 71 8159 2 646 1434 2 3 10320 220 1 
+0,129 6952 6395 24 577 0 258 527 2 130 97 1187 7 178 609 307 163 1292 1429 29 1144 1116 1290 4528 11059 6 10848 15 3 811 4 5120 5 185 361 1 
+0,130 172 70 19 85 488 3 363 439 3655 2534 4 16 178 609 307 163 5 496 916 5 838 15 19 10928 2 783 2659 143 911 2555 9035 23 0 520 811 500 163 6 114 2673 1 
+0,116 58 1598 29 337 0 2705 800 9 16 377 695 86 2186 0 2744 9 0 367 208 705 294 7 9447 1215 202 143 1085 116 105 2 516 21 1333 695 2 425 352 289 104 400 10708 39 1692 0 2705 1 
+0,129 6757 7494 4276 6862 24 27 62 8931 60 4 264 20 1370 6 0 372 137 240 44 16 943 6 220 137 8678 114 25 1516 3 1528 12 0 2104 3511 240 1 
+0,129 7942 6800 638 24 11 781 4 130 97 1187 7 313 927 4408 81 10 0 41 45 1 
+0,129 4127 4376 607 1915 745 24 6670 2 10855 3 170 1361 2711 3033 1065 4 324 2610 28 3 38 1251 1787 220 986 1087 2 1704 1721 10231 1466 568 0 794 9 730 9305 1 
+0,1059 208 705 294 116 58 1598 29 2214 49 61 1866 104 4 16 202 2 1965 58 4393 5 3 1528 10 10489 3 894 1028 9369 44 16 377 14 0 2548 4 9017 3 9969 524 1 
+0,129 6957 7834 74 205 24 464 121 491 900 8 4474 40 162 7 59 252 121 3237 356 34 8417 120 121 148 6 277 5 183 844 29 34 1856 6 9744 1 
+0,129 6760 7253 24 353 3 624 894 1138 3 829 62 10 582 16 5168 2501 0 3663 372 193 404 20 3100 434 1817 13 1993 1095 2 8540 450 23 3 930 1439 229 4 3663 372 371 8545 6 11237 4180 3853 745 206 5781 7029 7736 97 1 
+0,11 5171 72 9 8753 23 4959 5 111 332 635 13 95 14 268 451 486 9632 20 0 1077 190 860 4713 1716 4831 2 8308 7793 13 46 491 4855 908 860 210 256 2656 20 48 1077 190 2691 6138 3079 6 1207 2363 1404 1 
+0,129 2357 1047 6950 80 93 24 6386 2154 3 62 7 1199 6 2305 1174 3327 262 6177 5612 9771 1 
+0,2949 2 40 65 7 5489 23 208 705 2569 290 0 932 3024 2985 7670 1054 6587 7 3030 3024 5861 4200 3858 6073 4202 4243 9 7038 7503 2904 7223 2 7429 6427 4209 6445 7778 7280 2949 7779 116 7578 4200 7283 6287 3802 8 3035 7678 
+0,6294 5 2035 890 4 3 932 729 7901 10627 1122 3 5579 4 4729 50 0 94 6 193 291 118 50 245 932 729 8 0 899 1 
+0,6469 8315 434 1232 1767 1413 121 2 1082 185 621 1292 10649 8389 4 932 5158 1 
+0,7767 4 426 11259 10158 10 9595 185 220 568 5175 5919 9594 10859 6 11144 8672 12 1965 7 1778 3717 11130 1 
+0,116 6 1965 172 195 10 46 32 5284 48 4656 1267 197 0 275 1 
+0,2949 2 40 65 7 5489 23 208 705 2569 290 0 932 6919 7441 23 1396 1901 7674 575 2338 250 4202 5989 7414 116 2266 7107 353 3024 7592 851 7660 3 1642 1037 1396 6742 3731 6183 3052 7569 2952 14 7943 7521 3011 
+0,1713 6031 352 4 8699 1713 58 195 5237 3 352 113 1296 3309 2644 1319 5 575 1044 206 2502 3 557 367 392 4 2481 1069 6 2569 11 352 1319 31 5237 23 3 238 2502 4807 54 43 8642 1713 12 0 2147 1310 187 2 2502 134 577 2030 1 
+0,4005 5859 1396 1901 7221 7462 961 78 29 2501 0 182 3256 1834 6 2767 2 235 12 963 1094 4 455 1834 2767 6 8825 32 3 1848 3551 9 932 8823 11357 6 996 489 1539 11050 12 1472 0 729 51 2 1699 46 1094 436 3532 799 1 
+0,1396 1901 6996 7067 7447 7146 2 3835 1577 2 3 624 1026 8601 20 4297 3487 6 2907 5 1161 367 725 191 1756 2814 2 673 5632 153 12 5733 4 99 367 2709 9052 64 5220 1059 191 30 302 2 673 115 115 113 4 2228 725 191 5 1 
+0,2888 751 6740 7858 116 377 691 0 106 4 0 377 698 49 116 2497 116 2716 7 882 29 1976 5 21 2783 660 4 751 712 1352 0 620 31 1662 304 132 0 377 4970 0 751 620 140 32 4550 56 1945 2624 44 1296 377 5734 5242 1 
+0,116 377 3858 4193 116 7 377 1674 13 1489 520 116 6 16 5527 1099 2 520 0 695 9 0 400 4 116 2716 7 3896 27 1695 202 14 374 1566 3 486 6327 490 72 8 95 430 739 250 116 8257 793 32 3477 3 8256 2329 23 5023 4924 91 155 121 1 
+0,2864 6070 2807 104 377 6180 1070 6 2678 51 62 2864 97 25 3166 105 9 3 2807 104 586 183 394 416 2 545 143 980 19 18 5 3 3571 2564 8 70 2864 84 1247 16 3489 10318 5 1582 4 40 47 89 76 4751 0 393 1 
+0,116 377 7432 6364 961 116 25 192 35 377 2705 1700 14 195 9 485 2 1519 9 0 116 586 183 394 1373 11400 2 5164 5 116 7 586 183 394 32 631 139 10210 195 2 1519 9 0 695 14 5023 4924 91 187 2 3 679 1115 81 8 1 
+0,1870 0 116 377 462 0 421 1012 389 0 994 6 156 92 25 90 1856 116 5 4447 4 120 55 34 0 199 6727 4240 4189 4 202 5 551 464 121 30 5256 40 3201 6 32 76 90 1272 8 1381 9 0 421 47 121 1 
+0,3835 1396 1901 5971 11 356 234 2 1220 256 4512 20 0 208 2046 13 2 1220 405 138 10365 2767 10 64 1014 557 2 1733 173 10385 1176 155 185 1341 13 76 4512 20 405 138 4 0 2767 2454 19 140 34 10137 3164 5 0 1 
+0,116 2923 6239 2 34 5855 7735 723 10 0 575 2338 7040 1056 13 320 21 971 50 6001 4081 7 156 3824 3824 3487 963 52 7 90 10299 232 526 132 0 57 2081 11113 2 2736 60 40 72 13 10 4081 25 760 116 2 8908 35 2023 1 
+0,11 1596 7996 10 0 107 2636 392 10 2627 134 9587 1798 6 921 8351 5 21 1716 13 3732 2 2536 11037 1 
+0,407 7 176 8738 1454 32 3412 51 12 0 1984 240 195 5 1854 4 0 258 857 2 8785 367 1984 300 1 
+0,260 8303 32 1088 11330 10 1675 1409 1363 3698 42 184 1415 135 567 3699 2 1492 164 5 3 645 2 2139 11222 854 2 5334 301 1557 0 4375 1 
+0,11 3551 4 40 971 13 2 8069 3 183 4844 5 3 3417 9025 1024 0 4432 9284 20 21 3408 4732 163 6 564 277 138 10095 1659 4 0 2635 8259 5 715 2 478 1156 19 1 
+0,542 971 3261 468 3797 4242 3912 12 178 1949 1161 84 1415 3098 2 5240 344 371 132 884 203 585 392 854 20 43 9019 455 3442 30 51 2 1699 187 2 3 8227 142 1280 1 
+0,117 40 971 600 29 711 14 3 2099 2 4561 0 854 982 5289 4 3 62 7 10585 71 3334 542 385 13 2607 2 672 154 9 426 54 30 9719 1152 0 142 4 3 392 287 46 30 5 1262 9096 71 796 8639 1 
+0,542 194 6 131 368 4 0 567 10136 922 728 1004 14 230 2 5103 0 142 2714 5155 5 0 1503 8141 6 564 1004 14 138 2147 3642 3233 5 1458 4 0 6692 1 
+0,7669 503 23 99 0 11297 4970 11 57 1708 10 29 1416 34 0 356 1330 9 385 142 13 3 8573 142 2143 1 
+0,755 1935 66 9 609 10502 3 1312 1079 5 130 7 142 5199 1369 7 933 121 284 9481 0 9956 6 453 900 185 1334 1 
+0,27 820 47 266 343 25 5188 2066 2 9537 350 4 1227 6 41 45 107 1429 12 0 2851 326 931 1917 1610 4206 4 4039 7063 1 
+0,5799 21 4986 1700 3 809 79 14 3036 6137 4308 832 2016 3 48 1482 8080 10310 2 2690 10 3 1150 10626 229 55 277 4705 2466 235 204 739 
+0,602 25 3144 1531 2366 84 4714 173 898 905 72 2104 4926 3253 6 5011 12 2902 2876 1196 1368 6 943 3044 1196 170 14 3 8335 370 4 3769 204 169 
+0,4233 3045 4823 114 10962 9 2058 6 372 10 4714 905 72 3537 1554 5 57 4630 6 2606 2118 8555 2902 2876 1196 1368 2366 6 943 3044 1196 170 4233 3045 13 762 210 9 3 3679 725 370 4 73 3769 204 169 
+0,11 4973 2369 3042 210 25 49 61 57 104 2196 762 9 871 5 0 56 0 43 367 481 365 175 2 488 981 3 104 2740 8425 204 169 
+0,5833 0 1828 4 0 1465 2643 5 313 6770 1565 1865 11 1465 2643 13 5 1723 57 4 426 99 597 1515 4829 4 21 4975 671 9 0 466 143 165 13 3 671 10 771 562 1848 771 19 722 566 6 1004 5470 597 7175 4 455 4854 8 35 621 13 614 2 200 0 1465 2643 3 1828 35 8553 164 8787 6 4686 3392 4 9258 13 120 2095 0 3255 5788 5 138 1843 6 3691 1765 1543 1768 3 1541 10 1004 15 447 15 1379 7868 6 3 173 2074 4485 457 1070 581 10 25 4678 1206 5 10179 2580 6 11437 9 528 6 138 4 0 1517 9 0 1465 2643 3507 68 34 11228 2251 204 169 
+0,1 1865 262 135 943 8174 522 5 0 1034 370 1428 13 3168 4 1233 322 2754 1461 4686 2879 15 602 569 7 943 3044 1196 170 2251 1565 470 4134 1865 462 372 3544 590 2 494 35 10296 2 0 119 661 19 569 7 1160 3 1812 711 2251 204 186 
+0,2889 6994 7470 6 3799 7785 14 3040 254 2804 6462 2358 4188 853 1659 114 9 9209 4699 173 1146 8179 9873 961 138 818 161 2 323 161 4699 5 309 30 2358 206 204 170 
+0,5876 14 40 47 569 7 4227 254 10379 51 35 2753 385 163 2735 12 0 38 4227 186 226 6 8171 35 700 8000 164 2123 1283 204 170 
+0,6599 3010 712 172 188 0 919 7110 452 5 7135 10 19 25 8190 6527 0 712 569 7 5255 569 7 3186 9 2279 844 4 250 6793 1570 
+2,24 219 271 209 863 4 903 9 0 59 460 364 445 2183 33 6 564 280 60 52 89 90 964 2 0 126 79 9 0 1447 7 570 445 625 131 271 726 687 848 6 1622 1338 1706 1603 6 2221 2346 2344 64 877 61 0 1548 5 0 460 513 1502 1 
+2,24 7909 4154 7291 1260 1844 6 989 5 133 675 1744 0 2290 7426 355 0 575 1896 2998 351 139 8 136 110 575 1896 31 1758 60 4 3 911 4 0 1618 1561 862 231 12 0 386 6 461 7 224 44 1052 1167 991 5 0 106 1 
+2,125 212 2844 1037 797 734 2 108 3 384 699 59 112 144 9 0 88 128 32 9323 0 2791 51 1228 10 0 3948 1638 276 30 1711 278 155 2756 32 76 1 
+2,24 11 1358 952 802 601 57 137 4 0 1035 940 231 2179 133 675 5 0 43 340 6 972 0 835 1403 289 73 136 110 291 168 675 20 4353 4024 6 3069 1615 1 
+2,24 2345 6626 181 321 2718 9 1052 1167 6 6789 3874 89 48 2718 9 6160 15 0 6518 6 6199 901 2 3 226 226 2769 136 110 1 
+2,24 1399 829 29 504 650 4 16 43 420 403 5 3770 103 15 0 262 73 79 5 0 390 11 7791 1714 0 176 1538 5 0 2682 1036 378 2130 296 136 20 3567 1571 4 1345 43 422 2790 6 73 5722 1499 1 
+2,24 2928 7 6454 7293 124 0 4046 4235 330 9 189 43 3073 1959 403 972 6837 6841 4 0 2877 1393 289 139 170 115 115 170 136 39 409 330 336 5835 7106 5599 132 0 2183 213 4 3 10555 1471 1 
+2,74 17 219 271 26 43 112 144 10607 1734 5 772 164 0 119 2124 4 26 9920 59 2003 5 0 2591 4 0 460 666 445 33 42 0 9489 4 0 1833 274 31 2006 1 
+2,74 17 27 8856 11421 4 59 10315 31 4483 33 102 1914 92 2616 336 2830 4150 656 2 200 0 3738 9 26 3191 12 21 411 2114 1 
+2,1918 1208 17 2414 3081 656 2 200 0 182 4 3 2155 2072 453 5 0 41 45 839 534 194 346 8 136 780 2 3618 9 3 77 243 1788 3779 1 
+2,74 17 211 9 3 649 10 9016 189 613 6868 6157 387 32 90 0 1614 539 4 0 151 177 1 
+2,80 93 17 149 4190 10964 26 238 79 20 5262 3 48 145 1143 5 0 1240 340 2 988 5 0 222 274 675 84 1714 0 111 332 1648 3 115 186 108 44 0 1261 4288 6992 5 0 126 467 136 1 
+2,80 93 17 6129 539 4682 48 2596 6 6795 6476 7 48 145 546 5 0 1813 340 989 5 0 457 675 15 0 2751 826 1585 4778 0 1064 1379 1587 374 289 5 0 212 467 136 1 
+2,74 17 11 9431 4 205 5056 1210 998 1219 6 141 8197 8094 284 4631 19 685 2100 18 33 39 0 127 7 176 1219 64 2218 23 0 79 243 3 4614 4 1442 2760 1 
+2,74 17 11 88 128 355 1039 5 3 92 356 72 2 3554 9 0 131 4 0 146 7 59 3288 282 33 15 0 48 8711 2163 35 4872 10566 5 1458 4 3 10392 1083 14 7549 1 
+2,5886 0 589 2232 10 1758 135 8650 82 1331 1066 756 304 3 57 2752 231 44 834 1611 2587 53 0 131 346 4 0 1389 2313 27 534 1 
+2,149 4190 10543 5 26 626 2 7518 1903 20 10749 3 48 145 1143 10 3427 0 1648 6 3598 6378 6925 23 3 2506 1 
+2,2308 1041 5 0 6432 2099 4910 6 446 2 182 3342 4221 11127 263 0 2451 12 5591 4852 2086 7091 3141 6126 5376 1350 13 8283 8 0 10963 10836 8990 9698 1 
+2,11 1648 2827 7469 54 713 95 110 7 137 12 0 4894 6 2747 122 14 0 79 1747 109 1444 26 1216 4 3 133 137 5451 1115 9 26 9608 5 0 833 822 8333 14 828 6 611 5342 0 3503 573 124 69 34 8981 2 521 631 86 14 835 1 
+2,329 3080 1260 6 1414 133 1462 5 441 889 295 0 461 1175 44 0 3727 1052 1167 3944 66 226 109 1 
+2,577 4285 4386 4534 9 3 2680 2581 8842 4 26 258 4795 1471 0 471 331 75 110 1071 10216 2 592 8 568 180 15 6209 6892 9516 4612 2 10469 0 4615 11 331 2078 2601 46 85 200 21 2439 75 110 8 4386 7 1471 42 865 35 1792 188 0 43 340 4 0 137 1 
+2,3010 9593 14 0 131 6140 683 1635 6082 1826 2142 21 8194 3464 3074 4143 6 7401 4099 64 0 2561 4937 9 0 513 316 8510 14 0 2859 2308 7940 589 75 65 4143 7 2138 11439 4587 4 87 3074 3074 87 23 0 1083 4 1864 2086 4099 215 0 738 1100 3745 87 3077 13 52 87 11 5224 1 
+2,1911 1410 7299 173 8 0 8295 10074 2350 2362 1647 1634 13 3 791 11388 9683 15 710 3 3362 868 15 531 104 68 479 165 13 1081 12 10617 8945 10 55 546 15 10739 5372 6 1235 8373 10 284 5281 8979 1 
+2,74 7658 1318 2940 1608 6 1607 1639 64 2218 109 23 0 685 59 79 9 1279 872 1153 42 35 2559 31 215 5 0 1742 4 0 465 59 1359 1 
+2,7456 6524 0 2811 1189 6961 54 31 1758 60 4 0 7405 386 2 2342 635 75 2989 323 6 89 5351 2758 5 1928 31 374 9 818 9 3779 3751 5 3727 2374 7 1867 742 2050 108 44 0 575 1044 3774 75 110 1 
+2,95 110 506 15 337 2 21 266 72 1002 1221 9 11305 335 6109 186 139 6 289 15 600 491 362 2 1024 165 31 3 11109 2 1296 1997 1296 4753 6 278 1296 7810 276 15 0 77 1002 87 10041 87 5340 51 3 730 110 4 1317 5222 1352 471 331 1067 331 31 8 2286 186 7 87 2733 87 1540 2286 2255 19 1 
+2,273 4266 1581 2297 701 69 278 32 72 2 167 26 1560 174 10754 9622 10777 8 0 5373 7391 1954 482 75 110 132 10613 3 66 66 2769 9 0 3018 83 1361 938 88 6438 339 909 39 9020 0 137 15 3 589 5432 2297 3164 3 363 492 23 1265 0 3503 643 6 2055 26 79 7 5522 1 
+2,7193 98 1047 7603 7025 5906 950 802 503 5 0 3667 2 108 0 1344 161 6245 6618 14 7079 1194 109 9030 0 1331 159 9 139 9264 1 
+2,1911 1410 464 52 64 1969 26 1970 1376 1600 85 32 89 0 736 82 10525 8 26 9696 6 10400 0 1420 4 0 1486 6 1238 1885 1959 1 
+2,1399 829 4929 306 4 16 43 2682 262 73 3559 248 5654 12 8064 1 
+2,1911 1410 462 530 2070 7680 5847 901 9196 3362 5 109 7 194 346 133 8266 6 740 10113 165 1158 69 631 52 31 1711 2128 10 52 985 10 19 1158 69 15 447 3 106 15 52 89 924 10 52 89 181 3 10355 8304 1 
+2,6437 6 1013 9 3985 6 4067 879 903 9 3694 7 1371 1179 1624 1 
+2,11 1149 32 118 1101 35 131 10274 9 0 282 6 903 5 281 895 2224 61 109 1 
+2,11 4257 3972 1638 2529 4842 230 121 68 200 51 2134 14 0 1371 4597 1 
+2,11 4257 3972 1638 1513 164 16 5221 568 11415 50 0 11119 71 16 726 11 878 25 90 8 7689 282 76 3993 9311 1 
+2,3985 11083 6 4067 10675 42 141 646 9 2492 310 7281 7703 1720 0 3178 4 3993 14 0 282 1 
+2,27 384 158 231 10326 9199 9 0 4255 15 149 6811 1462 3 5712 3750 482 3360 136 2 777 0 3001 3 821 323 2116 108 1 
+2,6251 420 336 3062 1673 0 250 283 674 225 19 215 82 0 47 132 15 0 176 1514 79 5 0 24 2416 352 2130 1 
+2,639 64 547 3479 4961 188 1405 2411 7 43 5327 4 1245 1793 8 136 1 
+2,7784 6677 0 48 72 1350 7042 3010 4 0 47 2211 3 115 66 115 226 386 2 1570 47 266 9629 7006 433 0 5611 4551 4 238 5212 6 5661 4387 9172 6974 1 
+2,3837 204 740 1176 132 299 328 10383 51 0 1899 1887 6 611 2 703 16 8933 8 7317 7398 6194 0 1573 161 1350 5995 7651 9 2966 8339 9167 6 9815 89 90 1148 82 0 1331 7 355 51 11201 868 1 
+2,497 139 1452 268 5643 2681 63 3134 1054 6465 1004 49 166 0 3896 5750 4199 6944 7 8231 61 57 4 0 1354 1934 176 5229 1 
+2,11 41 45 288 7 6 146 7 3288 79 505 192 92 4506 5 0 59 5311 438 33 966 12 431 2119 7526 7159 6 5828 6612 1 
+2,219 271 13 1010 2 0 2258 364 445 625 79 10 29 2487 5 33 7 131 1472 3107 26 1797 9 3 719 441 59 112 665 1 
+1,17 962 735 60 382 6 5 710 1121 8 33 2 456 5 3 659 319 10 29 3289 5274 215 1159 101 261 251 23 489 71 777 180 3 38 2097 2 2579 9 0 119 48 103 1 
+1,17 227 1051 369 404 599 8403 5 764 3032 33 2 4755 14 214 289 161 5233 3718 0 107 2 5299 3 1431 649 2 804 49 196 2 100 1 
+1,17 1956 4 692 721 5 411 5026 611 3 1110 454 9 478 980 33 42 538 7 142 1120 18 52 701 69 863 155 46 3656 2 769 1 
+1,125 7645 3369 6 10690 209 422 1491 5 0 8885 1365 233 4 227 1192 5 11310 4 3 3346 1687 3749 321 1365 10669 1 
+1,24 2298 4 1627 810 3337 53 315 7 1556 298 233 33 0 43 881 1417 584 3252 5 0 643 248 743 1481 611 3 3558 83 1222 633 2558 515 350 1 
+1,17 1627 196 64 8043 8 33 2 315 7 298 15 0 43 881 584 160 2636 2 801 1779 1776 3 1686 616 197 0 576 107 6 907 5 0 1556 681 233 1 
+1,24 27 1423 2053 188 21 2331 1041 2667 5 605 7 1803 3474 33 515 14 214 818 94 360 3600 310 3 1308 475 188 3 2475 14 3 676 5 0 2184 233 4 4084 2604 632 134 18 1 
+1,125 325 7 881 1120 29 556 3 1515 816 2 413 689 40 65 9 349 8 16 1488 5221 39 1233 3 10900 292 20 1854 3 413 1051 1278 85 34 1209 2 246 1400 1 
+1,24 636 594 1998 149 329 771 76 318 2 816 312 5 0 4426 4 299 328 213 52 7 3210 26 581 9024 55 8891 23 2702 1095 52 18 136 1 
+1,24 636 2241 594 1998 149 2301 358 9 1230 2 3531 872 1507 6 10257 320 3448 149 329 7 388 6 107 175 14 21 10052 581 1302 292 136 1 
+1,125 1352 2131 452 0 56 594 282 3 1853 2025 1257 2295 149 329 1663 2 34 3352 21 1239 44 947 2424 152 335 0 257 3657 10 55 1435 0 10050 1 
+1,24 1360 793 1492 9714 2 801 0 390 7 8976 3367 155 19 1340 2 1220 10496 4 0 2514 10048 10 1135 290 441 3657 75 47 1946 6990 6072 1361 7887 18 136 1 
+1,24 7066 7 3125 996 5437 0 512 2543 354 8 95 595 19 13 291 8 1152 0 150 7 474 9 231 15 10504 20 893 1 
+1,744 17 4052 1249 5110 14 3 1276 225 269 1271 1277 2 1494 21 1752 420 1666 33 515 14 214 48 94 5 3 2581 3576 4 0 127 7 11134 3501 926 1086 1 
+1,638 17 1957 466 3460 1015 9426 14 3 1431 628 9 259 2025 33 0 5720 8082 4 313 7 379 1407 686 2214 3 592 10 3274 8075 23 936 9964 1 
+1,3895 197 56 196 6 432 1481 9041 86 15 0 56 10989 253 9 719 2729 1 
+1,691 33 962 29 1435 287 2 265 807 0 267 7 1151 84 13 416 2 157 5 1568 1 
+1,2271 3 10436 134 40 65 2693 0 8581 1699 9 3722 122 2 5665 1 
+1,7726 8 2186 51 21 269 1666 767 583 617 598 5 1170 6 0 4296 10931 4 253 1 
+1,27 319 13 243 234 5 272 2 1435 155 101 261 251 284 1019 5 489 1 
+1,428 8038 1062 149 414 686 123 1188 14 4119 0 409 628 5298 20 1942 6091 1 
+1,1627 196 4464 5 315 2 154 801 1687 2649 5 0 259 2698 298 233 1 
+1,687 848 3617 0 1246 72 5 0 3764 445 15 1544 7 176 2652 4694 53 0 2723 1455 1 
+1,1183 7 2830 7065 11402 23 0 177 1214 8615 44 26 1517 1 
+1,4280 5921 7 5345 2441 3 2084 155 52 68 247 0 345 4 21 5004 53 219 1916 7 3702 20 369 1 
+1,11 100 869 6 8847 44 7141 9758 44 0 56 3933 2397 7 1276 12 26 314 3221 1 
+1,1046 6 227 1192 5126 0 2513 4 3 2233 46 2568 583 723 4 3 1852 475 1 
+1,11 260 2614 54 31 9648 5 100 123 52 1450 9906 9083 6 1843 2 9044 26 8412 1 
+1,27 343 4662 4 10386 6 515 3 10637 13 9368 5 605 7 43 9082 5 1012 103 1 
+1,4032 7344 2447 6758 7486 8794 3 809 229 4 1100 20 884 189 43 183 5256 5 4284 9 133 103 1 
+1,533 7277 8 505 1816 30 975 19 0 182 10218 6 3393 237 5 272 7 551 3 594 424 319 195 10 29 1438 0 868 4 1086 292 6 55 8365 92 67 96 5986 4562 30 436 11281 3217 46 29 108 1 
+1,412 691 3 11007 430 1211 947 4379 11360 2069 26 8343 26 1317 9619 6 26 10375 53 3 3470 10652 4400 0 10303 11218 474 8599 262 2255 2471 15 19 3584 16 10594 198 2324 6663 5 1378 7 8381 179 976 1 
+1,1377 375 7460 3459 8496 5 681 375 109 2164 515 1569 94 6 10380 51 2492 50 142 15 0 127 3530 9 2623 2535 1 
+1,744 353 49 61 3 65 4 443 197 0 6968 1169 6 0 3202 56 6 269 444 5 253 1170 7 432 635 6 3 779 1093 135 4600 967 148 9966 1752 501 244 2929 2270 9988 1300 480 1383 334 432 89 245 447 3486 215 135 61 2 1326 1 
+1,7499 313 691 3 8522 1211 5 40 10753 2679 179 21 351 47 266 4920 8929 5 1421 6 8527 3 8327 8737 3735 53 0 9775 14 189 8977 676 7733 6 2208 9252 8151 0 1533 9329 788 9 189 9731 9955 89 9331 1749 9406 42 912 1512 8881 10 9511 2800 18 20 11280 3 547 1 
+1,7498 3864 6652 435 4162 444 43 506 2 3864 217 3 2027 380 46 3433 5 4953 3368 8584 10124 0 917 5 1546 6 1782 11410 585 3156 8131 1 
+1,4272 462 103 1610 7951 6784 892 189 3605 7899 8097 760 6 46 510 99 0 72 15 189 4577 15 21 8002 6 1280 5393 912 2732 4720 0 5314 262 912 31 76 3 9225 1950 31 1051 1 
+1,533 272 11 669 7 174 6 3153 581 2 2663 640 101 261 251 993 506 198 2 3 424 319 33 12 0 1762 294 1463 3 514 5515 335 0 902 29 553 180 5 558 3798 8 505 1816 192 82 4881 6 8276 10424 4 2469 365 2 2791 817 552 132 2024 1463 9 3 4890 4 382 2790 5 35 4864 1 
+1,1060 1045 831 353 776 3 43 711 14 0 2245 1434 215 291 20 299 328 312 1133 64 1450 12 0 4457 1833 4 5383 164 0 2250 6 9 138 916 44 333 1 
+1,6956 299 328 1505 29 76 646 6461 786 10273 2 10743 166 2810 7 1580 6 9156 1266 284 222 1985 23 2671 60 4701 302 2 673 0 3149 4 987 150 6 240 134 18 136 847 148 1503 10272 3396 6 10071 181 213 4 1580 0 182 4839 5123 1091 5 41 45 1 
+1,1060 1045 831 353 776 3 43 711 14 0 2245 1434 215 291 20 299 328 312 1133 64 1450 12 0 4457 1833 4 5383 164 0 2250 6 9 138 916 44 333 1 
+1,74 205 219 271 209 863 4 903 9 0 59 460 364 445 2183 33 6 564 280 60 52 89 90 964 2 0 126 79 9 0 1447 7 570 445 625 131 271 726 687 848 6 1622 1338 1706 1603 6 2221 2346 2344 64 877 61 0 1548 5 0 460 513 1502 1 
+1,533 272 11 669 7 174 6 3153 581 2 2663 640 101 261 251 993 506 198 2 3 424 319 33 12 0 1762 294 1463 3 514 5515 335 0 902 29 553 180 5 558 2990 23 263 0 92 360 1062 149 414 686 6 41 45 1 
+1,80 1892 27 1423 2053 188 21 2331 1041 2667 5 605 7 1803 3474 8 33 515 14 214 818 94 134 18 118 21 784 39 501 244 4140 756 10198 2 615 1836 11 10056 88 4112 6490 4 5863 31 2216 4 256 291 0 475 5 2275 150 6 3 63 57 789 5 0 643 18 2275 6731 537 4 1391 6832 4293 1 
+1,7199 7830 2308 6 2062 4 4118 4023 1734 136 9 3 9847 175 2 2162 0 1563 5590 476 763 31 5369 20 189 2598 310 912 10755 1870 2259 94 3124 0 175 360 0 1290 4 505 4118 4023 6 189 2598 957 54 25 90 977 12 189 2651 1 
+1,74 205 219 271 209 863 4 903 9 0 59 460 364 445 2183 33 6 564 280 60 52 89 90 964 2 0 126 79 9 0 1447 7 570 445 625 131 271 726 687 848 6 1622 1338 1706 1603 6 2221 2346 2344 64 877 61 0 1548 5 0 460 513 1502 1 
+1,1919 684 27 4903 1062 149 414 686 4575 21 520 474 1188 8 33 15 678 2597 2768 3513 232 5 5588 2020 2659 3 482 391 3 409 628 2 0 1207 2965 10 13 2443 12 5101 2022 4 0 913 11 1188 31 3 9411 4 0 1062 7 48 106 816 2 4119 3 1029 5 0 7377 225 1942 3888 11145 123 1064 1 
+1,100 29 804 196 2 253 2 1221 3 540 1925 1007 39 717 349 3200 197 0 107 6 1383 432 1 
+1,11 4857 23 7591 443 41 45 6993 5 3 484 179 55 11349 100 7 38 107 
+3,17 11 4556 6 0 4498 30 5 40 925 5326 35 2589 6 3745 120 7 549 2 2584 119 1 
+3,17 845 1282 1254 1753 5 223 716 1753 5 4524 837 3 38 4541 4 1254 13 394 0 864 2 621 3329 4465 1 
+3,17 2402 509 96 1790 458 50 0 221 6 0 1289 9 383 30 302 2 2583 44 0 202 143 40 65 188 0 2035 4 0 919 2525 1 
+3,273 17 11 4556 6 0 4498 30 5 40 925 5326 35 2589 6 3745 120 7 549 2 2584 119 1 
+3,80 93 17 845 1282 1254 1753 5 223 716 1753 5 4524 837 3 38 4541 4 1254 13 394 0 864 2 621 3329 4465 1 
+3,80 93 17 2402 509 96 1790 458 50 0 221 6 0 1289 9 383 30 302 2 2583 44 0 202 143 40 65 188 0 2035 4 0 919 2525 1 
+0,1855 91 3855 792 5056 0 5169 3 1178 220 320 178 3788 165 13 3677 23 4849 3193 156 8138 11219 2 673 0 434 429 57 4 11 433 1623 7 481 2521 366 1833 13 2 5139 0 3441 122 2 782 249 366 8539 4086 1041 68 854 189 440 519 6 992 3660 23 168 1 
+0,1967 1040 130 25 2149 3 1001 4 217 682 114 1507 10 5281 11206 5 715 2 235 12 16 182 624 178 527 1 
+0,24 435 3 10211 12 3 3133 8230 9938 4384 6254 9 26 2448 8965 19 31 1459 4637 9 0 1898 8262 10 0 1537 13 5 2232 1 
+0,17 11 9287 974 1675 10 1253 0 3136 9 361 140 436 34 404 2 1226 0 1155 3442 4 2653 41 45 724 18 8 86 1 
+0,24 27 314 2000 806 2632 12 1325 6 873 3154 82 23 0 3847 4670 5 2343 8 70 1108 9 0 587 201 1540 3 314 201 317 18 1 
+0,11 9845 4 366 168 220 31 366 9136 0 1134 4 780 3683 2 3 173 1146 156 4643 568 9566 3 3334 71 9393 426 10155 2794 2 871 2017 3716 9 178 6 579 931 114 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 11 433 1623 37 1668 36 37 1731 36 1 
+0,270 13 900 2 1654 3 347 4774 4 4466 1113 3789 2 1789 5310 2 200 203 281 49 2182 10477 3168 4 1005 3135 8770 188 3 591 842 30 1236 930 1696 3789 114 10 8999 3 293 661 4 3589 1113 1 
+0,577 49 61 186 161 242 360 1571 2987 1625 2247 780 1141 3 679 5148 0 152 610 7 309 4 1523 2432 40 237 47 13 1524 3 38 700 5 0 5528 1520 197 727 6 2676 1 
+0,129 6949 6221 24 299 328 7 1033 2108 584 209 2567 20 2759 6 1140 118 230 2725 3 727 19 184 13 2 3528 3 731 7 5016 278 12 99 0 258 1022 6 5245 193 2365 216 3386 477 216 20 549 164 3 1304 5015 1 
+0,117 21 773 2 3285 0 390 7 9960 978 8 5140 3 38 1128 1337 0 4858 2640 4 372 446 2 2466 16 3456 11 6415 1904 6103 3 2118 4 133 220 446 13 1087 2 11057 94 54 167 5140 1212 322 15 512 1102 5405 230 2 200 478 309 4 35 1102 1 
+0,7201 7241 6810 24 2286 10982 13 1485 82 0 4671 15 1632 97 1 
+2,24 577 1050 937 7 1313 761 814 2177 2 504 5 517 61 48 528 3651 13 5113 10 3546 30 590 9 3 234 2 4754 0 2005 39 3 728 4 10672 1 
+2,24 7322 1054 5908 31 8821 9 8148 20 0 960 2387 8 33 39 549 73 73 12 3 186 1344 6324 5 77 1322 1 
+2,24 2832 1286 16 650 4 16 2303 403 12 3 186 73 108 14 3967 8 33 1005 19 3771 1324 446 568 3 386 5 0 2384 467 1 
+2,24 1952 7 7402 7129 124 26 43 3795 3632 403 33 972 2273 7 2338 5784 115 170 115 186 5 0 131 14 0 4046 4235 330 1 
+2,24 947 7960 4042 13 122 9 255 492 14 0 41 45 5827 11 111 7947 13 57 4 77 2168 10361 9 0 630 84 1673 28 14 0 7912 6467 6565 2867 1610 852 4 1349 635 3808 6 3003 2860 4 6891 2424 7836 1081 0 5512 23 75 47 7 2004 10 31 124 20 4181 6453 4 325 1 
+2,74 17 1952 7 7285 6533 550 0 478 4 171 7 4394 4149 5 59 316 7 621 3191 4 0 4918 33 1 
+2,1918 1410 17 260 330 336 3058 4028 181 0 43 3629 592 5 0 41 45 839 534 131 346 33 15 1491 10093 1066 756 3529 9 21 1211 11068 82 1 
+2,74 17 1183 7 92 2616 336 2830 4150 10451 2 2487 83 21 411 33 11177 3 617 869 14 0 59 252 225 282 8708 71 648 30 4494 23 9581 5 1821 1 
+2,74 17 11 88 128 892 787 9183 8 868 9 3 194 1324 59 3640 112 144 102 19 11104 325 169 226 33 1 
+2,3844 17 219 1944 2021 2 3 159 3755 108 4 0 283 5 0 1371 1179 1624 8 33 2 1255 26 3986 79 3 1533 3676 4649 403 1 
+1,17 1956 4 41 45 196 5 375 140 32 656 2 2474 2377 1989 2349 42 46 30 3227 12 4811 667 4 2826 2 1519 9 0 127 7 659 237 5 838 1 
+1,17 4052 1249 5110 14 3 1276 225 269 1271 1277 2 1494 21 1752 420 1666 33 515 14 214 48 94 5 3 2581 3576 4 0 127 7 5492 3501 926 1086 1 
+1,24 1894 0 5408 8494 4 5109 10691 796 6 10150 598 5 0 127 232 1437 14 0 714 8 33 4 100 7 212 1889 304 60 619 10 40 43 3346 5461 4 1086 85 5433 1 
+1,24 11 41 45 1887 1367 8298 0 806 5 0 8480 1882 1857 6 611 4674 4688 1119 42 0 1121 510 76 664 426 1105 991 1 
+1,327 378 1930 1172 1039 7 369 4591 3543 3 5446 4 50 352 2002 8 3466 3686 6 156 3150 2 2501 0 439 4 11046 53 0 2011 4 812 6 135 4690 1 
+1,125 11 176 2295 8 0 1912 1113 3204 1030 3326 5044 76 2 1413 4637 2590 8 3 38 6048 1440 2 8815 1095 2 1490 56 5397 3103 1 
+1,533 272 17 962 2020 2677 2746 8 33 2 456 8 287 2 424 215 1159 101 261 251 71 122 26 2097 2 2579 0 92 7 262 139 67 3314 9 0 119 48 103 1 
+1,3050 17 6834 1259 21 1914 4742 5 100 29 10343 180 601 1571 552 155 1183 771 76 247 1034 721 1999 5 16 2808 2809 259 12 100 1914 150 921 18 33 1 
+1,1377 17 1956 4 41 45 196 5 375 140 32 656 2 2474 2377 1989 2349 42 46 30 3227 12 4811 667 4 2826 2 1519 9 0 127 7 659 237 5 838 1 
+1,7832 2947 17 11 94 4 2947 11276 5 3 38 2544 33 15 0 1155 5823 10291 5584 3 38 10592 6 1819 2 2188 16 707 15 3 3373 9 450 9694 1 
+1,1961 276 4 3174 2 9433 0 5814 5555 54 9727 1145 684 391 3 8507 1 
+1,1 6763 1983 7311 521 35 75 9303 925 14 3 730 4870 5 0 56 84 25 4476 350 4 1100 1 
+1,3895 12 5302 369 9038 15 350 1017 5 3032 83 105 2 804 196 2 154 56 554 444 5 100 1 
+1,299 328 57 4 0 182 1295 3661 5 0 390 7 551 1685 14 214 323 153 5 1434 5 312 1973 1 
+1,74 205 219 271 209 863 4 903 9 0 59 460 364 445 2183 33 6 564 280 60 52 89 90 964 2 0 126 79 9 0 1447 7 570 445 625 131 271 726 687 848 6 1622 1338 1706 1603 6 2221 2346 2344 64 877 61 0 1548 5 0 460 513 1502 1 
+1,1919 684 27 913 343 335 0 913 1062 149 414 686 2207 164 33 1188 14 3 409 628 10 2528 2511 94 1525 2645 2022 11 2262 47 266 3522 3354 3704 6 760 1655 9 154 188 0 66 73 66 784 175 5 3634 1257 1 
+1,1911 1410 1647 1634 13 10953 0 214 4 4450 7 458 1 
+1,533 272 7863 735 60 5 514 1121 33 2 1435 287 2 553 3524 101 261 251 5 558 71 2663 180 6 26 2195 2711 10 2019 148 25 5355 0 2102 1428 6 2572 11080 197 2171 6 902 3798 8 505 1816 192 82 514 3333 6 901 2700 2469 2196 2 2791 817 552 132 2024 1 
+1,273 6716 1121 4 212 1367 6 1394 196 54 32 1305 23 259 5 100 6 375 30 9001 38 3140 12 35 4601 8996 14 168 6787 64 8982 1420 2156 6 5227 9232 1 
+1,7390 745 27 1562 2572 11384 3587 164 21 266 5100 1029 391 7424 8 33 3254 323 887 6 3344 217 1864 1133 2 4887 134 18 1 
+1,1060 1045 831 429 312 1133 611 2219 51 0 2250 215 291 20 299 328 134 8 33 3306 4701 15 173 15 351 153 9 9574 887 1973 1 
+0,24 2863 58 0 156 206 1484 175 13 3 3562 5 1370 427 3 62 10 4786 4802 130 97 193 6 5271 8 19 4830 2 3700 801 6 986 0 2107 19 5276 2 300 2412 333 130 13 3 216 341 5 0 62 84 13 436 2665 20 1469 5428 6915 6534 1 
+0,1 3994 1586 84 701 69 10147 40 971 249 141 7266 9 121 5904 6525 6 7722 0 43 8428 3194 20 8482 5502 11027 3 3170 5497 5 575 1044 86 11 193 31 1088 20 6541 7543 6147 3 62 10 10295 3170 8516 1 
+2,125 6066 203 0 7169 1638 1 
+2,24 3991 285 2335 4263 1015 49 61 297 311 161 2 3 312 303 10 181 667 803 3147 5484 2 4914 893 2541 134 813 11 2383 4049 1 
+2,74 17 11 88 128 555 35 43 316 664 14 0 177 248 2809 8 33 102 1392 1395 775 203 3 3623 1574 1572 11185 1 
+2,74 17 227 1873 3594 219 271 7 3273 4 457 3 159 441 112 665 14 0 151 177 12 3 2754 224 5 0 146 7 2258 666 445 131 8 33 1 
+2,74 17 1183 8783 0 59 10833 8 33 20 10453 2 2495 3 2616 4531 12 21 411 14 0 151 252 1005 141 773 2 4960 0 1723 19 31 1796 10797 12 0 1621 132 112 665 1 
+2,74 17 6233 7625 8882 99 189 8156 2 1239 7476 274 4 0 88 128 5 0 59 288 7 1740 79 903 33 1 
+2,3991 976 2335 4263 1015 49 61 311 161 2 3 312 303 10 181 667 803 3147 5484 2 4914 893 2541 134 813 11 2383 4049 1 
+1,327 378 6510 7255 98 45 1172 501 244 414 2964 31 1277 12 3 10727 1017 33 15 52 3124 0 1692 4 0 379 5783 1360 5 7188 7556 1 
+1,7388 642 683 17 1956 4 692 721 5 411 5026 611 3 1110 454 9 478 980 8 33 42 538 7 142 1120 18 52 701 69 863 155 46 3656 2 769 1 
+1,1919 684 17 1062 149 414 3 913 343 335 0 913 11418 51 3 4806 816 2 40 2645 628 33 6 2207 12 9610 8834 2 1248 3 10668 5 715 2 4810 716 3677 263 180 1 
+1,74 205 957 1953 68 1020 1238 11 356 219 271 68 277 13 108 321 112 665 14 455 177 1 
+1,74 205 117 21 1335 10 31 15 659 15 19 31 2603 0 41 45 146 7 316 79 555 9 241 0 194 72 417 5 0 177 8 33 1574 1572 2 1392 1395 1 
+2,24 5987 49 20 0 1562 6 189 11430 11422 61 189 2114 870 112 3453 3068 852 10075 7035 6212 4 6681 115 73 115 66 5 0 714 664 4 0 151 252 1327 630 1581 1941 181 26 59 1085 12 3 115 170 289 115 186 224 44 6455 7540 4 1171 11003 459 7992 6 819 175 2247 1 
+2,74 17 126 7068 4142 775 79 3448 6831 7877 21 4975 5684 4510 3532 12 21 1238 224 44 313 5 35 714 664 4 0 177 1671 2789 630 8 33 1 
+2,11 88 128 146 7 316 79 555 2 1392 1395 1574 1572 11 386 55 396 0 126 112 144 1746 5 5028 1 
+1,533 272 17 962 2020 2677 2746 8 33 2 456 8 287 2 424 215 1159 101 261 251 71 122 26 2097 2 2579 0 92 7 262 139 67 3314 9 0 119 48 103 1 
+2,3844 17 219 1944 2021 2 3 159 3755 108 4 0 283 5 0 1371 1179 1624 8 33 2 1255 26 3986 79 3 1533 3676 4649 403 1 
+1,24 27 913 343 335 0 913 1062 149 414 686 2207 164 33 1188 14 3 409 628 10 2528 2511 94 1525 2645 2022 11 2262 47 266 3522 3354 3704 6 760 1655 9 154 188 0 66 73 66 784 175 5 3634 1257 1 
+2,24 7163 846 880 613 6238 3929 31 1249 33 44 7989 4 7977 995 2082 797 6 21 3121 613 3929 31 76 4984 5 0 1657 995 42 21 890 280 52 656 2 3443 3541 3216 4 0 3121 0 7634 1889 676 18 1 
+1,1060 1045 831 3017 215 2595 20 299 328 7 1033 2108 2799 1712 164 35 2698 887 8 33 2219 51 2726 3357 6 5282 120 46 55 15 101 152 1509 1304 1705 4 1091 1212 1 
+1,2943 3066 6675 4957 4182 2325 3447 3 111 332 450 2096 132 2024 33 5 21 9513 3176 187 2 228 3327 3477 20 11 1036 378 2325 323 3447 3058 1580 4146 1867 14 0 4104 4370 4380 3890 187 3 3894 2292 2638 1764 1 
+1,273 6561 83 8663 11398 0 637 1394 13 1301 999 1130 5 21 237 47 532 19 13 2901 3889 2268 6584 3 606 54 13 295 0 2005 278 1027 21 4993 606 5 0 1067 1912 13 2555 11322 1846 50 0 150 4 0 221 1 
+1,836 100 41 45 1546 6 196 1135 122 53 0 976 4 253 6 1670 12 540 1006 33 3572 598 5 0 484 179 118 15 1437 5 1170 1286 3 508 2639 2 34 3 2623 5 0 127 7 1483 926 1086 1 
+1,533 272 4322 20 2469 860 6 3333 667 4 962 735 60 5 1845 1121 33 2 456 8 287 2 584 1762 101 261 251 23 489 6912 10781 9 1675 5 11266 5128 225 2760 13 173 10 0 1762 294 105 3 942 3672 8845 6 5 0 10765 225 563 9 26 2711 9 0 902 13 9153 1 
+1,1184 11 411 343 14 0 976 4 111 1605 1909 1375 1362 1 
+2,74 17 100 7 9122 59 145 867 8 33 102 46 355 6191 7453 66 226 2 673 0 158 1455 4 0 151 252 1 
+2,227 1873 124 0 112 144 33 5 0 146 7 570 364 445 625 12 3 92 159 72 4 170 1008 739 632 909 1 
+0,1 78 486 2066 78 486 3 844 7321 1058 118 57 71 48 8677 4 3923 42 1505 888 4 350 4 987 1160 4 114 6591 51 366 1290 701 69 32 281 450 6101 76 2597 4 350 4 987 9 792 2 2458 8 114 78 31 3232 6 89 3 10118 9 1429 10 78 2499 69 982 78 2732 1339 2 1117 6 2499 69 8029 2 556 9 114 6940 78 510 40 102 78 31 3 9647 249 9438 78 124 69 34 10317 78 436 763 10 2546 78 510 31 9053 78 701 69 494 245 450 60 4 0 1742 4 0 114 240 6 78 928 502 8658 7870 49 2 0 240 61 3 547 2597 2768 5 114 191 542 13 5471 1 
+1,24 272 7 67 240 29 1242 3 38 1291 4 10866 6 449 39 33 7 319 8 261 251 7 5206 0 267 4 0 150 145 67 62 18 33 1 
+3,1 6145 2425 13 3 4907 5608 54 4679 1282 1254 9 17 5845 10020 5 0 3201 30 10796 426 4 2368 2425 1 
+2,24 944 7157 1839 3 1425 173 12 133 1462 360 3 222 274 546 5 1354 7 441 145 1240 340 6 0 1934 1799 9 21 351 289 224 44 0 849 827 1049 8 33 1 
+1,125 2928 31 8 173 8048 15 3 229 1474 2 1348 1626 2164 1154 2 475 10734 60 501 244 4303 2278 15 3 286 39 0 9098 4 3 1700 9 7477 2 1795 16 196 60 4 100 1 
+1,24 6078 21 2608 169 106 438 9 6744 7 321 1445 2790 636 2241 594 5133 149 2301 543 0 152 610 8 33 4 256 1999 20 872 6 786 9580 14 0 2548 4 934 3322 1 
+1,6440 5 253 1154 2 11227 3 508 2 4593 3 420 1666 1 
+2,24 3868 4348 1414 57 145 5 401 889 9 26 43 108 5 217 3 162 6 1258 575 1044 1081 3 77 137 1831 4 0 2383 3008 12 3 170 73 224 33 1 
+2,74 17 227 1873 3594 219 271 7 3273 4 457 3 159 441 112 665 5 3 1443 6 1431 106 4 1544 14 0 151 177 8 33 1 
+2,80 93 24 1939 747 209 3 133 2594 53 0 1013 340 2 154 0 1353 1895 157 3 1012 137 1274 2205 33 972 1201 4350 6 0 111 332 1382 66 226 1 
+2,6330 320 122 3873 7303 29 1774 26 5306 283 39 8879 3 5351 5 21 2050 137 83 0 111 536 4207 1 
+1,11 1845 456 31 5374 15 514 8720 3431 51 14 5567 8319 7444 64 76 302 631 268 451 486 71 789 1 
+2,1918 1208 17 126 834 1611 2747 8 1331 9 26 63 216 403 5 0 41 45 839 534 131 346 8 33 1485 48 2727 2006 12 1012 2070 2 521 1 
+1,1185 1163 2978 9318 64 1238 10283 9 3 546 4774 4 201 9596 11 10638 9157 10767 5810 3731 1 
+1,74 205 117 21 1335 15 659 15 19 31 2603 1958 2891 1876 4063 6 0 1020 4 0 41 45 316 79 555 1574 1572 2 1392 1395 8 33 241 0 194 59 386 417 9 324 6 16 43 248 3095 3544 1 
+2,24 6493 2312 5933 55 34 5355 631 0 3774 283 3482 2398 459 83 0 826 3975 12 3 173 215 4442 10849 1 
+2,24 5791 5946 6 4294 4009 1260 6 1201 6447 246 3 222 274 3013 1143 5 0 1240 340 15 0 1052 1167 3944 1799 9 21 268 139 224 44 0 461 1175 8 33 1 
+1,7811 7112 17 11 41 98 1199 2871 1277 5 1716 1527 8592 2857 8 33 879 0 5083 4 49 61 2807 4359 3909 2157 14 3 1681 5 681 2857 1 
+0,24 111 536 7 8270 2483 11393 140 34 658 8 94 42 46 32 90 278 4947 8 21 936 3404 10 7 1838 2 4727 9399 57 4 0 10724 3703 4 0 233 7 2061 1 
+2,24 7558 4070 3 632 47 266 23 826 1073 0 43 1222 2 108 0 5648 3632 2004 4 0 41 45 4341 1037 5992 7140 1 
+1,17 1627 196 970 5 298 33 15 0 43 881 584 160 2636 2 801 1779 1776 3 2725 1686 616 197 0 576 107 6 907 5 0 1803 681 233 1 
+1,6320 2900 315 17 1627 196 970 5 298 33 15 0 43 881 584 160 2636 2 801 1779 1776 3 2725 1686 616 197 0 576 107 6 907 5 0 1803 681 233 1 
+2,24 4120 6089 1839 0 2178 12 3 1013 340 1143 6 10178 2594 6230 6337 246 3 48 145 546 5 0 1863 340 10 554 0 312 6999 44 0 7060 6000 139 170 33 1 
+1,836 100 41 45 1546 6 196 1135 122 53 0 976 4 253 6 1670 12 540 1006 33 3572 598 5 0 484 179 118 15 1437 5 1170 1286 3 508 2639 2 34 3 2623 5 0 127 7 1483 926 1086 1 
+2,17 1939 747 918 60 740 3138 5 268 73 170 889 2 154 0 1353 1895 157 3 1012 137 1274 2205 12 3 66 226 108 44 0 111 332 1382 8 33 1 
+2,80 93 17 1939 747 918 60 740 3138 5 268 73 170 889 2 154 0 1353 1895 157 3 1012 137 1274 2205 12 3 66 226 108 44 0 111 332 1382 8 33 1 
+1,24 11 38 41 98 8015 9 6840 209 982 33 4 0 4291 1511 84 25 3575 8775 4766 966 9054 1473 248 0 157 4 3 3761 259 1 
+1,533 272 4322 20 2469 860 6 0 8836 4 514 3333 962 735 60 5 1845 1121 33 2 456 8 287 2 584 1762 101 261 251 23 489 845 1473 14 2677 1293 2055 9 73 352 1480 2754 278 3723 237 2649 1 
+2,80 93 17 2870 2939 4616 3 48 145 997 5 0 176 4 0 1863 340 2 154 0 835 1403 783 8 2 43 422 5 0 126 467 940 12 3 186 66 1307 108 44 0 1358 952 1 
+2,1647 1634 25 1518 9781 10547 188 0 1389 2313 27 534 84 13 3233 5 3 77 343 1498 197 834 1611 1066 756 6 1356 1595 1 
+2,1392 1395 11267 0 88 128 1574 1572 14 0 146 7 316 1502 8 33 0 43 386 14 0 252 9 0 77 72 870 112 9839 248 2809 1 
+2,11 3570 28 7087 8660 29 626 2 40 1786 119 65 15 29 0 5582 2130 1 
+2,2870 2939 246 3 48 145 997 5 0 1863 340 15 0 1403 10068 3 48 145 2031 9 3 186 66 108 33 884 203 3 48 137 231 44 0 63 422 952 1 
+2,497 821 3517 2390 7649 2087 160 7 184 72 2 9509 0 1354 1934 1458 489 573 118 1011 2 2203 8 0 9931 174 614 1 
+2,17 2870 2939 4616 3 48 145 997 5 0 176 4 0 1863 340 2 154 0 835 1403 783 8 2 43 422 5 0 126 467 940 12 3 186 66 1307 108 44 0 1358 952 1 
+1,533 272 17 962 2020 2677 2746 8 33 2 456 8 287 2 424 215 1159 101 261 251 71 122 26 2097 2 2579 0 92 7 262 139 67 3314 9 0 119 48 103 1 
+1,842 30 1236 3 1708 84 55 3394 0 2091 4 9649 8843 2122 1 
+3,533 272 17 750 655 20 272 32 76 90 4762 20 0 319 8 0 561 4 101 261 251 1531 1316 18 589 33 1 
+2,24 11 241 3151 1066 756 181 99 106 31 0 241 57 10 9831 1350 42 3220 60 4 0 839 534 12 3 10351 10 656 180 756 209 3102 4 3 589 1231 20 834 1611 2 167 53 3 77 234 1498 33 14 1647 1634 564 181 0 241 3151 44 0 77 876 2070 2 108 0 131 216 4 0 47 1 
+2,1066 756 10055 834 1611 6 1356 1595 5 3 77 234 1498 2 108 0 839 534 8 33 14 1647 1634 5 6635 1208 1 
+1,24 41 45 1546 6 196 1135 122 53 0 976 4 253 6 1670 12 540 1006 33 3572 598 5 0 484 179 118 15 1437 5 1170 1286 3 508 2639 2 34 3 2623 5 0 127 7 1483 926 1086 1 
+1,273 2882 2397 2296 7490 8344 26 314 3221 44 0 736 8 41 45 105 2 1139 16 444 2745 263 0 3359 5 138 1684 2680 1994 203 4612 2 752 7 8321 1 
+1,11 1676 1010 20 0 634 476 15 0 3754 3000 2351 5564 3 5469 5 2413 1 
+2,24 1356 1595 6 1356 1940 1450 2619 5037 1126 5 0 75 346 4 0 839 534 532 120 510 46 167 9 35 1095 1384 4 0 398 1595 6 1940 901 1328 53 0 1126 2666 753 535 12 35 3505 33 5 0 47 7 131 216 1595 1101 63 39 3 77 343 77 3381 1498 6 1940 31 384 427 447 614 2 9662 1376 1600 6 1201 6458 60 4 0 176 169 5 0 753 535 1125 3654 1 
+1,27 1295 1423 709 14 214 819 94 232 4 203 3600 6 1861 50 531 716 15 46 4470 9 21 2331 1041 2667 33 5 0 2655 150 4 2275 476 18 1 
+3,17 750 655 20 272 32 76 90 4762 20 0 319 8 0 561 4 101 261 251 1531 1316 18 589 33 1 
+2,24 219 271 5447 701 69 1986 9 40 1602 1797 2 108 441 112 665 13 44 3267 20 324 7 1564 1314 5 0 570 364 445 625 2314 2320 6791 13 11112 82 532 210 698 0 11140 282 4 99 83 687 848 1058 9069 120 271 89 5 2103 102 52 2029 2 2003 957 1953 1 
+2,24 1394 2570 7771 7162 1758 5 3 9386 57 1482 53 5417 72 295 0 2880 6028 2 3 73 226 224 44 0 43 422 7046 8 33 1 
+1,327 378 1035 2900 315 24 2298 4 1627 810 970 5 315 7 1556 298 233 33 0 43 881 1417 584 3252 5 0 643 248 743 1481 611 3 3558 83 1222 633 2558 515 350 1 
+3,17 11 702 442 2 3 617 133 65 363 5553 0 1718 8 28 39 3 11380 4 0 41 45 733 774 2 159 2628 1018 458 50 359 9543 5 0 88 128 6 3 719 2193 5 0 221 1 
+3,638 17 1206 7 1386 442 49 61 57 113 20 2642 1009 8 28 15 485 3658 51 803 4 2054 322 15 1960 1617 97 1214 448 706 4 0 2074 4 2213 67 96 8 0 483 221 1 
+3,4209 9 0 433 643 7 176 1098 560 2730 75 47 10546 0 8922 3705 10 192 5 12 0 10410 5 1567 1 
+3,7753 2 0 3023 3621 210 167 3 174 1151 3511 4 230 35 202 13 2524 84 2592 5 9059 875 556 1 
+3,462 1917 1047 7668 1078 6 612 875 4 6406 97 217 66 139 104 5 5318 75 47 31 118 0 2453 1 
+3,2962 4 6190 433 7 1098 6 1422 1119 32 808 3240 12 0 275 46 10077 1 
+3,6441 60 230 281 1098 167 1015 68 34 166 5204 0 1859 21 9533 727 1 
+3,637 5216 629 5 0 433 643 560 75 47 14 16 2593 904 248 0 2252 187 2 3 596 2 34 296 195 1233 2225 4 350 4 588 6 1264 652 449 4758 5 2372 1405 1 
+2,24 2414 3081 506 51 2539 5 3 216 333 42 52 7 184 0 262 73 1514 670 5 0 92 1 
+2,24 6931 7744 2411 288 7 316 613 1038 7928 25 473 2 3 38 77 47 768 12 0 3485 9 77 49 103 0 676 172 33 1 
+2,24 5994 7233 492 21 268 243 2260 8 33 2 108 0 4171 1959 7 3011 3921 6112 534 9 26 43 4171 224 1 
+1,24 7620 3661 9614 111 1649 7 384 580 179 4 6313 8 28 1692 216 2172 10715 3601 6 2557 6 9244 9972 9677 1 
+1,687 848 6 219 271 29 8460 112 5 0 146 7 3764 445 8 106 77 4 0 177 1 
+1,74 205 117 21 1335 15 659 15 19 31 2603 1958 2891 1876 4063 6 0 1020 4 0 41 45 316 79 555 1574 1572 2 1392 1395 8 33 241 0 194 59 386 417 9 324 6 16 43 248 3095 3544 1 
+2,4248 6488 404 3 133 1763 5601 2 8819 26 2966 79 7 3479 898 5 33 7 1793 8273 1 
+2,219 1944 124 0 1371 1179 1624 33 5 6017 2186 1032 255 159 20 1672 0 43 2906 469 4782 12 459 5558 5 3 283 1 
+2,435 19 698 2 903 9 0 790 2041 4 0 6887 7008 4341 2865 2391 6820 2087 10 3681 1545 10259 6 9248 30 396 8 783 1 
+1,11 508 31 5479 53 8880 20 1437 10870 1510 83 0 41 45 554 388 2113 5 253 1 
+1,11 2306 326 78 25 90 3555 2034 5 21 773 2 4899 1031 1510 14 0 606 212 1591 1 
+3,17 1206 7 1386 442 49 61 57 113 20 2642 1009 8 28 15 485 3658 51 803 4 2054 322 15 1960 1617 97 1214 448 706 4 0 2074 4 2213 67 96 8 0 483 221 1 
+2,24 826 3975 3248 157 3003 6829 5392 26 674 2619 5 1793 8 33 102 3 9737 1135 8 26 3423 1 
+2,24 542 13 120 1244 31 3734 9 23 2425 2971 353 820 389 4 10430 6 4675 2694 50 0 3665 4 26 215 2619 0 6022 3859 320 122 993 3547 138 8092 2971 89 0 222 274 2178 8 3 73 3750 145 5 26 1189 2682 1085 1744 0 3859 2 3 819 115 108 44 0 3936 3872 8 33 1 
+1,4276 6502 45 6588 2417 3043 6331 5534 1743 5251 6 1031 2799 10 9150 3703 6 10562 82 10579 33 5 0 4790 1882 310 888 4 94 1819 10448 5 10692 1 
+1,1185 1163 11 150 3125 996 7 489 25 3211 10 219 1916 31 76 9802 20 5345 7 8813 54 209 180 53 3234 75 47 8 2479 2647 647 1880 1056 147 33 11 3328 64 4653 5 3 77 1786 1763 2964 27 1 
+1,7718 98 1047 2324 661 111 1605 3935 18 33 10 1126 13 974 335 1119 4 1909 1 
+1,1060 1045 831 3017 215 2595 20 299 328 7 1033 2108 2799 1712 164 35 2698 887 8 33 2219 51 2726 3357 6 5282 120 46 55 15 101 152 1509 1304 1705 4 1091 1212 1 
+3,638 17 1206 7 1386 442 73 2261 113 20 5098 8 28 3315 2094 53 3 194 106 15 448 706 50 0 2074 4 2213 67 96 8 0 483 221 246 2054 322 15 1960 1617 97 1 
+0,17 830 58 1893 236 0 92 7 199 470 342 18 8 28 19 25 215 0 363 157 381 470 143 5 171 6 265 16 1784 449 286 9 0 127 40 47 416 2 5413 438 5 0 5331 1 
+2,4292 797 10060 1066 756 8 33 42 5132 5154 180 14 0 2263 1389 2313 27 534 1 
+3,571 17 88 5883 3082 7808 18 16 1119 14 2284 58 8 33 1157 2 5269 0 2013 873 342 7 768 2146 0 63 72 40 47 529 32 1157 83 3 2284 768 488 1 
+3,4274 17 7124 509 67 3351 37 6054 36 246 3 38 159 4 2812 3780 8 28 8 458 50 719 1846 6 4759 2 67 1543 15 962 1157 5 3 319 8 287 2 424 101 261 251 1 
+2,1918 1208 17 3988 7 1066 756 304 26 4629 2 108 0 2263 41 45 839 534 5 3 77 234 1498 8 33 4607 0 194 216 403 4 26 1425 1 
+1,17 413 689 13 1838 2 309 1836 83 0 227 3032 7 1113 354 18 5 3 1515 183 2433 8 28 6 680 227 1051 2480 5 171 6 3038 496 2 34 8 35 4936 1 
+1,5893 375 37 1677 36 1612 2788 1448 3100 5 40 7269 4368 1852 7404 5938 3569 2 456 528 380 6717 557 9022 13 249 173 9 0 2989 374 594 237 0 43 5 935 551 10 350 4 94 5 0 11286 32 2164 1 
+3,571 17 11 790 2236 14 2284 58 18 16 1119 1157 8 33 2 5269 0 2013 873 342 7 768 2146 0 63 72 40 47 0 529 32 1157 83 21 488 1 
+3,3025 17 537 1924 0 92 7 199 4479 29 453 803 16 43 181 5 171 6064 5 382 741 5 3 143 19 2051 2049 2 3089 9 3 1453 4 483 191 4 0 3440 1424 1 
+1,6034 745 6532 96 32 1444 217 139 2002 5 0 421 77 528 12 21 708 5 1543 42 1535 509 67 96 55 646 1130 2 627 333 1144 21 240 2438 18 33 1 
+2,4346 2312 3995 89 90 3563 9924 39 1279 0 265 20 3 2752 5 0 2263 839 534 8 95 160 31 778 9630 8 33 1 
+1,17 11 88 128 29 694 0 1310 4 133 3803 1278 650 5574 2 1401 119 47 12 1705 362 2 504 5 3765 3 1284 18 8 28 1 
+1,17 1206 7 1386 442 73 2261 113 20 5098 8 28 3315 2094 53 3 194 106 15 255 2212 5 67 96 8774 458 50 0 483 652 2074 6 1758 198 2054 322 15 1960 1617 97 1 
+1,4330 17 11 88 128 29 694 0 1310 4 133 3803 1278 650 5574 2 1401 119 47 12 1705 362 2 504 5 3765 3 1284 18 8 28 1 
+3,1964 1922 10934 20 2018 4 16 264 5457 13 934 2 3394 16 707 20 5444 183 8350 1 
+3,6934 166 0 1195 583 787 25 920 0 237 568 9076 598 216 3188 1846 71 11253 3006 4225 9039 6504 993 25 10 8984 1445 2097 6 0 474 4 2779 10 10154 4139 5 0 528 295 51 2 0 237 25 1105 234 2 9 138 21 474 4 3172 5149 71 9 716 2707 10 1285 417 1230 281 5 0 1195 6 10 0 347 9704 6 3349 1230 2654 2 1809 0 127 30 174 2727 1 
+2,24 7827 1548 7166 7837 1702 126 4106 6578 66 115 115 186 115 66 5 0 131 4 0 4369 330 8 33 1 
+1,11 2306 326 78 25 90 3555 2034 5 21 773 2 4899 1031 1510 14 0 606 1432 1 
+0,11 2023 6 344 114 62 13 192 2 2710 2513 8 3716 2 16 2059 395 6 10057 4244 105 1 
+0,7641 8 308 29 49 61 546 40 47 12 3 448 3110 1080 23 0 798 2180 1 
+0,941 13 192 2 425 2036 8 2291 7 5344 10328 49 438 9 463 1 
+2,4285 7472 1260 1844 2 477 0 43 1618 670 2 673 311 7389 40 283 15 0 1064 1379 1587 4778 0 826 1585 169 186 8 33 110 1 
+2,1585 537 4138 149 7552 10712 26 4708 49 61 2239 40 421 10013 249 2527 64 173 4320 3 2003 2095 40 283 99 0 5454 1 
+1,17 2413 7 63 9423 9918 0 3000 2351 25 5563 6899 8 3 1515 2776 2 0 4036 233 20 0 1272 6018 2798 2181 5 5660 31 5082 20 2497 39 0 9086 6226 2351 4594 255 1676 1 
+1,2940 1608 6 1607 1639 140 76 34 9186 614 2 8161 0 885 53 35 713 990 922 1 
+1,428 5532 10946 1025 2 0 56 1241 5 0 4136 359 2941 2951 5049 2 3 142 2719 1 
+0,1967 1040 639 30 9763 4 447 1517 9 487 4446 1 
+0,1967 1040 2991 165 7 72 9 135 500 3721 2 9216 471 6631 7 231 6 488 10123 4997 2148 2 35 300 1 
+0,1967 1040 6267 2 2090 2684 4416 1328 8952 5 2166 2 799 71 525 4717 568 785 5019 0 765 9925 193 55 154 574 553 51 12 984 9 447 697 1 
+1,27 10111 1527 416 2 32 3144 8 28 5 0 4137 25 90 3525 4994 1 
+1,125 4314 4102 7 7727 2414 907 18 46 29 10856 245 38 107 10316 2 10550 717 349 76 884 203 2182 561 15 10131 7184 31 192 2 495 2 9638 0 2111 1 
+3,117 3 438 8556 10996 6 2750 4675 4537 10805 30 1463 2 479 3 809 3512 4 116 5 21 695 40 65 1 
+1,2935 2945 17 7694 1560 1501 25 90 280 5 21 4818 10 970 14 0 41 45 1241 5 1380 6 77 1539 1119 10359 21 1241 317 6 369 18 8 28 27 1763 31 3250 10 89 3 2761 1501 5 19 18 21 1241 1820 5 2941 2951 11 1501 13 256 1025 9 1152 1 
+3,638 17 1206 7 1386 1981 442 226 1345 113 2 3 617 77 162 1692 363 8 28 15 509 67 96 333 246 159 1109 5 936 1332 552 8521 0 1289 9 0 483 221 1 
+2,24 4025 4324 89 3 447 2073 120 1757 4 41 45 79 52 85 494 2 0 753 535 1350 10 865 5 0 131 346 4 0 839 534 1 
+1,327 378 3025 171 24 3070 5 21 4790 1511 1712 12 3323 1554 28 2 208 9 822 94 1279 5 791 5040 5534 20 1962 3015 84 25 502 90 8274 9 3752 3241 6 13 0 1564 322 731 2 246 171 5 321 103 1 
+1,11 63 885 5 0 814 4 77 406 543 4 5493 6 320 3 3411 1673 5 2341 1 
+3,4274 17 3971 554 3 1245 5 936 400 8 28 15 67 96 192 38 1109 391 1343 6 301 1314 0 41 45 733 2031 3741 2 3 159 1018 617 2492 50 0 782 4 0 92 7 199 221 1 
+3,17 3971 554 3 1245 5 936 400 8 28 15 67 96 192 38 1109 391 297 1343 6 301 1314 0 41 45 733 2031 3741 2 3 159 1018 617 2492 50 0 782 4 0 92 7 199 221 1 
+1,17 538 1436 2692 259 8 1110 1827 692 721 8 28 595 19 85 3131 1479 796 35 1688 2 495 2 859 35 29 1 
+1,125 1961 634 11166 10612 32 90 9349 198 5 1145 1955 15 368 4 2184 598 84 25 1690 49 61 5686 2091 248 0 453 4 0 47 134 18 1 
+1,1184 17 538 1436 2692 259 8 1110 1827 692 721 8 28 595 19 85 3131 1479 796 35 1688 2 495 2 859 35 29 1 
+3,11 1914 2120 2790 83 257 2669 4 3 10443 318 2435 14 714 0 221 2 881 1754 1 
+3,6509 13 9597 10427 1317 4365 2006 2192 426 5359 8236 2 2042 23 10621 68 548 121 8 3 9416 320 11146 5 3 5033 8 0 1780 71 8 3 2580 9121 225 121 491 734 2 9116 3 5676 3750 988 53 0 1562 1 
+3,1646 31 0 75 5384 185 220 181 165 140 32 8992 3 8473 102 440 519 970 71 3 9402 11013 102 121 1823 51 178 1 
+3,435 6726 58 192 0 1550 4 16 586 183 394 5 2360 19 924 19 31 256 3213 9091 400 2 425 9 169 2 459 8102 1 
+3,3837 6984 6921 2620 26 1425 31 14 1319 11 738 3602 89 118 1101 235 8 21 1721 4570 5160 8 0 9289 4 3 10106 10 8429 9791 837 52 6 26 3417 1440 1450 3 2017 649 225 2 2210 0 971 9 10334 1 
+3,11 2085 7 184 60 8 287 3 220 68 417 3707 34 5012 42 160 7 141 1129 10 19 68 32 1777 5176 165 7 118 3 1769 4 114 1 
+2,464 0 7474 7673 64 3377 2 5607 0 11098 2740 9 40 47 7 471 331 0 229 387 1490 57 4 35 8497 4050 4 87 4282 87 46 55 222 12 87 7574 87 757 7 213 141 1769 230 658 46 495 0 331 68 69 167 141 5337 5 35 1797 9 0 176 1538 5 0 1561 862 282 1 
+2,165 7 3 507 52 55 32 181 49 8669 12 562 135 61 3 779 3682 11045 2 0 3517 9 0 131 60 4 1032 255 5 3 3609 4812 728 4 57 145 2094 9 0 471 331 211 4197 3883 1340 1100 2 763 40 462 99 0 9270 11011 52 25 181 5 26 48 65 9515 15 3 1 
+2,7132 1356 4236 4 0 1064 1379 2852 802 60 4 0 379 535 4 4037 109 12 21 11229 1471 6 31 10498 8 0 327 79 20 1376 5990 4 0 312 3001 4236 7 649 698 517 61 3 65 132 0 327 79 2660 1842 1681 11 79 29 1793 9 0 43 72 95 5 4199 1 
+2,3868 4348 1414 57 145 5 401 889 9 26 43 108 5 217 3 162 6 0 575 1044 1907 2007 5 8 0 1561 862 231 12 3 170 73 224 44 0 3008 109 5 2383 4626 3 77 137 1831 1 
+2,11 471 331 32 62 1979 0 1561 862 3654 117 3827 960 6882 7172 1260 6 989 5 77 675 360 3 11115 10604 3339 6 0 2387 355 0 3051 2846 3940 4253 115 66 109 2 1831 0 77 137 728 1 
+2,11 3184 737 51 0 2079 2425 2971 209 0 482 1604 7 10 9 1126 1 
+2,3973 2968 6905 555 26 8110 40 919 6 15 3 1134 2695 3 547 2681 42 52 701 69 2093 57 3490 4 26 11401 1 
+2,3018 613 1201 4183 13 76 348 1698 9 0 8991 4 1581 2297 211 4183 7 11030 3469 1258 422 2297 5 3 1792 2 2178 26 43 48 3361 15 3 2143 5 0 131 909 4 3 170 226 108 14 2880 70 6 8 26 43 2229 4 136 110 7 137 14 1361 938 88 9 0 3018 7 63 1 
+1,1384 576 2157 4887 290 0 1223 53 2285 879 723 4 2163 598 5 0 298 233 1 
+1,80 1892 27 1295 1423 709 14 214 819 94 232 4 203 3600 6 1861 50 531 716 15 46 4470 9 21 2331 1041 2667 109 5 0 2655 150 4 2275 476 18 1 
+1,1184 11 411 343 14 0 976 4 0 2707 4 111 1605 7 1737 1375 1362 4144 44 3 1735 4423 18 5 21 620 2149 109 10 52 13 1324 6 89 141 2073 14 43 10 26 2462 13 9429 1 
+1,5752 1636 6044 6216 6497 684 109 9432 810 360 2225 4 350 4 5798 54 5401 21 761 8 0 409 7461 1344 103 380 2 859 0 7144 4931 57 4 0 214 5273 388 1287 4 379 1407 686 1 
+1,1919 684 27 913 343 335 0 913 1062 149 414 686 2207 109 164 33 1188 14 3 409 628 10 2528 2511 94 1525 2645 2022 11 2262 47 266 3522 3354 3704 6 760 1655 9 154 188 0 66 784 175 5 0 3634 1257 1 
+1,533 272 17 1957 640 107 3460 18 8 28 101 261 251 89 3282 920 3 319 8 287 2 424 180 42 35 2010 8610 12 1743 669 4964 46 89 124 1 
+0,7364 9 3389 3169 6 3 2133 1519 30 5147 20 407 7 385 2244 1 
+0,156 206 1745 62 6338 123 63 158 3435 32 90 265 3206 12 0 398 1291 3 47 380 1 
+1,17 640 101 261 251 25 920 3 319 2 424 180 187 2 345 296 20 1445 476 8 28 12 2820 113 4 0 456 3220 1 
+1,125 4136 1716 399 10601 2 0 11117 3368 56 1241 5 2941 2951 39 3 1501 84 369 18 55 34 3112 31 280 5 3 1763 1 
+0,27 9238 372 137 1971 2 3285 978 50 0 483 615 83 1110 1 
+1,533 272 17 640 101 261 251 25 920 3 319 2 424 180 187 2 2136 345 296 20 0 127 7 176 1445 1781 8 28 1 
+3,17 750 96 1756 2 3 38 159 173 391 297 1343 8 28 12 3697 8 11082 9 0 1134 4 272 7 736 319 6 100 7 655 333 8888 20 3 1200 1268 2238 5 0 3648 1 
+2,74 2892 109 7 2475 4 0 3122 4 0 1207 2965 0 685 10043 8505 89 3 10894 3576 9 0 3488 4 0 59 252 262 1769 120 0 4422 6 10826 148 50 205 7 5104 1645 10038 40 127 13 184 0 3266 4 16 8768 2772 11213 150 10615 10481 7639 44 0 6033 4587 4 3 1 
+2,6983 6549 7 82 3130 492 215 26 1255 118 3 10838 63 132 0 131 8386 3278 164 0 8229 2 777 2273 21 3783 3782 224 44 4290 4157 109 8 0 43 106 4 59 146 7 316 5 151 1 
+2,693 861 7991 4265 6431 217 802 82 120 4459 85 32 90 0 580 1335 5 7111 551 3332 63 39 916 75 109 14 7889 6554 98 7938 465 1 
+2,11 56 146 7 6 288 7 3288 802 82 514 5558 5 109 7 59 5311 2591 1094 2186 92 4506 2 3101 4743 2 33 7 1455 1 
+2,74 11 1447 611 8 3 198 5137 9 0 56 1544 79 6 8817 23 160 1177 92 336 6776 4345 2207 168 1453 5 0 311 364 4560 2412 92 159 3380 2856 2321 31 3171 20 313 7 4095 2938 5 0 311 4538 6443 0 146 7 186 11435 311 445 625 1101 194 291 227 1873 6 0 1 
+2,1641 5889 11377 3 75 63 492 421 6201 9316 6477 7849 2 777 0 126 599 10228 79 3 289 115 224 5 16 630 3482 109 1 
+2,74 2884 313 7 4095 2938 859 0 1136 102 52 355 92 159 3380 2856 2321 20 632 5626 4 3 63 5 109 7 59 311 364 4538 5780 5005 2321 7 56 5463 54 1690 2938 10586 1337 0 1111 2526 5034 1 
+2,74 435 0 59 3640 2176 31 296 4113 3984 3365 3 9818 6 611 4674 198 0 276 87 542 137 13 1416 8 366 8390 87 912 18 87 78 32 3 11169 551 12 325 87 532 631 109 2650 3 9382 57 1 
+2,74 639 7 241 1806 9 57 6135 5 3 2464 1805 757 31 0 10606 3425 3816 5955 3422 75 110 15 26 174 1307 2 0 59 252 506 2 21 7974 157 14 0 4214 2854 2320 5 26 43 615 5 0 5585 1441 1177 615 10 13 3713 121 1156 120 19 209 9 180 1 
+1,6645 443 9040 5 1046 7 3159 227 1192 233 10687 3 48 106 1687 1 
+0,7246 7861 7101 5421 877 372 1790 3126 4585 2 2395 190 1 
+2,74 17 219 271 57 112 124 42 57 4 26 441 403 3179 210 555 1545 8 687 848 6 1622 1338 1706 1603 5 35 174 3128 1814 5 0 460 666 445 131 8 28 1 
+1,2947 7 4229 6620 3799 1742 44 558 2 26 2739 6 9605 0 5589 390 2 3 9276 622 1 
+1,533 272 101 261 251 1071 2 32 920 3 557 319 2 2663 180 187 2 382 345 28 310 272 7 669 11008 1690 995 5907 4 0 1762 3524 267 192 82 4881 6 611 4576 5 0 917 4 0 359 5 0 1296 2024 8750 5539 885 0 306 23 1044 6087 267 4 0 212 6352 2871 1 
+1,1060 1045 831 2421 908 1149 786 2430 6 212 1367 196 64 1810 290 312 2 154 1133 1302 23 0 4552 4 299 328 0 1564 731 2 246 0 150 5 3 1093 103 1 
+1,24 27 409 126 2614 25 3259 966 12 26 269 11159 5 0 1145 269 179 4 7136 0 10329 8814 1737 18 28 1 
+1,125 27 2494 554 20 260 2605 7373 7252 25 1444 16 645 9 3 216 227 1051 150 3120 9798 62 1 
+1,24 101 152 7 318 2 1228 2225 4 350 4 41 45 196 168 23 499 6 496 55 1733 180 237 47 3115 23 388 3322 42 124 69 1715 0 5419 8 810 184 4497 1031 2556 5 100 6 375 1 
+1,1961 2800 1549 21 431 228 46 1311 850 8196 2382 6408 49 61 3 65 39 52 3259 1 
+0,3942 3463 4 21 659 806 30 2016 15 368 4 16 10527 2 922 155 19 29 10993 0 800 1 
+1,24 2967 1198 310 5445 4 101 152 7 491 237 30 9886 4 0 5146 2016 284 149 329 355 180 5 1928 7 237 0 150 7 43 1947 11283 5 48 2028 6 3 864 2 859 0 636 5107 8 16 459 1772 2507 5 1360 1 
+3,56 134 4754 582 4 2220 9061 9726 14 11344 7696 20 48 238 1717 1 
+3,7616 2098 5 0 850 30 3670 5 0 279 4 1535 67 96 6 293 999 1130 0 6040 123 1 
+3,130 13 5418 94 776 783 4 3 257 142 527 1557 487 206 992 2187 525 1 
+0,7196 6698 2649 9018 5131 178 525 30 545 5 9878 8380 1 
+1,24 11 1676 4595 20 0 634 107 15 0 10460 4 0 3000 2351 31 4929 11335 20 2039 4 9919 102 52 5563 2413 7 182 10603 5469 150 921 147 884 1515 10336 2 57 4 26 8107 1 
+0,5932 58 4 7023 124 3 459 104 768 23 0 4351 1199 1578 2 504 0 194 2124 5 0 354 7 522 2 1082 3 1320 4977 862 9 41 45 5504 1717 1 
+0,637 5216 629 5 0 433 643 560 75 47 14 16 2593 904 248 0 2252 187 2 3 596 2 34 296 195 1233 2225 4 350 4 588 6 1264 652 449 4758 5 2372 1405 1 
+0,11 1701 44 287 275 284 2231 202 3486 15 21 2548 2528 3462 5254 23 433 643 1227 138 4 84 32 3 662 2 2093 155 0 3090 561 1230 1 
+1,1184 1384 61 3 2768 692 721 737 8 3 1110 454 2 490 9 478 980 42 538 5287 20 5483 9556 5289 6 3 3882 1120 18 52 5605 69 863 155 46 3656 2 769 1870 73 2259 721 918 8 33 1 
+3,11 603 1174 6 101 947 152 30 5 3 259 4 5602 44 3674 1105 2 1656 1275 2267 6 2280 1 
+1,638 17 117 3 4521 4913 3748 4 0 2478 336 52 1014 31 1878 2309 105 2 5581 3 133 72 313 778 5 0 619 4 4484 2508 168 2 0 88 128 225 52 13 1339 9 1224 1808 1 
+0,4264 5232 10 8242 166 426 1474 2 11282 6 3811 7 68 777 11436 3 3681 3102 1 
+2,24 11 835 1403 215 1358 8509 2 43 422 5 0 1035 940 427 6 502 590 274 2 35 119 1814 12 0 10987 952 1 
+2,74 17 685 1318 1890 1608 6 1607 1639 124 3 48 106 5279 14 0 151 177 102 3 4750 885 2611 35 713 990 1153 31 8012 631 70 1 
+0,2282 1056 91 7 6248 7020 1720 933 4662 931 6 2248 8177 30 49 362 2 277 2867 2901 61 658 72 1 
+1,125 1061 2496 2 563 1374 3407 5 4084 6 29 309 3458 644 2 3401 3481 717 349 12 524 605 605 7 950 1923 18 5 16 578 218 1 
+3,17 1053 7 1592 1609 98 0 63 199 168 1112 523 291 950 1593 58 1368 98 8 28 18 559 207 560 42 2773 874 1 
+3,80 93 17 3041 30 1137 714 419 865 8 28 15 509 96 1019 173 42 1266 140 4741 8 458 50 35 623 3426 39 3 1464 918 312 8 95 1 
+3,1869 17 1053 7 1592 37 1609 98 36 0 63 199 168 1112 523 291 950 1593 58 37 1368 98 36 8 28 18 559 207 560 42 2773 874 1 
+3,412 17 750 96 10729 391 617 1109 28 278 1027 382 723 4 224 9 640 101 261 251 5 3 319 8 26 561 772 706 10 1846 55 1335 0 127 7 67 655 1 
+3,260 1165 529 318 2 454 9 822 552 188 0 430 683 6665 736 3713 3 556 2715 13 3619 0 6505 2236 18 1 
+2,74 17 27 1442 1297 10 25 4769 0 1261 127 7 59 252 5403 926 9128 8 28 102 3 885 2082 205 7 48 176 1318 31 3525 333 1 
+3,24 950 1112 523 1053 7 1592 147 28 10 3 3590 1748 143 989 63 158 383 293 47 44 47 42 345 656 2 1771 641 541 1449 1 
+3,3082 14 4314 4102 7 790 67 62 157 3 48 106 454 304 5 1017 14 107 105 2 425 49 4 0 62 1 
+0,1643 91 117 0 481 1970 0 119 356 923 2 256 160 387 34 40 3 409 190 393 1 
+0,2957 27 561 5223 344 190 1536 2431 20 0 41 45 637 1888 1174 1596 991 40 162 1092 69 3539 190 2 190 3692 2641 42 1596 134 763 0 38 561 3202 12 3 739 47 266 893 284 801 41 45 344 190 300 83 5536 1694 440 519 1 
+0,24 27 254 4 6825 3538 13 3553 2 192 51 3 1835 5 325 10 2208 55 309 1557 0 156 1 
+0,6237 140 235 5 0 1251 1243 114 143 9856 6 2428 30 302 2 708 0 8523 4 295 1250 1 
+0,3036 6 130 8613 381 8101 15 46 857 2 815 137 8631 53 9942 2048 4911 1 
+2,1918 1410 17 3988 7 1066 756 304 26 4629 2 108 0 2263 41 45 839 534 5 3 77 234 1498 8 33 4607 0 194 216 403 4 26 1425 1 
+2,74 17 126 5396 6878 852 18 8 28 912 85 32 141 1298 2488 8 0 398 79 15 1614 539 5 0 59 2258 666 625 1 
+2,74 17 2404 194 2721 3886 7098 11036 3 11012 1562 6 4769 3047 2114 7758 6366 2 673 0 63 346 4 0 59 146 7 3632 12 3 289 115 115 289 374 289 224 28 1 
+1,125 407 13 10752 53 3 5450 5378 213 4 107 105 2 5020 2329 3169 6 3 2133 1519 0 390 7 385 2244 31 2696 15 595 1 
+1,17 1877 5 3474 605 1030 1584 8 28 76 2 2618 50 311 1561 4797 10 32 5420 290 0 1223 6 2578 8 3 3558 515 739 94 6 3403 2039 49 1 
+1,1184 17 538 1436 2692 259 8 1110 1827 692 721 28 595 19 85 3131 1479 796 35 1688 2 495 2 859 35 29 1 
+3,17 955 1180 97 954 236 8 28 147 16 194 699 559 207 15 439 1084 181 51 9 1818 191 6 16 437 1495 1106 2 50 297 66 115 153 1 
+3,571 17 955 1180 97 37 954 236 36 8 28 147 16 194 699 559 207 15 439 1084 181 51 9 1818 191 6 16 437 1495 1106 2 50 66 115 153 1 
+3,80 93 17 641 541 13 1137 714 419 865 8 28 15 509 96 1019 173 42 1266 140 4741 8 458 50 35 623 3426 39 3 1464 918 312 8 95 1 
+3,80 93 17 4329 97 37 7514 98 36 0 199 41 45 8895 4 1102 2 10524 6 9443 8 28 18 559 207 560 15 21 876 65 5 0 1291 6 2428 1258 2112 0 2044 4 293 1102 96 1 
+3,1869 17 1053 7 1592 37 1609 98 36 0 262 66 168 1112 523 291 950 1593 58 8 28 147 21 530 113 708 5 63 158 207 42 345 2773 874 15 191 5578 5 688 1 
+3,24 2885 523 955 1180 97 147 28 10 0 62 11016 2 3 207 5 0 63 158 42 398 1541 191 64 184 5 3245 1 
+0,125 260 344 190 1288 3071 18 19 25 1678 5 21 875 23 466 524 2981 2887 2 781 51 16 3670 466 927 3071 2339 2339 1 
+0,130 7 857 2 108 44 38 371 12 3 265 198 2534 4 178 698 5 9 347 2018 1 
+3,17 1964 1922 2405 58 3072 98 8 28 9785 16 430 191 1252 6 18 984 31 3497 347 5 0 41 45 3038 75 65 15 94 10902 51 8 8940 11042 10582 6 599 5 5205 9 299 328 1 
+3,571 17 955 1180 97 37 954 236 36 8 28 147 16 194 699 559 207 15 439 1084 181 51 9 1818 191 6 16 437 1495 1106 2 50 66 115 153 1 
+0,11 62 291 171 7 295 500 2482 1237 471 3989 844 408 690 13 10694 16 790 878 2 16 910 307 163 6 1243 300 6 900 2 3310 191 3493 0 62 7 1969 267 18 5 21 620 8 95 1 
+0,1199 62 7701 97 105 2 694 0 247 4 3 2670 663 671 8 28 10 19 123 29 3295 809 6 2640 2734 1227 7509 2 2203 8 176 4 114 5568 1 
+1,640 101 261 251 124 3129 23 3776 113 4 817 12 2820 113 4 1445 3593 3220 5 3 319 8 287 2 424 180 1 
+2,24 6585 23 0 59 146 7 316 630 225 324 13 1839 9 75 422 12 2829 325 111 1649 171 6 0 870 92 2476 23 4290 4157 1 
+1,428 59 10609 13 977 12 0 9807 4 3 6006 54 770 39 256 246 20 3 861 5 151 1 
+0,5785 8 28 29 247 21 1334 2 16 4270 114 3212 663 2118 10 0 62 3116 1122 3 10971 4438 2 992 206 4270 525 1 
+0,949 958 408 29 247 0 131 1794 1334 9 16 5821 518 4 2420 805 8 28 1 
+0,497 16 578 463 379 5543 508 5 461 40 65 949 958 408 463 29 694 3 393 4 3299 2 16 463 4363 307 163 1087 2 3470 0 774 197 0 1229 4 0 62 7 4053 6 463 5740 805 1 
+1,17 3070 2168 2 35 887 5 315 7 298 233 30 1272 5 1451 4 0 743 1481 54 2078 989 203 503 0 88 749 18 5 3 218 1518 20 17 28 1 
+1,17 7406 20 576 444 6 743 1481 32 11417 3 2511 10737 5 298 1021 1107 148 42 0 88 749 18 4092 31 1005 1812 1095 2 553 10199 2 8729 0 598 1 
+1,4080 315 17 3070 2168 2 35 887 5 315 7 298 233 30 1272 5 1451 4 0 743 1481 54 2078 989 203 503 0 88 749 18 5 3 218 1518 20 17 28 1 
+1,638 17 117 3 4521 4913 3748 4 0 2478 336 52 1014 31 1878 2309 105 2 5581 3 133 72 466 778 5 0 619 4 4484 2508 168 2 0 88 128 225 52 13 1339 9 1224 1808 1 
+1,2943 3066 4182 2325 0 1747 4957 6 10790 25 1839 0 9663 12 189 8238 5 3 589 110 3176 187 2 228 3327 3477 20 11 1036 378 2325 323 3447 111 332 450 2096 3058 1580 4146 1867 14 0 4104 4370 4380 3890 382 33 187 3 3894 2292 2638 1764 1 
+1,74 1932 27 229 4 8071 13 348 21 1333 1147 2 200 16 718 3728 3375 549 2 0 2131 5834 9281 10286 1328 8 35 10468 4 107 6 16 8175 1 
+3,80 93 17 5815 97 37 5766 98 36 0 262 66 41 45 168 6 8178 2605 8 28 18 623 2094 10891 23 299 328 55 1749 16 545 345 1 
+3,125 11 1371 764 696 265 16 4505 999 904 20 682 3136 1499 23 351 139 113 2 351 226 113 0 696 18 5 3 679 720 8 16 1341 1 
+3,17 1178 58 4000 98 0 220 342 10 1800 5353 16 1689 4 3758 1323 29 504 803 16 2036 14 2291 58 7 2664 789 40 65 0 48 275 18 8 28 1 
+3,80 93 17 1178 58 37 4000 98 36 0 220 342 10 1800 5353 16 1689 4 3758 1323 29 504 803 16 2036 14 2291 58 7 2664 789 40 65 0 48 275 18 8 28 1 
+0,125 4107 344 190 371 64 1030 2 2089 35 5191 11073 14 168 71 3683 82 5 1017 14 120 13 8628 57 4 0 8684 399 5 0 92 1 
+0,11 1588 1835 25 1137 514 883 3195 5 201 5 0 1411 4 3455 2 1253 3 791 9274 2008 1 
+1,24 1877 30 1152 3 2761 1501 5076 2 0 41 45 3962 5 1380 2 1438 287 19 13 3112 134 18 28 39 0 63 322 2719 14 3 41 45 591 5 496 601 3 65 1 
+1,2935 2945 1380 17 428 4818 8643 2761 1560 1501 6 11106 83 406 31 3250 2 0 41 45 1241 5 1380 10311 782 3183 8 77 1539 0 1241 6 369 18 28 1 
+1,4064 17 6722 101 7031 7685 734 2 3443 2648 5 3 1553 237 1221 18 28 189 107 89 10867 0 221 6 4682 198 658 8 1006 6 5338 1 
+1,6214 17 27 7704 228 25 473 2 247 8 8199 3 5225 1021 7998 54 1722 814 8 647 4 11049 0 707 4 0 743 150 20 3548 9126 385 57 4 26 1761 18 28 1 
+3,571 17 955 1180 97 37 954 236 36 8 28 147 16 194 699 559 207 583 1818 191 15 19 265 588 6 2156 1070 6 1710 1993 16 437 1495 2 66 115 153 1 
+3,80 93 17 7491 2071 1288 6904 6672 58 37 6865 236 36 25 473 2 1652 524 4239 6639 408 37 7259 98 36 9 73 1864 153 5 437 6 202 2 4548 16 9295 673 0 275 18 8 28 1 
+1,125 6593 50 11148 13 368 4 0 234 4 361 42 57 4625 25 2029 2 277 562 50 19 20 916 26 621 5248 1 
+1,538 9691 2692 259 83 888 4 692 5002 8 1110 454 9 478 980 1 
+2,24 2877 1393 420 9470 3379 79 613 6 238 7324 7292 613 6753 6660 770 28 39 10998 1812 5000 5 3 861 2503 21 317 18 1 
+3,80 93 17 641 541 13 192 2 520 419 865 8 28 15 509 96 1019 391 35 159 1109 42 1266 140 2736 8 458 50 35 623 3426 39 3 1464 918 312 8 95 1 
+0,11 43 3399 10 0 1849 4 4244 1798 9245 4977 29 2553 4508 0 1543 11332 31 3626 75 65 20 7900 3 1299 4 567 5065 6 2011 193 1 
+0,1388 97 105 2 247 0 258 565 4 16 1881 1234 1520 663 969 601 0 119 48 389 15 368 4 21 3481 527 4 16 1362 3875 3045 1 
+1,17 6589 3573 8 28 8710 1851 2650 288 6 978 4 3 2184 2459 5 2655 605 310 888 4 4440 94 1799 83 0 2713 8288 1 
+1,11 476 5 259 1841 2288 694 437 9300 9 2465 2461 8 0 5369 294 7 4510 1 
+0,3982 6715 4082 5792 5 3014 6426 3982 58 195 172 0 516 4 16 3014 3324 2100 522 11 941 2716 7 3014 208 9145 29 1144 4652 8653 1547 1069 23 4082 58 3 295 1299 4 10844 5062 9 208 345 6 1232 5158 5825 10650 140 10024 2 556 1 
+0,116 377 6181 1894 751 5978 116 7 377 1513 53 255 1298 75 65 12 485 256 1414 2 453 1674 8 0 62 75 95 116 2716 7 882 29 1976 5 21 2783 660 4 751 712 84 29 34 8 0 5344 40 95 1352 0 620 31 1662 304 1 
+2,74 17 685 1318 1890 1608 6 1607 1639 124 3 48 106 5279 14 0 151 177 28 249 46 55 1976 5 1493 14 3 4750 885 53 35 713 990 1153 6 1017 35 5003 1 
+3,4329 7 1728 384 158 207 560 819 113 416 2 21 876 65 5 0 158 1234 175 4999 6 307 8967 1 
+1,11 267 124 0 3129 4 3776 113 4 817 134 18 42 0 669 18 10 0 107 89 8464 1 
+3,80 93 17 41 45 223 1286 419 865 8 28 15 509 96 1019 391 35 159 1109 42 382 723 4 224 9 0 640 267 5 3 319 8 26 561 772 706 50 0 127 7 67 655 1 
+1,327 378 2935 2945 1380 24 1877 30 1152 3 2761 1501 5076 2 0 41 45 3962 5 1380 2 1438 287 19 7 3112 134 18 28 165 7 0 63 322 2719 14 3 41 45 591 5 496 601 3 65 1 
+2,24 11 111 536 4207 1141 43 346 2040 1494 5949 4379 8 28 2540 0 1553 157 7 5051 9421 1 
+3,2924 860 8 0 465 7077 6495 2 154 10507 16 8763 15 2057 2710 3 2193 5 652 449 1 
+3,750 96 442 821 2002 2 2812 1868 3 1072 39 640 261 251 124 3 424 319 8109 2802 704 896 1 
+3,5832 3495 2661 32 2142 5565 345 42 40 62 25 5416 0 922 4 72 1 
+0,1967 1040 11 62 40 65 29 3714 49 1507 6 1834 1087 2 1715 371 4 16 173 157 805 2659 16 4053 518 84 1337 953 7 2260 1074 4062 1794 1 
+0,24 603 1975 1801 18 28 46 32 2055 35 2170 4 3 338 197 130 97 6 1402 1408 58 2 200 579 2673 114 925 1 
+3,17 41 45 1266 8 28 5203 10 299 328 55 34 0 182 2015 41 45 731 248 1580 5 2810 12 312 134 2078 9051 652 2094 14 297 818 153 1 
+0,165 1004 166 116 7 183 202 394 29 5214 141 1769 230 658 0 62 3706 2 10702 787 5 0 2060 1058 241 10 16 2781 8116 2 0 377 137 55 1134 5 21 1490 4 512 3607 2159 37 572 279 929 748 841 843 856 886 493 66 648 824 36 37 326 36 1855 91 37 1668 36 37 1731 36 1 
+0,844 239 5270 118 84 1507 30 780 1267 12 16 755 1935 66 527 9 178 609 1 
+1,533 272 17 272 7 215 1159 101 261 251 8 28 1436 224 5 3 659 424 319 8 26 561 42 26 3484 3730 2 2003 456 345 46 1802 15 3 4919 995 1 
+1,1911 1410 1066 756 492 3 186 44 3780 2 8969 53 3 77 234 1498 33 564 181 26 241 3151 4 0 106 8 0 43 4 77 876 2070 2 355 834 1611 6 1356 1595 5 0 839 534 14 1647 1634 1 
+1,1060 1045 831 2421 908 1149 786 2430 6 212 1367 196 1810 290 312 28 2 154 1133 1302 23 0 4552 4 299 328 0 1564 731 2 246 0 150 5 3 1093 103 1 
+1,80 93 641 541 10693 293 28 15 1986 4969 3627 82 10744 67 96 6 2463 400 879 21 2782 191 218 23 1964 1922 2405 6 3 3160 1289 23 1053 7 11 1042 539 3400 1981 31 51 2262 5623 71 226 374 113 14 374 5741 2257 39 4792 226 73 113 293 75 65 1 
+1,27 6163 1936 11374 8290 54 2016 3 420 3613 5 171 20 5235 3092 26 10972 4 11128 317 4669 25 90 4633 9 1224 622 1136 6 2115 2 277 3 1081 2182 9071 476 172 1 
+3,571 17 955 1180 97 37 954 236 36 8 28 147 16 194 699 559 207 583 1818 191 15 19 265 588 6 2156 1070 6 1710 1993 16 437 1495 2 66 115 153 1 
+3,2409 17 2893 10337 7759 97 37 7714 4334 36 18 8 28 19 29 479 922 1299 6079 23 6346 7751 755 0 258 5 3 10925 4 2428 1087 2 4888 60 16 671 5142 1 
+3,1053 7 147 3 632 374 113 708 5 487 1468 188 16 63 158 42 345 713 641 541 1449 1 
+0,27 38 114 1330 23 4382 58 68 1538 4904 434 1817 404 5 367 8619 788 15 3508 5322 187 2 3 679 23 0 62 1 
+3,1869 17 1053 7 1592 37 1609 98 36 0 262 66 168 1112 523 291 950 1593 58 8 28 147 21 530 113 708 5 63 158 207 42 345 2773 874 15 191 5578 5 688 1 
+0,1 37 347 36 7450 37 347 36 2888 76 857 92 8907 42 
+0,2175 1937 463 3498 12 2256 2576 342 1575 1322 1531 1346 2110 697 2821 91 2 474 8 1197 7 4323 1 
+0,4256 55 184 516 21 890 53 0 48 5485 2427 4 2869 3 481 1021 62 1 
+0,11 250 1635 177 30 2235 5 151 6 0 932 13 168 2 2129 4 385 10 2095 1856 0 446 278 49 2081 1 
+3,17 955 1180 97 954 236 8 28 147 16 194 699 559 207 583 1818 191 15 19 265 588 6 2156 1070 6 1710 1993 16 437 1495 2 297 66 115 153 1 
+3,412 17 750 96 772 23 38 159 1109 8 28 15 224 9 640 101 261 251 5 3 319 8 26 561 772 706 10 1846 55 1335 0 127 7 67 655 1 
+3,80 93 17 41 45 223 2065 8 28 15 67 96 2516 39 723 4 224 9 0 640 267 5 3 319 8 26 561 772 706 50 0 127 7 67 655 1 
+0,6389 1834 55 1826 0 9043 4 483 1559 9 682 103 6 235 8 4983 203 68 504 9488 187 2 21 1659 20 4230 254 242 4230 254 490 247 1 
+0,4035 1858 5 461 3002 6 4799 5 413 324 6 499 29 477 49 2608 49 3348 6 1003 3419 5 0 3767 8439 187 2 3 38 9908 596 20 48 242 14 0 212 1884 9 3831 840 4165 4165 490 247 1 
+0,462 0 63 72 5 77 103 3 9465 2025 5613 25 1732 82 0 764 1933 1887 165 7 515 2565 8696 6 135 2637 361 6 295 724 2 763 10 3 3349 1079 140 34 348 422 5 1780 980 5 0 1487 1193 2376 1933 846 254 490 247 1 
+0,949 958 29 4729 3299 2 16 5562 395 0 463 4363 3753 2420 307 163 6 16 5822 40 65 188 16 463 379 250 508 5 461 1 
+2,1376 1600 1073 0 63 2657 670 2 34 8 0 753 535 79 2613 1202 1357 15 0 48 1075 7 2125 28 9 3 41 45 79 10 29 495 2 2158 0 3231 23 499 119 162 1 
+2,6009 5122 7955 6771 13 21 1448 1563 5 205 2143 316 670 1280 1637 6 1326 452 285 6 151 8687 1 
+2,41 45 5396 4349 2301 1071 9 3 885 8 189 1442 507 10 29 1438 287 0 5268 311 364 92 336 8574 5 0 151 252 1 
+1,461 7787 7337 1321 30 141 10912 2 1982 4500 7617 9348 3594 1747 10580 322 9898 1782 30 8484 44 789 11434 51 5 0 11026 3506 2630 8 9606 9 1434 1058 249 5 752 225 0 183 2051 16 5174 2 2162 10 46 2792 5 7732 7 9222 1 
+1,327 378 1930 1172 1039 5014 2 626 2 0 9224 8 633 700 1267 2453 12 3 2163 773 14 1744 0 1556 3214 4934 60 4 5201 123 2903 244 4248 6571 1 
+1,17 11 107 13 302 2 5370 8736 6 9079 4784 8 5177 395 5 3 645 2 982 2077 568 9460 0 10302 4 67 1250 21 67 240 317 18 8 28 1 
+1,327 378 2409 1172 27 420 10160 522 3350 20 0 512 107 85 4995 1134 5 10707 3734 9745 7216 2384 6227 7021 18 28 1 
+1,1052 1167 2415 847 555 14 2676 14 21 382 1970 7602 564 46 928 10388 181 0 398 581 1775 1844 1 
+1,4044 17 1061 7 388 28 901 198 3 218 5 0 258 2534 4 1402 712 4724 120 19 358 3 5439 4 812 304 5 2360 5 9697 11172 1217 8320 375 1 
+1,2337 6819 2930 6706 0 126 543 4 320 3 513 9678 579 1551 2113 6 798 1504 5 375 11089 5 228 28 10 52 55 2690 41 45 6 935 476 64 9257 4486 4 26 4410 1 
+1,273 101 152 7 318 2 1228 2225 4 350 4 41 45 196 168 23 499 6 496 55 1733 180 237 47 3115 23 388 3322 42 124 69 1715 0 5419 8 810 184 4497 1031 2556 5 100 6 375 1 
+1,744 100 3934 14 100 7 212 1889 358 28 9 1300 540 480 1385 334 432 2 3087 26 2238 83 41 45 6 269 196 6 1795 26 1454 60 4 3 484 628 5 253 1 
+1,74 205 219 271 184 10769 39 324 7 1335 386 2 227 1873 5 0 570 445 625 1305 2 0 2132 28 6 10356 1246 9 0 59 460 364 4560 271 1797 2 859 957 1953 7 159 4 321 112 665 14 0 3759 4160 252 31 4704 12 0 406 1995 144 5 0 625 33 110 1 
+1,4128 1376 1600 1073 0 63 2657 670 2 34 8 0 753 535 79 2613 1202 1357 15 0 48 1075 7 2125 28 9 3 41 45 79 10 29 495 2 2158 0 3231 23 499 119 162 1 
+1,1060 1045 831 2421 908 1149 786 2430 6 212 1367 196 1810 290 312 28 2 8848 60 1851 4 299 328 6 1704 599 6 135 1325 2 350 4 94 215 2595 1 
+1,80 93 27 703 5 67 96 6 2782 10058 23 1964 1922 6 1053 7 2145 38 1986 9456 8 641 541 28 1526 223 1530 293 11 1042 4606 49 61 5631 1499 5 1009 1332 1 
+3,11 7092 6466 1068 928 550 3 450 1129 6 0 4939 5 0 119 8727 4948 69 3 3196 230 2 8091 1068 794 11191 385 211 225 2 453 1 
+3,412 6400 17 6225 18 19 1018 49 61 3785 104 20 803 16 169 139 113 1319 5 6685 1617 408 690 8 28 3 10487 4 16 8805 3108 2 1226 3 483 8419 8995 1 
+0,24 3823 18 28 46 32 280 3 3173 225 46 763 149 0 3850 4444 232 4 26 4749 427 3 514 4598 12 1162 1825 295 2 21 3710 2132 4 599 1 
+1,17 4058 501 244 2929 2270 1030 21 269 228 28 2 1146 51 5215 83 5490 294 3027 2920 6 26 337 1655 1 
+3,434 190 175 13 9100 532 3450 1917 6141 9746 0 240 7 9268 
+3,1068 68 11147 5983 7 38 163 2 521 2196 121 166 2332 19 1160 0 370 1 
+3,11 6242 2665 20 16 8022 140 858 2821 45 4529 518 211 29 46 982 1232 1 
+0,125 227 633 2566 518 5107 7742 3730 2 222 274 12 105 2 265 186 161 588 583 3 228 9555 3344 0 150 145 1243 2 5330 759 12 0 3711 1 
+1,24 171 1018 16 317 769 11129 23 1962 3015 2 14 214 1865 8 28 15 5560 12 3323 1554 1712 164 791 5040 3603 9 531 94 184 1279 1 
+1,744 17 4058 501 244 2929 2270 1030 21 269 228 28 2 1146 51 5215 83 5490 294 3027 2920 6 26 337 1655 1 
+1,409 2173 285 5956 6071 55 626 168 23 4114 2 1081 26 3411 1151 9 515 26 9766 1 
+1,27 1778 1804 5439 4 812 10376 706 4 3 617 318 2 475 0 41 45 542 8311 6 5179 55 34 3 257 670 
+3,80 93 17 41 45 223 2065 8 28 776 3 858 23 895 67 96 39 306 0 640 267 920 3 424 772 706 50 0 127 7 67 655 1 
+0,17 3083 308 6 700 629 5 0 1446 482 4 2653 284 627 50 169 113 40 47 2 297 268 115 153 3 308 303 18 8 28 1 
+0,24 6168 50 2064 6 902 8162 29 76 1826 105 2 1225 3 2515 4833 2 2396 635 14 1644 1633 254 4 2342 6 2396 635 134 18 1 
+0,607 1043 17 3083 308 6 700 629 5 0 1446 482 4 2653 284 627 50 169 113 40 47 2 268 115 153 3 308 303 18 8 28 1 
+1,17 27 343 8 3 3148 1552 14 214 133 3157 53 0 2491 4 3 3440 1747 5 1190 7 359 8 28 42 0 2456 1685 141 4569 476 18 1 
+1,24 6271 4 38 921 758 23 101 152 2 504 4429 28 1 
+3,5708 6368 550 9642 8479 75 65 3885 19 167 122 51 82 0 9827 1 
+0,1855 91 1373 5 0 62 10 7 734 2 979 82 15 4696 15 3 4557 55 548 1328 1259 3 4883 4 5561 155 0 882 4 116 58 767 2 8457 35 9036 868 1 
+0,125 746 1741 30 1472 51 12 0 526 6 30 1796 21 2600 193 8808 92 14 1306 0 1078 4 295 56 206 604 142 303 7018 18 1 
+0,603 1975 1801 18 28 46 32 2055 35 2170 4 3 338 197 130 97 6 1402 1408 58 2 200 579 2673 114 925 1 
+1,4079 1190 17 27 343 8 3 3148 1552 14 214 133 3157 53 0 2491 4 3 3440 1747 5 1190 7 359 28 42 0 2456 1685 141 4569 476 18 1 
+3,80 93 17 1630 4 344 168 2098 10788 8 28 8 1746 4 1751 264 39 0 4728 299 328 11425 5 312 44 0 736 21 2438 18 1 
+0,252 6 142 1507 30 8 3 174 1001 4 114 10 771 76 235 566 12 0 178 959 527 1 
+0,3083 308 6 700 629 5 0 1446 482 4 2653 284 627 50 169 113 40 47 2 268 115 153 3 308 303 18 8 28 1 
+1,125 27 8910 8420 355 3 7338 5 3 282 304 5 3 1487 7342 179 44 0 736 2 1024 84 9907 4 5504 55 356 4943 861 4640 1 
+1,125 11 576 1977 2552 669 2 0 5538 4 21 633 1644 2148 591 1778 970 5 0 11339 681 233 4 298 53 3 585 8292 10132 584 1 
+3,17 41 45 223 2065 8 28 776 3 858 23 895 67 96 39 306 0 640 267 920 3 424 772 706 50 0 127 7 67 655 1 
+3,17 4055 690 0 92 7 63 199 1022 1288 18 8 28 19 473 2 34 2463 20 3 2494 4 133 798 3303 1250 9 297 139 153 360 3122 4 297 66 153 4 3242 1 
+3,80 93 17 4055 690 0 92 7 63 199 1022 1288 18 8 28 19 473 2 34 2463 20 3 2494 4 133 798 3303 1250 9 139 153 360 3122 4 66 153 4 3242 1 
+3,11 367 5197 175 1122 3 864 2 396 1292 2126 8 317 5405 1 
+3,11 62 1311 21 1112 5 16 5500 42 29 40 143 34 1757 2 0 8356 1 
+0,17 11 3639 10027 9314 4 3 1695 1882 2498 698 76 23 0 2498 787 42 23 4489 10 2631 1265 19 41 45 242 18 8 86 1 
+0,1 37 347 36 463 379 37 347 36 1906 5820 5537 2 1824 8776 
+0,11 238 105 2 79 51 16 3408 5209 193 12 0 5042 7 5569 663 114 1 
+1,24 11 88 749 284 548 11348 2 8799 718 3397 2158 0 9970 19 555 39 0 2398 351 644 6 478 8567 12 0 92 7 73 153 7104 3 176 41 98 317 18 28 1 
+1,3022 2953 745 2298 4 219 1916 7 1100 3533 83 3 1689 5061 2563 796 228 28 10873 60 3726 10847 552 132 0 2678 285 31 2 4464 2 279 82 83 0 3545 54 25 3552 180 9 103 8 2479 2647 647 1 
+1,6804 7721 538 3823 18 28 46 32 280 3 3173 225 46 763 149 0 3850 4444 232 4 26 4749 3 514 4598 12 1162 1825 295 2 21 3710 2132 4 599 2892 21 3308 1555 4 0 3173 20 11 1036 378 8124 10277 5572 8423 46 18 1549 0 1542 4 0 4873 111 7750 10258 15 566 15 3 3659 46 763 31 404 9 8440 2060 11341 1 
+1,836 100 41 45 1546 1135 53 0 1191 635 4 253 926 3 484 540 628 225 1006 64 9410 28 15 5163 14 3 420 508 1157 2 804 3 2507 292 2 495 2 5126 21 157 2 0 443 1 
+1,273 101 152 8 28 172 105 2 1139 1344 161 2 825 161 41 45 196 54 30 210 2745 5 499 6 496 5 57 4 0 199 11183 5257 248 0 157 4 0 3900 1407 1 
+0,1855 91 1640 167 21 2073 4 230 178 550 2 34 322 3 5092 732 4 19 15 3 1262 10 31 4555 8 21 2083 5 0 2102 4 3 9675 1235 5 0 8690 1 
+0,178 527 1092 69 521 566 12 99 969 278 7425 7 1351 2282 7451 8 142 235 184 2 34 1711 1 
+2,117 255 1138 2 0 2688 5 0 1050 937 1313 761 507 0 228 9787 1807 10 1415 385 50 0 7990 7 1529 361 53 228 1 
+3,80 93 17 41 45 223 4606 8 28 776 3 858 23 895 67 96 39 272 7 267 1436 52 89 920 3 424 319 3283 706 50 0 127 7 67 655 1 
+3,571 17 955 2326 97 37 27 376 305 373 341 397 91 410 415 402 954 236 286 223 423 418 36 954 236 37 27 36 8 28 147 16 194 699 559 207 6 2460 16 437 1495 2 66 115 153 15 19 265 629 8 1070 6 1710 1526 16 202 51 740 113 1 
+0,6418 6470 2009 3020 3005 29 34 367 2 1326 50 11 433 1623 7 730 8739 218 1 
+0,7325 4308 58 4 4180 3853 829 8 28 172 0 4481 4 7384 6452 1196 21 3408 5209 114 671 9 178 3442 10 8861 71 9912 2056 4 130 97 178 6 0 156 3970 1372 434 2467 10 30 3348 3684 4 2635 220 1741 6 931 11432 1 
+3,603 1975 1801 2553 35 2170 4 130 1402 1408 338 1351 1178 105 2 425 2036 14 2291 2068 3071 7 466 927 9387 21 875 23 524 2981 2887 2068 6 49 1 
+3,1373 30 8944 1150 223 12 3087 429 9 7236 4317 19 13 3 8383 1 
+0,2884 0 398 685 10165 54 1436 0 43 2233 9 59 438 32 0 9227 2 4982 3 2237 76 118 225 232 3222 55 8534 42 436 3 2237 5141 20 232 322 9354 2801 1 
+0,470 379 6268 525 29 34 762 5 5674 1323 290 0 41 45 1 
+3,1869 17 1053 7 1592 37 27 376 305 373 341 397 91 410 415 402 1609 98 286 223 423 418 36 1609 98 37 27 36 0 92 7 262 66 168 1112 523 8 28 147 21 530 113 708 5 63 158 207 6 18 383 9 0 1020 4 0 47 85 176 545 874 1 
+3,80 93 17 1630 4 176 5497 342 7014 58 37 27 376 305 373 341 397 91 410 415 402 4124 98 286 223 423 418 36 4124 98 37 27 36 6 524 6628 58 37 27 376 305 373 341 397 91 410 415 402 4019 98 286 223 423 418 36 4019 98 37 27 36 442 2 35 3437 2628 14 214 740 389 8 28 39 3 8917 20 6888 6010 84 4599 8833 240 980 1 
+0,7845 268 4 0 363 2141 5218 2118 1263 3 2200 1973 8846 6 3 7347 1654 8 1 
+0,2175 1937 3861 2510 2672 1736 352 104 1576 9646 82 513 4860 1317 463 3498 12 2256 2576 342 1 
+0,4256 55 184 516 21 890 53 0 48 5485 2427 4 2869 40 162 1 
+0,2296 4026 57 286 4 3 1428 1653 2755 1972 994 4988 14 1985 4224 1489 3143 3879 1 
+0,117 255 1138 0 3185 123 3 1794 9 710 1023 3540 2410 124 69 979 60 40 47 1 
+0,7700 58 296 3 38 565 4 16 968 700 1330 8 28 358 7346 169 6 5155 105 9 21 1334 2 16 2023 114 5787 1949 3963 84 29 222 53 4507 14 0 157 4 40 162 6 13 416 9 247 119 47 1 
+0,542 5160 3261 0 309 4 7151 3 3698 5219 1776 193 762 8 232 5308 9 309 5 0 382 4732 4 2249 2197 6 135 7972 392 3094 5 710 1243 897 6 175 2691 1 
+2,24 847 5575 2473 396 51 5198 6 3641 1204 10699 2 2152 0 450 2 34 292 435 19 31 993 72 9 7851 6376 2 615 26 1724 6 2062 5584 180 53 0 4458 20 11347 26 3511 6 4586 3891 3891 1 
+2,24 7764 4 0 1313 761 507 83 1052 1167 6870 285 1050 937 1 
+2,461 6709 693 468 11 461 5970 28 473 2 1550 8 3 57 47 768 12 8781 6013 5951 429 1016 79 1791 994 1550 4 0 338 64 76 4751 1 
+1,1182 2441 1584 2 3649 0 2091 4 263 311 4797 84 32 5420 290 0 1223 1 
+1,1391 2611 0 2651 4 9973 6107 7730 200 21 3117 1 
+1,101 152 195 172 105 2 1139 1344 161 2 825 161 41 45 196 54 30 210 2745 5 499 6 496 1 
+1,6051 101 152 8 28 172 105 2 1139 15 232 15 825 161 41 45 196 54 30 210 2745 5 1645 499 6 496 5 57 4 0 199 5257 248 0 157 4 0 3900 1407 1 
+1,3022 2953 745 3976 4 219 1916 3304 5 3184 28 15 0 2678 285 2538 23 3 546 3244 1555 2471 6 737 53 228 9 3 1814 12 0 3545 54 25 3552 180 9 103 8 2479 2647 647 1 
+0,470 379 4382 7 38 114 582 2 3388 4904 1817 1 
+0,1057 11 4186 1619 7119 5725 13 21 9034 3521 4942 1708 10 5010 807 3692 2109 6 2079 2641 1229 12 99 4 0 2056 6 9260 10 57 5136 85 1099 2 548 5 3 2030 6512 6508 2001 190 1 
+0,1057 1632 1619 6394 13 348 16 344 301 399 2 0 119 661 2766 505 981 6 264 300 15 0 1076 3593 60 0 1632 1637 2109 175 6 3097 0 557 3860 440 519 968 2 16 5193 1 
+0,1057 577 3127 871 4 130 7 1055 1187 3300 959 142 2670 2 0 178 609 307 163 192 2 504 0 240 184 11327 2 1156 16 10377 950 371 10 32 35 10267 192 2 5260 307 163 3716 15 46 30 181 762 20 130 140 34 5448 2 1117 10 138 4 0 114 46 502 145 8 35 525 55 34 4744 20 959 71 140 145 930 4739 1 
+1,24 2831 8146 1980 3 3218 4 4357 11187 4713 1325 2 41 45 444 5 100 6 209 48 4357 2529 3382 35 62 18 28 1 
+0,24 2298 4 242 318 2 8664 14 0 254 4 7932 2 1326 50 642 7170 931 599 1500 6 135 1267 2706 2 8528 8239 9865 4781 1 
+0,129 3922 4021 273 24 11 43 1708 2 5274 1421 8519 23 0 3158 4 94 3677 3668 3 38 3702 3485 10 55 1809 2091 6 10688 0 2815 153 578 1988 2 2231 3668 25 90 1664 20 0 107 117 3781 113 4 3668 3 1421 5554 5 0 4535 3143 3193 20 3 1421 8518 1264 0 864 4 10676 8860 71 769 1 
+0,2175 1937 130 10697 3715 7926 6402 10338 5 819 104 3861 2510 2672 1736 352 104 1 
+0,2296 4026 57 286 4 3 1428 1653 2755 1972 994 4988 14 1985 4224 1489 3143 3879 1 
+0,840 117 2366 7 3860 440 519 6 301 399 29 34 4567 8 1632 7 897 1 
+3,80 93 17 41 45 223 5259 8 28 15 67 96 5295 23 35 1109 310 2782 383 8083 23 2709 2572 0 2075 3615 1 
+3,80 93 17 11 702 2065 83 0 1718 8 28 39 3 218 8 3338 53 41 45 2442 1140 614 4 3 627 5 881 3409 2 2112 0 545 3089 774 9 0 162 1 
+0,24 462 459 47 266 5840 4293 619 9 3 478 361 8135 5 189 902 9130 1852 77 276 3 65 8 3 3148 10556 10 4568 3 220 12 3 173 1146 567 156 4643 1 
+0,129 5758 6506 2943 3066 24 11 11326 1663 11025 5 0 8695 8853 1806 9671 12 67 10100 10871 3357 6 4964 4 8470 6 10003 3855 10413 0 106 7 730 9853 5024 52 10425 0 10523 7 8840 2 3323 617 1127 6 11136 60 3 9926 50 16 9950 8465 3866 7208 1 
+0,129 6496 6501 1185 1163 24 2863 58 0 156 206 1484 175 13 3 3562 5 1370 3 62 10 4786 4802 130 97 1187 193 6 5271 8 19 4830 2 3700 801 6 986 0 2107 19 5276 2 300 1 

fluid/PaddleRec/tagspace/utils.py

 import re
 import sys
 import collections
+import os
 import six
 import time
 import numpy as np
@@ -23,30 +24,39 @@ def to_lodtensor(data, place):
     res.set_lod([lod])
     return res
 
-def prepare_data(train_filename,
-                 test_filename,
+def get_vocab_size(vocab_path):
+    with open(vocab_path, "r") as rf:
+        line = rf.readline()
+        return int(line.strip())
+
+
+def prepare_data(file_dir,
+                 vocab_text_path,
+                 vocab_tag_path,
                  batch_size,
-                 neg_size=1,
-                 buffer_size=1000,
-                 word_freq_threshold=0,
-                 enable_ce=False):
+                 neg_size,
+                 buffer_size,
+                 is_train=True):
     """ prepare the AG's News Topic Classification data """
-    print("start constuct word dict")
-    vocab_text = build_dict(2, word_freq_threshold, train_filename, test_filename)
-    vocab_tag = build_dict(0, word_freq_threshold, train_filename, test_filename)
-    print("construct word dict done\n")
-    train_reader = sort_batch(
-        paddle.reader.shuffle(
-            train(
-                train_filename, vocab_text, vocab_tag, neg_size,
-                buffer_size, data_type=DataType.SEQ),
-            buf_size=buffer_size),
-        batch_size, batch_size * 20)
-    test_reader = sort_batch(
-        test(
-            test_filename, vocab_text, vocab_tag, neg_size, buffer_size, data_type=DataType.SEQ),
-        batch_size, batch_size * 20)
-    return vocab_text, vocab_tag, train_reader, test_reader
+    print("start read file")
+    if is_train:
+        vocab_text_size = get_vocab_size(vocab_text_path)
+        vocab_tag_size = get_vocab_size(vocab_tag_path)
+        reader = sort_batch(
+            paddle.reader.shuffle(
+                train(
+                    file_dir, vocab_tag_size, neg_size,
+                    buffer_size, data_type=DataType.SEQ),
+                buf_size=buffer_size),
+            batch_size, batch_size * 20)
+    else:
+        vocab_tag_size = get_vocab_size(vocab_tag_path)
+        vocab_text_size = 0
+        reader = sort_batch(
+            test(
+                file_dir, vocab_tag_size, buffer_size, data_type=DataType.SEQ),
+            batch_size, batch_size * 20)
+    return vocab_text_size, vocab_tag_size, reader
 
 def sort_batch(reader, batch_size, sort_group_size, drop_last=False):
     """
@@ -97,82 +107,57 @@ def sort_batch(reader, batch_size, sort_group_size, drop_last=False):
 class DataType(object):
     SEQ = 2
 
-def word_count(column_num, input_file, word_freq=None):
-    """
-    compute word count from corpus
-    """
-    if word_freq is None:
-        word_freq = collections.defaultdict(int)
-    data_file = csv.reader(input_file)
-    for row in data_file:
-        for w in re.split(r'\W+',row[column_num].strip()):
-            word_freq[w]+= 1
-    return word_freq
-
-def build_dict(column_num=2, min_word_freq=50, train_filename="", test_filename=""):
-    """
-    Build a word dictionary from the corpus,  Keys of the dictionary are words,
-    and values are zero-based IDs of these words.
-    """
-    with open(train_filename) as trainf:
-        with open(test_filename) as testf:
-            word_freq = word_count(column_num, testf, word_count(column_num, trainf))
-
-    word_freq = [x for x in six.iteritems(word_freq) if x[1] > min_word_freq]
-    word_freq_sorted = sorted(word_freq, key=lambda x: (-x[1], x[0]))
-    words, _ = list(zip(*word_freq_sorted))
-    word_idx = dict(list(zip(words, six.moves.range(len(words)))))
-    return word_idx
-
-def train_reader_creator(filename, text_idx, tag_idx, neg_size, n, data_type):
+def train_reader_creator(file_dir, tag_size, neg_size, n, data_type):
     def reader():
-        with open(filename) as input_file:
-            data_file = csv.reader(input_file)
-            for row in data_file:
-                text_raw = re.split(r'\W+', row[2].strip())
-                text = [text_idx.get(w) for w in text_raw]
-                tag_raw = re.split(r'\W+', row[0].strip())
-                pos_index = tag_idx.get(tag_raw[0])
-                pos_tag=[]
-                pos_tag.append(pos_index)
-                neg_tag=[]
-                max_iter = 100
-                now_iter = 0
-                sum_n = 0
-                while(sum_n < neg_size) :
-                    now_iter += 1
-                    if now_iter > max_iter:
-                        print("error : only one class")
-                        sys.exit(0)
-                    rand_i = np.random.randint(0, len(tag_idx))
-                    if rand_i != pos_index:
-                        neg_index=rand_i
-                        neg_tag.append(neg_index)
-                        sum_n += 1
-                if n > 0 and len(text) > n: continue
-                yield text, pos_tag, neg_tag
+        files = os.listdir(file_dir)
+        for fi in files:
+            with open(file_dir + '/' + fi, "r") as f:
+                for l in f:
+                    l = l.strip().split(",")
+                    pos_index = int(l[0])
+                    pos_tag = []
+                    pos_tag.append(pos_index)
+                    text_raw = l[1].split()
+                    text = [int(w) for w in text_raw]
+                    neg_tag = []
+                    max_iter = 100
+                    now_iter = 0
+                    sum_n = 0
+                    while(sum_n < neg_size) :
+                        now_iter += 1
+                        if now_iter > max_iter:
+                            print("error : only one class")
+                            sys.exit(0)
+                        rand_i = np.random.randint(0, tag_size)
+                        if rand_i != pos_index:
+                            neg_index = rand_i
+                            neg_tag.append(neg_index)
+                            sum_n += 1
+                    if n > 0 and len(text) > n: continue
+                    yield text, pos_tag, neg_tag
     return reader
 
-def test_reader_creator(filename, text_idx, tag_idx, n, data_type):
+def test_reader_creator(file_dir, tag_size, n, data_type):
     def reader():
-        with open(filename) as input_file:
-            data_file = csv.reader(input_file)
-            for row in data_file:
-                text_raw = re.split(r'\W+', row[2].strip())
-                text = [text_idx.get(w) for w in text_raw]
-                tag_raw = re.split(r'\W+', row[0].strip())
-                pos_index = tag_idx.get(tag_raw[0])
-                pos_tag = []
-                pos_tag.append(pos_index)
-                for ii in range(len(tag_idx)):
-                    tag = []
-                    tag.append(ii)
-                    yield text, tag, pos_tag
+        files = os.listdir(file_dir)
+        for fi in files:
+            with open(file_dir + '/' + fi, "r") as f:
+                for l in f:
+                    l = l.strip().split(",")
+                    pos_index = int(l[0])
+                    pos_tag = []
+                    pos_tag.append(pos_index)
+                    text_raw = l[1].split()
+                    text = [int(w) for w in text_raw]
+                    for ii in range(tag_size):
+                        tag = []
+                        tag.append(ii)
+                        yield text, tag, pos_tag
     return reader
 
 
-def train(filename, text_idx, tag_idx, neg_size, n, data_type=DataType.SEQ):
-    return train_reader_creator(filename, text_idx, tag_idx, neg_size, n, data_type)
+def train(train_dir, tag_size, neg_size, n, data_type=DataType.SEQ):
+    return train_reader_creator(train_dir, tag_size, neg_size, n, data_type)
 
-def test(filename, text_idx, tag_idx, neg_size, n, data_type=DataType.SEQ):
-    return test_reader_creator(filename, text_idx, tag_idx, n, data_type)
+def test(test_dir, tag_size, n, data_type=DataType.SEQ):
+    return test_reader_creator(test_dir, tag_size, n, data_type)

fluid/PaddleRec/tagspace/vocab_tag.txt

+4

fluid/PaddleRec/tagspace/vocab_text.txt

+11447

fluid/PaddleRec/word2vec/README.cn.md

+
+# 基于skip-gram的word2vector模型
+
+## 介绍
+
+
+## 运行环境
+需要先安装PaddlePaddle Fluid
+
+## 数据集
+数据集使用的是来自1 Billion Word Language Model Benchmark的(http://www.statmt.org/lm-benchmark)的数据集.
+
+下载数据集：
+```bash
+cd data && ./download.sh && cd ..
+```
+
+## 模型
+本例子实现了一个skip-gram模式的word2vector模型。
+
+
+## 数据准备
+对数据进行预处理以生成一个词典。
+
+```bash
+python preprocess.py --data_path ./data/1-billion-word-language-modeling-benchmark-r13output/training-monolingual.tokenized.shuffled --dict_path data/1-billion_dict
+```
+
+## 训练
+训练的命令行选项可以通过`python train.py -h`列出。
+
+### 单机训练：
+
+```bash
+python train.py \
+        --train_data_path ./data/1-billion-word-language-modeling-benchmark-r13output/training-monolingual.tokenized.shuffled \
+        --dict_path data/1-billion_dict \
+        2>&1 | tee train.log
+```
+
+### 分布式训练
+
+本地启动一个2 trainer 2 pserver的分布式训练任务，分布式场景下训练数据会按照trainer的id进行切分，保证trainer之间的训练数据不会重叠，提高训练效率
+
+```bash
+sh cluster_train.sh
+```
+
+## 预测
+在infer.py中我们在`build_test_case`方法中构造了一些test case来评估word embeding的效果：
+我们输入test case（ 我们目前采用的是analogical-reasoning的任务：找到A - B = C - D的结构，为此我们计算A - B + D，通过cosine距离找最近的C，计算准确率要去除候选中出现A、B、D的候选 ）然后计算候选和整个embeding中所有词的余弦相似度，并且取topK（K由参数 --rank_num确定，默认为4）打印出来。
+
+如：
+对于：boy - girl + aunt = uncle  
+0 nearest aunt:0.89
+1 nearest uncle:0.70
+2 nearest grandmother:0.67
+3 nearest father:0.64
+
+您也可以在`build_test_case`方法中模仿给出的例子增加自己的测试
+
+训练中预测：
+
+```bash
+python infer.py --infer_during_train 2>&1 | tee infer.log
+```
+使用某个model进行离线预测：
+
+```bash
+python infer.py --infer_once --model_output_dir ./models/[具体的models文件目录] 2>&1 | tee infer.log
+```
+## 在百度云上运行集群训练
+1. 参考文档 [在百度云上启动Fluid分布式训练](https://github.com/PaddlePaddle/FluidDoc/blob/develop/doc/fluid/user_guides/howto/training/train_on_baidu_cloud_cn.rst) 在百度云上部署一个CPU集群。
+1. 用preprocess.py处理训练数据生成train.txt。
+1. 将train.txt切分成集群机器份，放到每台机器上。
+1. 用上面的 `分布式训练` 中的命令行启动分布式训练任务.

fluid/PaddleRec/word2vec/README.md

+
+# Skip-Gram Word2Vec Model
+
+## Introduction
+
+
+## Environment
+You should install PaddlePaddle Fluid first.
+
+## Dataset
+The training data for the 1 Billion Word Language Model Benchmark的(http://www.statmt.org/lm-benchmark).
+
+Download dataset:
+```bash
+cd data && ./download.sh && cd ..
+```
+
+## Model
+This model implement a skip-gram model of word2vector.
+
+
+## Data Preprocessing method
+
+Preprocess the training data to generate a word dict.
+
+```bash
+python preprocess.py --data_path ./data/1-billion-word-language-modeling-benchmark-r13output/training-monolingual.tokenized.shuffled --dict_path data/1-billion_dict
+```
+
+## Train
+The command line options for training can be listed by `python train.py -h`.
+
+### Local Train:
+```bash
+python train.py \
+        --train_data_path ./data/1-billion-word-language-modeling-benchmark-r13output/training-monolingual.tokenized.shuffled \
+        --dict_path data/1-billion_dict \
+        2>&1 | tee train.log
+```
+
+
+### Distributed Train
+Run a 2 pserver 2 trainer distribute training on a single machine.
+In distributed training setting, training data is splited by trainer_id, so that training data
+ do not overlap among trainers
+
+```bash
+sh cluster_train.sh
+```
+
+## Infer
+
+In infer.py we construct some test cases in the `build_test_case` method to evaluate the effect of word embeding:
+We enter the test case (we are currently using the analogical-reasoning task: find the structure of A - B = C - D, for which we calculate A - B + D, find the nearest C by cosine distance, the calculation accuracy is removed Candidates for A, B, and D appear in the candidate) Then calculate the cosine similarity of the candidate and all words in the entire embeding, and print out the topK (K is determined by the parameter --rank_num, the default is 4).
+
+Such as:
+For: boy - girl + aunt = uncle
+0 nearest aunt: 0.89
+1 nearest uncle: 0.70
+2 nearest grandmother: 0.67
+3 nearest father:0.64
+
+You can also add your own tests by mimicking the examples given in the `build_test_case` method.
+
+Forecast in training:
+
+```bash
+Python infer.py --infer_during_train 2>&1 | tee infer.log
+```
+Use a model for offline prediction:
+
+```bash
+Python infer.py --infer_once --model_output_dir ./models/[specific models file directory] 2>&1 | tee infer.log
+```
+
+## Train on Baidu Cloud
+1. Please prepare some CPU machines on Baidu Cloud following the steps in [train_on_baidu_cloud](https://github.com/PaddlePaddle/FluidDoc/blob/develop/doc/fluid/user_guides/howto/training/train_on_baidu_cloud_cn.rst)
+1. Prepare dataset using preprocess.py.
+1. Split the train.txt to trainer_num parts and put them on the machines.
+1. Run training with the cluster train using the command in `Distributed Train` above.

fluid/PaddleRec/word2vec/cluster_train.sh

+#!/bin/bash
+
+echo "WARNING: This script only for run PaddlePaddle Fluid on one node..."
+echo ""
+
+export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib
+export PADDLE_PSERVER_PORTS=36001,36002
+export PADDLE_PSERVER_PORT_ARRAY=(36001 36002)
+export PADDLE_PSERVERS=2
+
+export PADDLE_IP=127.0.0.1
+export PADDLE_TRAINERS=2
+
+export CPU_NUM=2
+export NUM_THREADS=2
+export PADDLE_SYNC_MODE=TRUE
+export PADDLE_IS_LOCAL=0
+
+export FLAGS_rpc_deadline=3000000
+export GLOG_logtostderr=1
+
+
+export TRAIN_DATA=data/enwik8
+export DICT_PATH=data/enwik8_dict
+export IS_SPARSE="--is_sparse"
+
+
+echo "Start PSERVER ..."
+for((i=0;i<$PADDLE_PSERVERS;i++))
+do
+    cur_port=${PADDLE_PSERVER_PORT_ARRAY[$i]}
+    echo "PADDLE WILL START PSERVER "$cur_port
+    GLOG_v=0 PADDLE_TRAINING_ROLE=PSERVER CUR_PORT=$cur_port PADDLE_TRAINER_ID=$i python -u train.py $IS_SPARSE &> pserver.$i.log &
+done
+
+echo "Start TRAINER ..."
+for((i=0;i<$PADDLE_TRAINERS;i++))
+do
+    echo "PADDLE WILL START Trainer "$i
+    GLOG_v=0 PADDLE_TRAINER_ID=$i PADDLE_TRAINING_ROLE=TRAINER python -u train.py $IS_SPARSE --train_data_path $TRAIN_DATA --dict_path $DICT_PATH &> trainer.$i.log &
+done
\ No newline at end of file

fluid/PaddleRec/word2vec/data/download.sh

+#!/bin/bash
+
+wget http://www.statmt.org/lm-benchmark/1-billion-word-language-modeling-benchmark-r13output.tar.gz
+tar -zxvf 1-billion-word-language-modeling-benchmark-r13output.tar.gz
+

fluid/PaddleRec/word2vec/infer.py

+import paddle
+import time
+import os
+import paddle.fluid as fluid
+import numpy as np
+from Queue import PriorityQueue
+import logging
+import argparse
+from sklearn.metrics.pairwise import cosine_similarity
+
+word_to_id = dict()
+id_to_word = dict()
+
+logging.basicConfig(format='%(asctime)s - %(levelname)s - %(message)s')
+logger = logging.getLogger("fluid")
+logger.setLevel(logging.INFO)
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(
+        description="PaddlePaddle Word2vec infer example")
+    parser.add_argument(
+        '--dict_path',
+        type=str,
+        default='./data/1-billion_dict',
+        help="The path of training dataset")
+    parser.add_argument(
+        '--model_output_dir',
+        type=str,
+        default='models',
+        help="The path for model to store (with infer_once please set specify dir to models) (default: models)"
+    )
+    parser.add_argument(
+        '--rank_num',
+        type=int,
+        default=4,
+        help="find rank_num-nearest result for test (default: 4)")
+    parser.add_argument(
+        '--infer_once',
+        action='store_true',
+        required=False,
+        default=False,
+        help='if using infer_once, (default: False)')
+    parser.add_argument(
+        '--infer_during_train',
+        action='store_true',
+        required=False,
+        default=True,
+        help='if using infer_during_train, (default: True)')
+
+    return parser.parse_args()
+
+
+def BuildWord_IdMap(dict_path):
+    with open(dict_path + "_word_to_id_", 'r') as f:
+        for line in f:
+            word_to_id[line.split(' ')[0]] = int(line.split(' ')[1])
+            id_to_word[int(line.split(' ')[1])] = line.split(' ')[0]
+
+
+def inference_prog():
+    fluid.layers.create_parameter(
+        shape=[1, 1], dtype='float32', name="embeding")
+
+
+def build_test_case(emb):
+    emb1 = emb[word_to_id['boy']] - emb[word_to_id['girl']] + emb[word_to_id[
+        'aunt']]
+    desc1 = "boy - girl + aunt = uncle"
+    emb2 = emb[word_to_id['brother']] - emb[word_to_id['sister']] + emb[
+        word_to_id['sisters']]
+    desc2 = "brother - sister + sisters = brothers"
+    emb3 = emb[word_to_id['king']] - emb[word_to_id['queen']] + emb[word_to_id[
+        'woman']]
+    desc3 = "king - queen + woman = man"
+    emb4 = emb[word_to_id['reluctant']] - emb[word_to_id['reluctantly']] + emb[
+        word_to_id['slowly']]
+    desc4 = "reluctant - reluctantly + slowly = slow"
+    emb5 = emb[word_to_id['old']] - emb[word_to_id['older']] + emb[word_to_id[
+        'deeper']]
+    desc5 = "old - older + deeper = deep"
+    return [[emb1, desc1], [emb2, desc2], [emb3, desc3], [emb4, desc4],
+            [emb5, desc5]]
+
+
+def inference_test(scope, model_dir, args):
+    BuildWord_IdMap(args.dict_path)
+    logger.info("model_dir is: {}".format(model_dir + "/"))
+    emb = np.array(scope.find_var("embeding").get_tensor())
+    test_cases = build_test_case(emb)
+    logger.info("inference result: ====================")
+    for case in test_cases:
+        pq = topK(args.rank_num, emb, case[0])
+        logger.info("Test result for {}".format(case[1]))
+        pq_tmps = list()
+        for i in range(args.rank_num):
+            pq_tmps.append(pq.get())
+        for i in range(len(pq_tmps)):
+            logger.info("{} nearest is {}, rate is {}".format(i, id_to_word[
+                pq_tmps[len(pq_tmps) - 1 - i].id], pq_tmps[len(pq_tmps) - 1 - i]
+                                                              .priority))
+        del pq_tmps[:]
+
+
+class PQ_Entry(object):
+    def __init__(self, cos_similarity, id):
+        self.priority = cos_similarity
+        self.id = id
+
+    def __cmp__(self, other):
+        return cmp(self.priority, other.priority)
+
+
+def topK(k, emb, test_emb):
+    pq = PriorityQueue(k + 1)
+    if len(emb) <= k:
+        for i in range(len(emb)):
+            x = cosine_similarity([emb[i]], [test_emb])
+            pq.put(PQ_Entry(x, i))
+        return pq
+
+    for i in range(len(emb)):
+        x = cosine_similarity([emb[i]], [test_emb])
+        pq_e = PQ_Entry(x, i)
+        if pq.full():
+            pq.get()
+        pq.put(pq_e)
+    pq.get()
+    return pq
+
+
+def infer_during_train(args):
+    model_file_list = list()
+    exe = fluid.Executor(fluid.CPUPlace())
+    Scope = fluid.Scope()
+    inference_prog()
+    solved_new = True
+    while True:
+        time.sleep(60)
+        current_list = os.listdir(args.model_output_dir)
+        # logger.info("current_list is : {}".format(current_list))
+        # logger.info("model_file_list is : {}".format(model_file_list))
+        if set(model_file_list) == set(current_list):
+            if solved_new:
+                solved_new = False
+                logger.info("No New models created")
+            pass
+        else:
+            solved_new = True
+            increment_models = list()
+            for f in current_list:
+                if f not in model_file_list:
+                    increment_models.append(f)
+            logger.info("increment_models is : {}".format(increment_models))
+            for model in increment_models:
+                model_dir = args.model_output_dir + "/" + model
+                if os.path.exists(model_dir + "/_success"):
+                    logger.info("using models from " + model_dir)
+                    with fluid.scope_guard(Scope):
+                        fluid.io.load_persistables(
+                            executor=exe, dirname=model_dir + "/")
+                        inference_test(Scope, model_dir, args)
+            model_file_list = current_list
+
+
+def infer_once(args):
+    # check models file has already been finished
+    if os.path.exists(args.model_output_dir + "/_success"):
+        logger.info("using models from " + args.model_output_dir)
+        exe = fluid.Executor(fluid.CPUPlace())
+        Scope = fluid.Scope()
+        inference_prog()
+        with fluid.scope_guard(Scope):
+            fluid.io.load_persistables(
+                executor=exe, dirname=args.model_output_dir + "/")
+            inference_test(Scope, args.model_output_dir, args)
+
+
+if __name__ == '__main__':
+    args = parse_args()
+    # while setting infer_once please specify the dir to models file with --model_output_dir
+    if args.infer_once:
+        infer_once(args)
+    if args.infer_during_train:
+        infer_during_train(args)

fluid/PaddleRec/word2vec/network_conf.py

+# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""
+neural network for word2vec
+"""
+
+from __future__ import print_function
+
+import math
+import numpy as np
+
+import paddle.fluid as fluid
+
+
+def skip_gram_word2vec(dict_size,
+                       word_frequencys,
+                       embedding_size,
+                       max_code_length=None,
+                       with_hsigmoid=False,
+                       with_nce=True,
+                       is_sparse=False):
+    def nce_layer(input, label, embedding_size, num_total_classes,
+                  num_neg_samples, sampler, word_frequencys, sample_weight):
+
+        w_param_name = "nce_w"
+        b_param_name = "nce_b"
+        w_param = fluid.default_main_program().global_block().create_parameter(
+            shape=[num_total_classes, embedding_size],
+            dtype='float32',
+            name=w_param_name)
+        b_param = fluid.default_main_program().global_block().create_parameter(
+            shape=[num_total_classes, 1], dtype='float32', name=b_param_name)
+
+        cost = fluid.layers.nce(input=input,
+                                label=label,
+                                num_total_classes=num_total_classes,
+                                sampler=sampler,
+                                custom_dist=word_frequencys,
+                                sample_weight=sample_weight,
+                                param_attr=fluid.ParamAttr(name=w_param_name),
+                                bias_attr=fluid.ParamAttr(name=b_param_name),
+                                num_neg_samples=num_neg_samples,
+                                is_sparse=is_sparse)
+
+        return cost
+
+    def hsigmoid_layer(input, label, path_table, path_code, non_leaf_num,
+                       is_sparse):
+        if non_leaf_num is None:
+            non_leaf_num = dict_size
+
+        cost = fluid.layers.hsigmoid(
+            input=input,
+            label=label,
+            num_classes=non_leaf_num,
+            path_table=path_table,
+            path_code=path_code,
+            is_custom=True,
+            is_sparse=is_sparse)
+
+        return cost
+
+    datas = []
+
+    input_word = fluid.layers.data(name="input_word", shape=[1], dtype='int64')
+    predict_word = fluid.layers.data(
+        name='predict_word', shape=[1], dtype='int64')
+    datas.append(input_word)
+    datas.append(predict_word)
+
+    if with_hsigmoid:
+        path_table = fluid.layers.data(
+            name='path_table',
+            shape=[max_code_length if max_code_length else 40],
+            dtype='int64')
+        path_code = fluid.layers.data(
+            name='path_code',
+            shape=[max_code_length if max_code_length else 40],
+            dtype='int64')
+        datas.append(path_table)
+        datas.append(path_code)
+
+    py_reader = fluid.layers.create_py_reader_by_data(
+        capacity=64, feed_list=datas, name='py_reader', use_double_buffer=True)
+
+    words = fluid.layers.read_file(py_reader)
+
+    emb = fluid.layers.embedding(
+        input=words[0],
+        is_sparse=is_sparse,
+        size=[dict_size, embedding_size],
+        param_attr=fluid.ParamAttr(
+            name='embeding',
+            initializer=fluid.initializer.Normal(scale=1 /
+                                                 math.sqrt(dict_size))))
+
+    cost, cost_nce, cost_hs = None, None, None
+
+    if with_nce:
+        cost_nce = nce_layer(emb, words[1], embedding_size, dict_size, 5,
+                             "uniform", word_frequencys, None)
+        cost = cost_nce
+    if with_hsigmoid:
+        cost_hs = hsigmoid_layer(emb, words[1], words[2], words[3], dict_size,
+                                 is_sparse)
+        cost = cost_hs
+    if with_nce and with_hsigmoid:
+        cost = fluid.layers.elementwise_add(cost_nce, cost_hs)
+
+    avg_cost = fluid.layers.reduce_mean(cost)
+
+    return avg_cost, py_reader

fluid/PaddleRec/word2vec/preprocess.py

+# -*- coding: utf-8 -*
+
+import re
+import argparse
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(
+        description="Paddle Fluid word2 vector preprocess")
+    parser.add_argument(
+        '--data_path',
+        type=str,
+        required=True,
+        help="The path of training dataset")
+    parser.add_argument(
+        '--dict_path',
+        type=str,
+        default='./dict',
+        help="The path of generated dict")
+    parser.add_argument(
+        '--freq',
+        type=int,
+        default=5,
+        help="If the word count is less then freq, it will be removed from dict")
+    parser.add_argument(
+        '--is_local',
+        action='store_true',
+        required=False,
+        default=False,
+        help='Local train or not, (default: False)')
+
+    return parser.parse_args()
+
+
+def text_strip(text):
+    return re.sub("[^a-z ]", "", text)
+
+
+def build_Huffman(word_count, max_code_length):
+
+    MAX_CODE_LENGTH = max_code_length
+    sorted_by_freq = sorted(word_count.items(), key=lambda x: x[1])
+    count = list()
+    vocab_size = len(word_count)
+    parent = [-1] * 2 * vocab_size
+    code = [-1] * MAX_CODE_LENGTH
+    point = [-1] * MAX_CODE_LENGTH
+    binary = [-1] * 2 * vocab_size
+    word_code_len = dict()
+    word_code = dict()
+    word_point = dict()
+    i = 0
+    for a in range(vocab_size):
+        count.append(word_count[sorted_by_freq[a][0]])
+
+    for a in range(vocab_size):
+        word_point[sorted_by_freq[a][0]] = [-1] * MAX_CODE_LENGTH
+        word_code[sorted_by_freq[a][0]] = [-1] * MAX_CODE_LENGTH
+
+    for k in range(vocab_size):
+        count.append(1e15)
+
+    pos1 = vocab_size - 1
+    pos2 = vocab_size
+    min1i = 0
+    min2i = 0
+    b = 0
+
+    for r in range(vocab_size):
+        if pos1 >= 0:
+            if count[pos1] < count[pos2]:
+                min1i = pos1
+                pos1 = pos1 - 1
+            else:
+                min1i = pos2
+                pos2 = pos2 + 1
+        else:
+            min1i = pos2
+            pos2 = pos2 + 1
+        if pos1 >= 0:
+            if count[pos1] < count[pos2]:
+                min2i = pos1
+                pos1 = pos1 - 1
+            else:
+                min2i = pos2
+                pos2 = pos2 + 1
+        else:
+            min2i = pos2
+            pos2 = pos2 + 1
+
+        count[vocab_size + r] = count[min1i] + count[min2i]
+
+        #record the parent of left and right child
+        parent[min1i] = vocab_size + r
+        parent[min2i] = vocab_size + r
+        binary[min1i] = 0  #left branch has code 0
+        binary[min2i] = 1  #right branch has code 1
+
+    for a in range(vocab_size):
+        b = a
+        i = 0
+        while True:
+            code[i] = binary[b]
+            point[i] = b
+            i = i + 1
+            b = parent[b]
+            if b == vocab_size * 2 - 2:
+                break
+
+        word_code_len[sorted_by_freq[a][0]] = i
+        word_point[sorted_by_freq[a][0]][0] = vocab_size - 2
+
+        for k in range(i):
+            word_code[sorted_by_freq[a][0]][i - k - 1] = code[k]
+
+            # only non-leaf nodes will be count in
+            if point[k] - vocab_size >= 0:
+                word_point[sorted_by_freq[a][0]][i - k] = point[k] - vocab_size
+
+    return word_point, word_code, word_code_len
+
+
+def preprocess(data_path, dict_path, freq, is_local):
+    """
+    proprocess the data, generate dictionary and save into dict_path.
+    :param data_path: the input data path.
+    :param dict_path: the generated dict path. the data in dict is "word count"
+    :param freq:
+    :return:
+    """
+    # word to count
+    word_count = dict()
+
+    if is_local:
+        for i in range(1, 100):
+            with open(data_path + "/news.en-000{:0>2d}-of-00100".format(
+                    i)) as f:
+                for line in f:
+                    line = line.lower()
+                    line = text_strip(line)
+                    words = line.split()
+                    for item in words:
+                        if item in word_count:
+                            word_count[item] = word_count[item] + 1
+                        else:
+                            word_count[item] = 1
+    item_to_remove = []
+    for item in word_count:
+        if word_count[item] <= freq:
+            item_to_remove.append(item)
+    for item in item_to_remove:
+        del word_count[item]
+
+    path_table, path_code, word_code_len = build_Huffman(word_count, 40)
+
+    with open(dict_path, 'w+') as f:
+        for k, v in word_count.items():
+            f.write(str(k) + " " + str(v) + '\n')
+
+    with open(dict_path + "_ptable", 'w+') as f2:
+        for pk, pv in path_table.items():
+            f2.write(str(pk) + ":" + ' '.join((str(x) for x in pv)) + '\n')
+
+    with open(dict_path + "_pcode", 'w+') as f3:
+        for pck, pcv in path_table.items():
+            f3.write(str(pck) + ":" + ' '.join((str(x) for x in pcv)) + '\n')
+
+
+if __name__ == "__main__":
+    args = parse_args()
+    preprocess(args.data_path, args.dict_path, args.freq, args.is_local)

fluid/PaddleRec/word2vec/reader.py

+# -*- coding: utf-8 -*
+
+import numpy as np
+import preprocess
+
+import logging
+
+logging.basicConfig(format='%(asctime)s - %(levelname)s - %(message)s')
+logger = logging.getLogger("fluid")
+logger.setLevel(logging.INFO)
+
+
+class Word2VecReader(object):
+    def __init__(self,
+                 dict_path,
+                 data_path,
+                 filelist,
+                 trainer_id,
+                 trainer_num,
+                 window_size=5):
+        self.window_size_ = window_size
+        self.data_path_ = data_path
+        self.filelist = filelist
+        self.num_non_leaf = 0
+        self.word_to_id_ = dict()
+        self.id_to_word = dict()
+        self.word_to_path = dict()
+        self.word_to_code = dict()
+        self.trainer_id = trainer_id
+        self.trainer_num = trainer_num
+
+        word_all_count = 0
+        word_counts = []
+        word_id = 0
+
+        with open(dict_path, 'r') as f:
+            for line in f:
+                word, count = line.split()[0], int(line.split()[1])
+                self.word_to_id_[word] = word_id
+                self.id_to_word[word_id] = word  #build id to word dict
+                word_id += 1
+                word_counts.append(count)
+                word_all_count += count
+
+        with open(dict_path + "_word_to_id_", 'w+') as f6:
+            for k, v in self.word_to_id_.items():
+                f6.write(str(k) + " " + str(v) + '\n')
+
+        self.dict_size = len(self.word_to_id_)
+        self.word_frequencys = [
+            float(count) / word_all_count for count in word_counts
+        ]
+        print("dict_size = " + str(
+            self.dict_size)) + " word_all_count = " + str(word_all_count)
+
+        with open(dict_path + "_ptable", 'r') as f2:
+            for line in f2:
+                self.word_to_path[line.split(":")[0]] = np.fromstring(
+                    line.split(':')[1], dtype=int, sep=' ')
+                self.num_non_leaf = np.fromstring(
+                    line.split(':')[1], dtype=int, sep=' ')[0]
+        print("word_ptable dict_size = " + str(len(self.word_to_path)))
+
+        with open(dict_path + "_pcode", 'r') as f3:
+            for line in f3:
+                self.word_to_code[line.split(":")[0]] = np.fromstring(
+                    line.split(':')[1], dtype=int, sep=' ')
+        print("word_pcode dict_size = " + str(len(self.word_to_code)))
+
+    def get_context_words(self, words, idx, window_size):
+        """
+        Get the context word list of target word.
+
+        words: the words of the current line
+        idx: input word index
+        window_size: window size
+        """
+        target_window = np.random.randint(1, window_size + 1)
+        # need to keep in mind that maybe there are no enough words before the target word.
+        start_point = idx - target_window if (idx - target_window) > 0 else 0
+        end_point = idx + target_window
+        # context words of the target word
+        targets = set(words[start_point:idx] + words[idx + 1:end_point + 1])
+        return list(targets)
+
+    def train(self, with_hs):
+        def _reader():
+            for file in self.filelist:
+                with open(self.data_path_ + "/" + file, 'r') as f:
+                    logger.info("running data in {}".format(self.data_path_ +
+                                                            "/" + file))
+                    count = 1
+                    for line in f:
+                        if self.trainer_id == count % self.trainer_num:
+                            line = preprocess.text_strip(line)
+                            word_ids = [
+                                self.word_to_id_[word] for word in line.split()
+                                if word in self.word_to_id_
+                            ]
+                            for idx, target_id in enumerate(word_ids):
+                                context_word_ids = self.get_context_words(
+                                    word_ids, idx, self.window_size_)
+                                for context_id in context_word_ids:
+                                    yield [target_id], [context_id]
+                        else:
+                            pass
+                        count += 1
+
+        def _reader_hs():
+            for file in self.filelist:
+                with open(self.data_path_ + "/" + file, 'r') as f:
+                    logger.info("running data in {}".format(self.data_path_ +
+                                                            "/" + file))
+                    count = 1
+                    for line in f:
+                        if self.trainer_id == count % self.trainer_num:
+                            line = preprocess.text_strip(line)
+                            word_ids = [
+                                self.word_to_id_[word] for word in line.split()
+                                if word in self.word_to_id_
+                            ]
+                            for idx, target_id in enumerate(word_ids):
+                                context_word_ids = self.get_context_words(
+                                    word_ids, idx, self.window_size_)
+                                for context_id in context_word_ids:
+                                    yield [target_id], [context_id], [
+                                        self.word_to_code[self.id_to_word[
+                                            context_id]]
+                                    ], [
+                                        self.word_to_path[self.id_to_word[
+                                            context_id]]
+                                    ]
+                        else:
+                            pass
+                        count += 1
+
+        if not with_hs:
+            return _reader
+        else:
+            return _reader_hs
+
+
+if __name__ == "__main__":
+    window_size = 10
+
+    reader = Word2VecReader("data/enwik9_dict", "data/enwik9", window_size)
+    i = 0
+    for x, y in reader.train()():
+        print("x: " + str(x))
+        print("y: " + str(y))
+        print("\n")
+        if i == 10:
+            exit(0)
+        i += 1

fluid/PaddleRec/word2vec/train.py

+from __future__ import print_function
+
+import argparse
+import logging
+import os
+import time
+
+import numpy as np
+
+# disable gpu training for this example
+os.environ["CUDA_VISIBLE_DEVICES"] = ""
+
+import paddle
+import paddle.fluid as fluid
+from paddle.fluid.executor import global_scope
+
+import reader
+from network_conf import skip_gram_word2vec
+from infer import inference_test
+
+logging.basicConfig(format='%(asctime)s - %(levelname)s - %(message)s')
+logger = logging.getLogger("fluid")
+logger.setLevel(logging.INFO)
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(
+        description="PaddlePaddle Word2vec example")
+    parser.add_argument(
+        '--train_data_path',
+        type=str,
+        default='./data/1-billion-word-language-modeling-benchmark-r13output/training-monolingual.tokenized.shuffled',
+        help="The path of training dataset")
+    parser.add_argument(
+        '--dict_path',
+        type=str,
+        default='./data/1-billion_dict',
+        help="The path of data dict")
+    parser.add_argument(
+        '--test_data_path',
+        type=str,
+        default='./data/text8',
+        help="The path of testing dataset")
+    parser.add_argument(
+        '--batch_size',
+        type=int,
+        default=100,
+        help="The size of mini-batch (default:100)")
+    parser.add_argument(
+        '--num_passes',
+        type=int,
+        default=10,
+        help="The number of passes to train (default: 10)")
+    parser.add_argument(
+        '--model_output_dir',
+        type=str,
+        default='models',
+        help='The path for model to store (default: models)')
+    parser.add_argument(
+        '--embedding_size',
+        type=int,
+        default=64,
+        help='sparse feature hashing space for index processing')
+
+    parser.add_argument(
+        '--with_hs',
+        action='store_true',
+        required=False,
+        default=False,
+        help='using hierarchical sigmoid, (default: False)')
+
+    parser.add_argument(
+        '--with_nce',
+        action='store_true',
+        required=False,
+        default=False,
+        help='using negtive sampling, (default: True)')
+
+    parser.add_argument(
+        '--max_code_length',
+        type=int,
+        default=40,
+        help='max code length used by hierarchical sigmoid, (default: 40)')
+
+    parser.add_argument(
+        '--is_sparse',
+        action='store_true',
+        required=False,
+        default=False,
+        help='embedding and nce will use sparse or not, (default: False)')
+
+    parser.add_argument(
+        '--with_Adam',
+        action='store_true',
+        required=False,
+        default=False,
+        help='Using Adam as optimizer or not, (default: False)')
+
+    parser.add_argument(
+        '--is_local',
+        action='store_true',
+        required=False,
+        default=False,
+        help='Local train or not, (default: False)')
+
+    parser.add_argument(
+        '--with_speed',
+        action='store_true',
+        required=False,
+        default=False,
+        help='print speed or not , (default: False)')
+
+    parser.add_argument(
+        '--with_infer_test',
+        action='store_true',
+        required=False,
+        default=False,
+        help='Do inference every 100 batches , (default: False)')
+
+    parser.add_argument(
+        '--rank_num',
+        type=int,
+        default=4,
+        help="find rank_num-nearest result for test (default: 4)")
+
+    return parser.parse_args()
+
+
+def train_loop(args, train_program, reader, py_reader, loss, trainer_id):
+    train_reader = paddle.batch(
+        paddle.reader.shuffle(
+            reader.train((args.with_hs or (not args.with_nce))),
+            buf_size=args.batch_size * 100),
+        batch_size=args.batch_size)
+
+    py_reader.decorate_paddle_reader(train_reader)
+
+    place = fluid.CPUPlace()
+
+    exe = fluid.Executor(place)
+    exe.run(fluid.default_startup_program())
+
+    exec_strategy = fluid.ExecutionStrategy()
+
+    print("CPU_NUM:" + str(os.getenv("CPU_NUM")))
+    exec_strategy.num_threads = int(os.getenv("CPU_NUM"))
+
+    build_strategy = fluid.BuildStrategy()
+    if int(os.getenv("CPU_NUM")) > 1:
+        build_strategy.reduce_strategy = fluid.BuildStrategy.ReduceStrategy.Reduce
+
+    train_exe = fluid.ParallelExecutor(
+        use_cuda=False,
+        loss_name=loss.name,
+        main_program=train_program,
+        build_strategy=build_strategy,
+        exec_strategy=exec_strategy)
+
+    profile_state = "CPU"
+    profiler_step = 0
+    profiler_step_start = 20
+    profiler_step_end = 30
+
+    for pass_id in range(args.num_passes):
+        epoch_start = time.time()
+        py_reader.start()
+        batch_id = 0
+        start = time.clock()
+
+        try:
+            while True:
+
+                if profiler_step == profiler_step_start:
+                    fluid.profiler.start_profiler(profile_state)
+
+                loss_val = train_exe.run(fetch_list=[loss.name])
+                loss_val = np.mean(loss_val)
+
+                if profiler_step == profiler_step_end:
+                    fluid.profiler.stop_profiler('total', 'trainer_profile.log')
+                    profiler_step += 1
+                else:
+                    profiler_step += 1
+
+                if batch_id % 50 == 0:
+                    logger.info(
+                        "TRAIN --> pass: {} batch: {} loss: {} reader queue:{}".
+                        format(pass_id, batch_id,
+                               loss_val.mean() / args.batch_size,
+                               py_reader.queue.size()))
+                if args.with_speed:
+                    if batch_id % 1000 == 0 and batch_id != 0:
+                        elapsed = (time.clock() - start)
+                        start = time.clock()
+                        samples = 1001 * args.batch_size * int(
+                            os.getenv("CPU_NUM"))
+                        logger.info("Time used: {}, Samples/Sec: {}".format(
+                            elapsed, samples / elapsed))
+                # calculate infer result each 100 batches when using --with_infer_test
+                if args.with_infer_test:
+                    if batch_id % 1000 == 0 and batch_id != 0:
+                        model_dir = args.model_output_dir + '/batch-' + str(
+                            batch_id)
+                        inference_test(global_scope(), model_dir, args)
+
+                if batch_id % 500000 == 0 and batch_id != 0:
+                    model_dir = args.model_output_dir + '/batch-' + str(
+                        batch_id)
+                    fluid.io.save_persistables(executor=exe, dirname=model_dir)
+                    with open(model_dir + "/_success", 'w+') as f:
+                        f.write(str(batch_id))
+                batch_id += 1
+
+        except fluid.core.EOFException:
+            py_reader.reset()
+            epoch_end = time.time()
+            logger.info("Epoch: {0}, Train total expend: {1} ".format(
+                pass_id, epoch_end - epoch_start))
+
+            model_dir = args.model_output_dir + '/pass-' + str(pass_id)
+            if trainer_id == 0:
+                fluid.io.save_persistables(executor=exe, dirname=model_dir)
+                with open(model_dir + "/_success", 'w+') as f:
+                    f.write(str(pass_id))
+
+
+def GetFileList(data_path):
+    return os.listdir(data_path)
+
+
+def train(args):
+
+    if not os.path.isdir(args.model_output_dir):
+        os.mkdir(args.model_output_dir)
+
+    filelist = GetFileList(args.train_data_path)
+    word2vec_reader = None
+    if args.is_local or os.getenv("PADDLE_IS_LOCAL", "1") == "1":
+        word2vec_reader = reader.Word2VecReader(
+            args.dict_path, args.train_data_path, filelist, 0, 1)
+    else:
+        trainer_id = int(os.environ["PADDLE_TRAINER_ID"])
+        trainers = int(os.environ["PADDLE_TRAINERS"])
+        word2vec_reader = reader.Word2VecReader(args.dict_path,
+                                                args.train_data_path, filelist,
+                                                trainer_id, trainer_num)
+
+    logger.info("dict_size: {}".format(word2vec_reader.dict_size))
+    loss, py_reader = skip_gram_word2vec(
+        word2vec_reader.dict_size,
+        word2vec_reader.word_frequencys,
+        args.embedding_size,
+        args.max_code_length,
+        args.with_hs,
+        args.with_nce,
+        is_sparse=args.is_sparse)
+
+    optimizer = None
+    if args.with_Adam:
+        optimizer = fluid.optimizer.Adam(learning_rate=1e-3)
+    else:
+        optimizer = fluid.optimizer.SGD(learning_rate=1e-3)
+
+    optimizer.minimize(loss)
+
+    # do local training 
+    if args.is_local or os.getenv("PADDLE_IS_LOCAL", "1") == "1":
+        logger.info("run local training")
+        main_program = fluid.default_main_program()
+
+        with open("local.main.proto", "w") as f:
+            f.write(str(main_program))
+
+        train_loop(args, main_program, word2vec_reader, py_reader, loss, 0)
+    # do distribute training
+    else:
+        logger.info("run dist training")
+
+        trainer_id = int(os.environ["PADDLE_TRAINER_ID"])
+        trainers = int(os.environ["PADDLE_TRAINERS"])
+        training_role = os.environ["PADDLE_TRAINING_ROLE"]
+
+        port = os.getenv("PADDLE_PSERVER_PORT", "6174")
+        pserver_ips = os.getenv("PADDLE_PSERVER_IPS", "")
+        eplist = []
+        for ip in pserver_ips.split(","):
+            eplist.append(':'.join([ip, port]))
+        pserver_endpoints = ",".join(eplist)
+        current_endpoint = os.getenv("PADDLE_CURRENT_IP", "") + ":" + port
+
+        config = fluid.DistributeTranspilerConfig()
+        config.slice_var_up = False
+        t = fluid.DistributeTranspiler(config=config)
+        t.transpile(
+            trainer_id,
+            pservers=pserver_endpoints,
+            trainers=trainers,
+            sync_mode=True)
+
+        if training_role == "PSERVER":
+            logger.info("run pserver")
+            prog = t.get_pserver_program(current_endpoint)
+            startup = t.get_startup_program(
+                current_endpoint, pserver_program=prog)
+
+            with open("pserver.main.proto.{}".format(os.getenv("CUR_PORT")),
+                      "w") as f:
+                f.write(str(prog))
+
+            exe = fluid.Executor(fluid.CPUPlace())
+            exe.run(startup)
+            exe.run(prog)
+        elif training_role == "TRAINER":
+            logger.info("run trainer")
+            train_prog = t.get_trainer_program()
+
+            with open("trainer.main.proto.{}".format(trainer_id), "w") as f:
+                f.write(str(train_prog))
+
+            train_loop(args, train_prog, word2vec_reader, py_reader, loss,
+                       trainer_id)
+
+
+def env_declar():
+    print("********  Rename Cluster Env to PaddleFluid Env ********")
+
+    print("Content-Type: text/plain\n\n")
+    for key in os.environ.keys():
+        print("%30s %s \n" % (key, os.environ[key]))
+
+    if os.environ["TRAINING_ROLE"] == "PSERVER" or os.environ[
+            "PADDLE_IS_LOCAL"] == "0":
+        os.environ["PADDLE_TRAINING_ROLE"] = os.environ["TRAINING_ROLE"]
+        os.environ["PADDLE_PSERVER_PORT"] = os.environ["PADDLE_PORT"]
+        os.environ["PADDLE_PSERVER_IPS"] = os.environ["PADDLE_PSERVERS"]
+        os.environ["PADDLE_TRAINERS"] = os.environ["PADDLE_TRAINERS_NUM"]
+        os.environ["PADDLE_CURRENT_IP"] = os.environ["POD_IP"]
+        os.environ["PADDLE_TRAINER_ID"] = os.environ["PADDLE_TRAINER_ID"]
+        # we set the thread number same as CPU number
+        os.environ["CPU_NUM"] = "12"
+
+    print("Content-Type: text/plain\n\n")
+    for key in os.environ.keys():
+        print("%30s %s \n" % (key, os.environ[key]))
+
+    print("******  Rename Cluster Env to PaddleFluid Env END ******")
+
+
+if __name__ == '__main__':
+    args = parse_args()
+    if args.is_local:
+        pass
+    else:
+        env_declar()
+    train(args)

fluid/README.cn.rst

-Fluid 模型库
+`Fluid 模型库 <https://github.com/PaddlePaddle/models/tree/develop/fluid>`__
 ============
 
 图像分类