add ce for image classification (#4030)

PaddleCV/image_classification/reader.py

@@ -287,17 +287,25 @@ class ImageNetReader:
                     full_lines = [line.strip() for line in flist]
                     if mode != "test" and len(full_lines) < settings.batch_size:
                         print(
-                            "Warning: The number of the whole data ({}) is smaller than the batch_size ({}), and drop_last is turnning on, so nothing  will feed in program, Terminated now. Please reset batch_size to a smaller number or feed more data!"
-                            .format(len(full_lines), settings.batch_size))
+                            "Warning: The number of the whole data ({}) is smaller than the batch_size ({}), and drop_last is turnning on, so nothing  will feed in program, Terminated now. Please reset batch_size to a smaller number or feed more data!".
+                            format(len(full_lines), settings.batch_size))
                         os._exit(1)
                     if num_trainers > 1 and mode == "train":
                         assert self.shuffle_seed is not None, "multiprocess train, shuffle seed must be set!"
                         np.random.RandomState(self.shuffle_seed).shuffle(
                             full_lines)
                     elif shuffle:
+                        if not settings.enable_ce or settings.same_feed:
                             np.random.shuffle(full_lines)
 
                 batch_data = []
+                if settings.same_feed:
+                    temp_file = full_lines[0]
+                    print("Same images({},nums:{}) will feed in the net".format(
+                        str(temp_file), settings.same_feed))
+                    full_lines = []
+                    for i in range(settings.same_feed):
+                        full_lines.append(temp_file)
                 for line in full_lines:
                     img_path, label = line.split()
                     img_path = os.path.join(data_dir, img_path)

PaddleCV/image_classification/train.py

@@ -54,7 +54,7 @@ def build_program(is_train, main_prog, startup_prog, args):
     else:
         model = models.__dict__[args.model]()
     with fluid.program_guard(main_prog, startup_prog):
-        if args.random_seed:
+        if args.random_seed or args.enable_ce:
             main_prog.random_seed = args.random_seed
             startup_prog.random_seed = args.random_seed
         with fluid.unique_name.guard():
@@ -79,8 +79,14 @@ def build_program(is_train, main_prog, startup_prog, args):
     return loss_out
 
 
-def validate(args, test_iter, exe, test_prog, test_fetch_list, pass_id,
-             train_batch_metrics_record):
+def validate(args,
+             test_iter,
+             exe,
+             test_prog,
+             test_fetch_list,
+             pass_id,
+             train_batch_metrics_record,
+             train_batch_time_record=None):
     test_batch_time_record = []
     test_batch_metrics_record = []
     test_batch_id = 0
@@ -96,12 +102,11 @@ def validate(args, test_iter, exe, test_prog, test_fetch_list, pass_id,
         test_batch_metrics_avg = np.mean(np.array(test_batch_metrics), axis=1)
         test_batch_metrics_record.append(test_batch_metrics_avg)
 
-        print_info(pass_id, test_batch_id, args.print_step,
-                   test_batch_metrics_avg, test_batch_elapse, "batch")
+        print_info("batch", test_batch_metrics_avg, test_batch_elapse, pass_id,
+                   test_batch_id, args.print_step)
         sys.stdout.flush()
         test_batch_id += 1
 
-    #train_epoch_time_avg = np.mean(np.array(train_batch_time_record))
     train_epoch_metrics_avg = np.mean(
         np.array(train_batch_metrics_record), axis=0)
 
@@ -109,9 +114,18 @@ def validate(args, test_iter, exe, test_prog, test_fetch_list, pass_id,
     test_epoch_metrics_avg = np.mean(
         np.array(test_batch_metrics_record), axis=0)
 
-    print_info(pass_id, 0, 0,
+    print_info(
+        "epoch",
         list(train_epoch_metrics_avg) + list(test_epoch_metrics_avg),
-               test_epoch_time_avg, "epoch")
+        test_epoch_time_avg,
+        pass_id=pass_id)
+    if args.enable_ce:
+        device_num = fluid.core.get_cuda_device_count() if args.use_gpu else 1
+        print_info(
+            "ce",
+            list(train_epoch_metrics_avg) + list(test_epoch_metrics_avg),
+            train_batch_time_record,
+            device_num=device_num)
 
 
 def train(args):
@@ -207,8 +221,8 @@ def train(args):
                 np.array(train_batch_metrics), axis=1)
             train_batch_metrics_record.append(train_batch_metrics_avg)
             if trainer_id == 0:
-                print_info(pass_id, train_batch_id, args.print_step,
-                           train_batch_metrics_avg, train_batch_elapse, "batch")
+                print_info("batch", train_batch_metrics_avg, train_batch_elapse,
+                           pass_id, train_batch_id, args.print_step)
                 sys.stdout.flush()
             train_batch_id += 1
             t1 = time.time()
@@ -232,7 +246,7 @@ def train(args):
                 print('ExponentialMovingAverage validate over!')
 
             validate(args, test_iter, exe, test_prog, test_fetch_list, pass_id,
-                     train_batch_metrics_record)
+                     train_batch_metrics_record, train_batch_time_record)
             #For now, save model per epoch.
             if pass_id % args.save_step == 0:
                 save_model(args, exe, train_prog, pass_id)

PaddleCV/image_classification/utils/utility.py

@@ -136,7 +136,9 @@ def parse_args():
     add_arg('label_smoothing_epsilon',  float,  0.1,                    "The value of label_smoothing_epsilon parameter")
     #NOTE: (2019/08/08) temporary disable use_distill
     #add_arg('use_distill',              bool,   False,                  "Whether to use distill")
+    add_arg("enable_ce",                bool,   False,                  "Whether to enable ce")
     add_arg('random_seed',              int,    None,                   "random seed")
+
     add_arg('use_ema',                  bool,   False,                  "Whether to use ExponentialMovingAverage.")
     add_arg('ema_decay',                float,  0.9999,                 "The value of ema decay rate")
     add_arg('padding_type',             str,    "SAME",                 "Padding type of convolution")
@@ -146,6 +148,7 @@ def parse_args():
     add_arg('profiler_path',            str,    './',                   "the profiler output file path.(used for benchmark)")
     add_arg('max_iter',                 int,    0,                    "the max train batch num.(used for benchmark)")
     add_arg('validate',                 int,    1,                      "whether validate.(used for benchmark)")
+    add_arg('same_feed',                int,    0,                      "whether to feed same images")
 
 
     # yapf: enable
@@ -263,6 +266,10 @@ def check_args(args):
         args.data_dir
     ), "Data doesn't exist in {}, please load right path".format(args.data_dir)
 
+    if args.enable_ce:
+        args.random_seed = 0
+        print("CE is running now!")
+
     #check gpu
 
     check_gpu()
@@ -344,7 +351,13 @@ def create_data_loader(is_train, args):
         return data_loader, [feed_image, feed_label]
 
 
-def print_info(pass_id, batch_id, print_step, metrics, time_info, info_mode):
+def print_info(info_mode,
+               metrics,
+               time_info,
+               pass_id=0,
+               batch_id=0,
+               print_step=1,
+               device_num=1):
     """print function
 
     Args:
@@ -355,6 +368,7 @@ def print_info(pass_id, batch_id, print_step, metrics, time_info, info_mode):
         time_info: time infomation
         info_mode: mode
     """
+    #XXX: Use specific name to choose pattern, not the length of metrics. 
     if info_mode == "batch":
         if batch_id % print_step == 0:
             #if isinstance(metrics,np.ndarray):
@@ -402,11 +416,34 @@ def print_info(pass_id, batch_id, print_step, metrics, time_info, info_mode):
                        "%.5f" % test_acc5))
         sys.stdout.flush()
     elif info_mode == "ce":
-        raise Warning("CE code is not ready")
+        assert len(
+            metrics
+        ) == 7, "Enable CE: The Metrics should contain train_loss, train_acc1, train_acc5, test_loss, test_acc1, test_acc5, and train_speed"
+        assert len(
+            time_info
+        ) > 10, "0~9th batch statistics will drop when doing benchmark or ce, because it might be mixed with startup time, so please make sure training at least 10 batches."
+        print_ce(device_num, metrics, time_info)
+        #raise Warning("CE code is not ready")
     else:
         raise Exception("Illegal info_mode")
 
 
+def print_ce(device_num, metrics, time_info):
+    """ Print log for CE(for internal test).
+    """
+    train_loss, train_acc1, train_acc5, _, test_loss, test_acc1, test_acc5 = metrics
+
+    train_speed = np.mean(np.array(time_info[10:]))
+
+    print("kpis\ttrain_cost_card{}\t{}".format(device_num, train_loss))
+    print("kpis\ttrain_acc1_card{}\t{}".format(device_num, train_acc1))
+    print("kpis\ttrain_acc5_card{}\t{}".format(device_num, train_acc5))
+    print("kpis\ttest_loss_card{}\t{}".format(device_num, test_loss))
+    print("kpis\ttest_acc1_card{}\t{}".format(device_num, test_acc1))
+    print("kpis\ttest_acc5_card{}\t{}".format(device_num, test_acc5))
+    print("kpis\ttrain_speed_card{}\t{}".format(device_num, train_speed))
+
+
 def best_strategy_compiled(args, program, loss, exe):
     """make a program which wrapped by a compiled program
     """