add regularizer api (#27292)

python/paddle/__init__.py

@@ -49,6 +49,7 @@ import paddle.optimizer
 import paddle.metric
 import paddle.device
 import paddle.incubate.complex as complex
+import paddle.regularizer
 
 # TODO: define alias in tensor and framework directory
 

python/paddle/fluid/tests/unittests/test_regularizer_api.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.
+
+from __future__ import print_function
+
+import unittest
+from functools import partial
+import contextlib
+import numpy as np
+import paddle
+import paddle.fluid.core as core
+import paddle.fluid as fluid
+import paddle.fluid.framework as framework
+import paddle.fluid.optimizer as optimizer
+import paddle.regularizer as regularizer
+from paddle.fluid.backward import append_backward
+
+
+def bow_net(data,
+            label,
+            dict_dim,
+            is_sparse=False,
+            emb_dim=8,
+            hid_dim=8,
+            hid_dim2=6,
+            class_dim=2):
+    """
+    BOW net
+    This model is from https://github.com/PaddlePaddle/models:
+    fluid/PaddleNLP/text_classification/nets.py
+    """
+    emb = fluid.layers.embedding(
+        input=data, is_sparse=is_sparse, size=[dict_dim, emb_dim])
+    bow = fluid.layers.sequence_pool(input=emb, pool_type='sum')
+    bow_tanh = fluid.layers.tanh(bow)
+    fc_1 = fluid.layers.fc(input=bow_tanh, size=hid_dim, act="tanh")
+    fc_2 = fluid.layers.fc(input=fc_1, size=hid_dim2, act="tanh")
+    prediction = fluid.layers.fc(input=[fc_2], size=class_dim, act="softmax")
+    cost = fluid.layers.cross_entropy(input=prediction, label=label)
+    avg_cost = fluid.layers.mean(x=cost)
+
+    return avg_cost
+
+
+class TestRegularizer(unittest.TestCase):
+    def setUp(self):
+        self.word_dict = paddle.dataset.imdb.word_dict()
+        reader = paddle.batch(
+            paddle.dataset.imdb.train(self.word_dict), batch_size=1)()
+        self.train_data = [next(reader) for _ in range(1)]
+
+    def get_places(self):
+        places = [core.CPUPlace()]
+        if core.is_compiled_with_cuda():
+            places.append(core.CUDAPlace(0))
+        return places
+
+    @contextlib.contextmanager
+    def scope_prog_guard(self, main_prog, startup_prog):
+        scope = fluid.core.Scope()
+        with fluid.unique_name.guard():
+            with fluid.scope_guard(scope):
+                with fluid.program_guard(main_prog, startup_prog):
+                    yield
+
+    def run_program(self, place, feed_list):
+        exe = fluid.Executor(place)
+        feeder = fluid.DataFeeder(feed_list=feed_list, place=place)
+        exe.run(fluid.default_startup_program())
+
+        main_prog = fluid.default_main_program()
+        param_list = [var.name for var in main_prog.block(0).all_parameters()]
+
+        param_sum = []
+        for data in self.train_data:
+            out = exe.run(main_prog,
+                          feed=feeder.feed(data),
+                          fetch_list=param_list)
+            p_sum = 0
+            for v in out:
+                p_sum += np.sum(np.abs(v))
+            param_sum.append(p_sum)
+        return param_sum
+
+    def check_l2decay_regularizer(self, place, model):
+        paddle.manual_seed(1)
+        paddle.framework.random._manual_program_seed(1)
+        main_prog = fluid.framework.Program()
+        startup_prog = fluid.framework.Program()
+        with self.scope_prog_guard(
+                main_prog=main_prog, startup_prog=startup_prog):
+            data = fluid.layers.data(
+                name="words", shape=[1], dtype="int64", lod_level=1)
+            label = fluid.layers.data(name="label", shape=[1], dtype="int64")
+
+            avg_cost = model(data, label, len(self.word_dict))
+
+            optimizer = fluid.optimizer.Adagrad(
+                learning_rate=0.1,
+                regularization=paddle.regularizer.L2Decay(1.0))
+            optimizer.minimize(avg_cost)
+            param_sum = self.run_program(place, [data, label])
+        return param_sum
+
+    def check_l2decay(self, place, model):
+        paddle.manual_seed(1)
+        paddle.framework.random._manual_program_seed(1)
+        main_prog = fluid.framework.Program()
+        startup_prog = fluid.framework.Program()
+
+        with self.scope_prog_guard(
+                main_prog=main_prog, startup_prog=startup_prog):
+            data = fluid.layers.data(
+                name="words", shape=[1], dtype="int64", lod_level=1)
+            label = fluid.layers.data(name="label", shape=[1], dtype="int64")
+
+            avg_cost_l2 = model(data, label, len(self.word_dict))
+
+            param_list = fluid.default_main_program().block(0).all_parameters()
+            para_sum = []
+            for para in param_list:
+                para_mul = fluid.layers.square(x=para)
+                para_sum.append(fluid.layers.reduce_sum(input=para_mul))
+            avg_cost_l2 += fluid.layers.sums(para_sum) * .5
+
+            optimizer = fluid.optimizer.Adagrad(learning_rate=0.1)
+            optimizer.minimize(avg_cost_l2)
+            param_sum = self.run_program(place, [data, label])
+        return param_sum
+
+    def test_l2(self):
+        for place in self.get_places():
+            dense_sparse_p_sum = []
+            for sparse in [True, False]:
+                model = partial(bow_net, is_sparse=sparse)
+                framework_l2 = self.check_l2decay_regularizer(place, model)
+                l2 = self.check_l2decay(place, model)
+                assert len(l2) == len(framework_l2)
+                for i in range(len(l2)):
+                    assert np.isclose(a=framework_l2[i], b=l2[i], rtol=5e-5)
+                dense_sparse_p_sum.append(framework_l2)
+
+            assert len(dense_sparse_p_sum[0]) == len(dense_sparse_p_sum[1])
+            for i in range(len(dense_sparse_p_sum[0])):
+                assert np.isclose(
+                    a=dense_sparse_p_sum[0][i],
+                    b=dense_sparse_p_sum[1][i],
+                    rtol=5e-5)
+
+    def test_repeated_regularization(self):
+        l1 = paddle.regularizer.L1Decay(0.1)
+        l2 = paddle.regularizer.L2Decay(0.01)
+        fc_param_attr = fluid.ParamAttr(regularizer=l1)
+        with fluid.program_guard(fluid.Program(), fluid.Program()):
+            x = fluid.layers.uniform_random([2, 2, 3])
+            out = fluid.layers.fc(x, 5, param_attr=fc_param_attr)
+            loss = fluid.layers.reduce_sum(out)
+            sgd = fluid.optimizer.SGD(learning_rate=0.1, regularization=l2)
+            sgd.minimize(loss)
+        with fluid.dygraph.guard():
+            input = fluid.dygraph.to_variable(
+                np.random.randn(3, 2).astype('float32'))
+            paddle.manual_seed(1)
+            paddle.framework.random._manual_program_seed(1)
+
+            linear1 = fluid.dygraph.Linear(
+                2, 2, param_attr=fc_param_attr, bias_attr=fc_param_attr)
+            linear2 = fluid.dygraph.Linear(
+                2, 2, param_attr=fc_param_attr, bias_attr=fc_param_attr)
+
+            loss1 = linear1(input)
+            loss1.backward()
+            # set l2 regularizer in optimizer, but l1 in fluid.ParamAttr
+
+            fluid.optimizer.SGD(parameter_list=linear1.parameters(),
+                                learning_rate=1e-2,
+                                regularization=l2).minimize(loss1)
+            # only set l1 in fluid.ParamAttr
+            loss2 = linear2(input)
+            loss2.backward()
+            fluid.optimizer.SGD(parameter_list=linear2.parameters(),
+                                learning_rate=1e-2).minimize(loss2)
+            # they should both be applied by l1, and keep the same
+            self.assertTrue(
+                np.allclose(linear1.weight.numpy(), linear2.weight.numpy()),
+                "weight should use the regularization in fluid.ParamAttr!")
+            self.assertTrue(
+                np.allclose(linear1.bias.numpy(), linear2.bias.numpy()),
+                "bias should use the regularization in fluid.ParamAttr!")
+
+
+if __name__ == '__main__':
+    unittest.main()

python/paddle/regularizer.py

@@ -12,8 +12,134 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-# TODO: define the regularizer functions 
-# __all__ = ['L1Decay',
-#            'L1DecayRegularizer',
-#            'L2Decay',
-#            'L2DecayRegularizer']
+__all__ = ['L1Decay', 'L2Decay']
+
+import paddle.fluid as fluid
+
+
+class L1Decay(fluid.regularizer.L1Decay):
+    """
+    Implement the L1 Weight Decay Regularization, which encourages the weights to be sparse.
+    
+    It can be set in :ref:`api_fluid_ParamAttr` or ``optimizer`` (such as :ref:`api_paddle_optimizer_Momentum` ). 
+    When set in ``ParamAttr`` , it only takes effect for trainable parameters in this layer. When set in 
+    ``optimizer`` , it takes effect for all trainable parameters. When set together, ``ParamAttr`` has 
+    higher priority than ``optimizer`` , which means that for a trainable parameter, if regularizer is defined 
+    in its ParamAttr, then the regularizer in Optimizer will be ignored. Otherwise the  regularizer
+    in Optimizer will be used.
+    
+    In the implementation, the formula of L1 Weight Decay Regularization is as follows:
+	
+    .. math::
+
+        L1WeightDecay = reg\_coeff * sign(parameter)
+
+    Args:
+        coeff(float, optional): regularization coeff. Default:0.0.
+	
+    Examples:
+        .. code-block:: python
+
+            # Example1: set Regularizer in optimizer
+            import paddle
+            from paddle.regularizer import L1Decay
+            import numpy as np
+            paddle.disable_static()
+            inp = np.random.uniform(-0.1, 0.1, [10, 10]).astype("float32")
+            linear = paddle.nn.Linear(10, 10)
+            inp = paddle.to_tensor(inp)
+            out = linear(inp)
+            loss = paddle.mean(out)
+            beta1 = paddle.to_tensor([0.9], dtype="float32")
+            beta2 = paddle.to_tensor([0.99], dtype="float32")
+            momentum = paddle.optimizer.Momentum(
+                learning_rate=0.1,
+                parameters=linear.parameters(),
+                weight_decay=L1Decay(0.0001))
+            back = out.backward()
+            momentum.step()
+            momentum.clear_grad()
+
+            # Example2: set Regularizer in parameters
+            # Set L1 regularization in parameters.
+            # Global regularizer does not take effect on my_conv2d for this case.
+            from paddle.nn import Conv2d
+            from paddle import ParamAttr
+            from paddle.regularizer import L2Decay
+
+            my_conv2d = Conv2d(
+                    in_channels=10,
+                    out_channels=10,
+                    kernel_size=1,
+                    stride=1,
+                    padding=0,
+                    weight_attr=ParamAttr(regularizer=L2Decay(coeff=0.01)),
+                    bias_attr=False)
+    """
+
+    def __init__(self, coeff=0.0):
+        super(L1Decay, self).__init__(coeff)
+
+
+class L2Decay(fluid.regularizer.L2Decay):
+    """
+    Implement the L2 Weight Decay Regularization, which helps to prevent the model over-fitting.
+    
+    It can be set in :ref:`api_fluid_ParamAttr` or ``optimizer`` (such as :ref:`api_paddle_optimizer_Momentum` ). 
+    When set in ``ParamAttr`` , it only takes effect for trainable parameters in this layer. When set in 
+    ``optimizer`` , it takes effect for all trainable parameters. When set together, ``ParamAttr`` has 
+    higher priority than ``optimizer`` , which means that for a trainable parameter, if regularizer is defined 
+    in its ParamAttr, then the regularizer in Optimizer will be ignored. Otherwise the  regularizer
+    in Optimizer will be used.
+    
+    In the implementation, the formula of L2 Weight Decay Regularization is as follows:
+
+    .. math::
+
+        L2WeightDecay = reg\_coeff * parameter
+
+    Args:
+        regularization_coeff(float, optional): regularization coeff. Default:0.0
+	
+    Examples:
+        .. code-block:: python
+
+            # Example1: set Regularizer in optimizer
+            import paddle
+            from paddle.regularizer import L2Decay
+            import numpy as np
+            paddle.disable_static()
+            inp = np.random.uniform(-0.1, 0.1, [10, 10]).astype("float32")
+            linear = paddle.nn.Linear(10, 10)
+            inp = paddle.to_tensor(inp)
+            out = linear(inp)
+            loss = paddle.mean(out)
+            beta1 = paddle.to_tensor([0.9], dtype="float32")
+            beta2 = paddle.to_tensor([0.99], dtype="float32")
+            momentum = paddle.optimizer.Momentum(
+                learning_rate=0.1,
+                parameters=linear.parameters(),
+                weight_decay=L2Decay(0.0001))
+            back = out.backward()
+            momentum.step()
+            momentum.clear_grad()
+
+            # Example2: set Regularizer in parameters
+            # Set L2 regularization in parameters.
+            # Global regularizer does not take effect on my_conv2d for this case.
+            from paddle.nn import Conv2d
+            from paddle import ParamAttr
+            from paddle.regularizer import L2Decay
+
+            my_conv2d = Conv2d(
+                    in_channels=10,
+                    out_channels=10,
+                    kernel_size=1,
+                    stride=1,
+                    padding=0,
+                    weight_attr=ParamAttr(regularizer=L2Decay(coeff=0.01)),
+                    bias_attr=False)
+    """
+
+    def __init__(self, coeff=0.0):
+        super(L2Decay, self).__init__(coeff)

python/paddle/utils/__init__.py

@@ -16,12 +16,13 @@ from .profiler import ProfilerOptions
 from .profiler import Profiler
 from .profiler import get_profiler
 from .deprecated import deprecated
+from ..fluid.framework import unique_name
+from ..fluid.framework import load_op_library
+from ..fluid.framework import require_version
 
 from . import download
 
 __all__ = ['dump_config', 'deprecated', 'download']
 
 #TODO: define new api under this directory
-# __all__ = ['unique_name',
-#            'load_op_library',
-#            'require_version']
+__all__ += ['unique_name', 'load_op_library', 'require_version']