Merge pull request #2115 from PaddlePaddle/dygraph-resnet

resnet at Dygraph mode

fluid/dygraph/resnet/README.md

+DyGraph模式下Residual Network实现
+========
+
+简介
+--------
+Residual Network（ResNet）是常用的图像分类模型。我们实现了在paddlepaddle的DyGraph模式下相应的实现。可以对比原先静态图下实现（[Residual Network](https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification/models)）来了解paddle中DyGraph模式。
+运行本目录下的程序示例需要使用PaddlePaddle develop最新版本。如果您的PaddlePaddle安装版本低于此要求，请按照[安装文档](http://www.paddlepaddle.org/docs/develop/documentation/zh/build_and_install/pip_install_cn.html)中的说明更新PaddlePaddle安装版本。
+
+
+## 代码结构
+```
+└── train.py     # 训练脚本。
+```
+
+## 使用的数据
+
+教程中使用`paddle.dataset.flowers`数据集作为训练数据，该数据集通过`paddle.dataset`模块自动下载到本地。
+
+## 训练测试Residual Network
+
+在GPU单卡上训练Residual Network:
+
+```
+env CUDA_VISIBLE_DEVICES=0 python train.py
+```
+
+这里`CUDA_VISIBLE_DEVICES=0`表示是执行在0号设备卡上，请根据自身情况修改这个参数。
+
+## 输出
+执行训练开始后，将得到类似如下的输出。每一轮`batch`训练将会打印当前epoch、step以及loss值。当前默认执行`epoch=10`, `batch_size=8`。您可以调整参数以得到更好的训练效果，同时也意味着消耗更多的内存（显存）以及需要花费更长的时间。
+```text
+epoch id: 0, batch step: 0, loss: 4.951202
+epoch id: 0, batch step: 1, loss: 5.268410
+epoch id: 0, batch step: 2, loss: 5.123999
+```

fluid/dygraph/resnet/train.py

+# Copyright (c) 2019 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 numpy as np
+
+import paddle
+import paddle.fluid as fluid
+from paddle.fluid.layer_helper import LayerHelper
+from paddle.fluid.dygraph.nn import Conv2D, Pool2D, BatchNorm, FC
+from paddle.fluid.dygraph.base import to_variable
+
+batch_size = 8
+epoch = 10
+
+
+def optimizer_setting():
+    return fluid.optimizer.SGD(learning_rate=0.01)
+
+
+class ConvBNLayer(fluid.dygraph.Layer):
+    def __init__(self,
+                 name_scope,
+                 num_channels,
+                 num_filters,
+                 filter_size,
+                 stride=1,
+                 groups=1,
+                 act=None):
+        super(ConvBNLayer, self).__init__(name_scope)
+
+        self._conv = Conv2D(
+            self.full_name(),
+            num_channels=num_channels,
+            num_filters=num_filters,
+            filter_size=filter_size,
+            stride=stride,
+            padding=(filter_size - 1) // 2,
+            groups=groups,
+            act=None,
+            bias_attr=None)
+
+        self._batch_norm = BatchNorm(self.full_name(), num_filters, act=act)
+
+    def forward(self, inputs):
+        y = self._conv(inputs)
+        y = self._batch_norm(y)
+
+        return y
+
+
+class BottleneckBlock(fluid.dygraph.Layer):
+    def __init__(self,
+                 name_scope,
+                 num_channels,
+                 num_filters,
+                 stride,
+                 shortcut=True):
+        super(BottleneckBlock, self).__init__(name_scope)
+
+        self.conv0 = ConvBNLayer(
+            self.full_name(),
+            num_channels=num_channels,
+            num_filters=num_filters,
+            filter_size=1,
+            act='relu')
+        self.conv1 = ConvBNLayer(
+            self.full_name(),
+            num_channels=num_filters,
+            num_filters=num_filters,
+            filter_size=3,
+            stride=stride,
+            act='relu')
+        self.conv2 = ConvBNLayer(
+            self.full_name(),
+            num_channels=num_filters,
+            num_filters=num_filters * 4,
+            filter_size=1,
+            act=None)
+
+        if not shortcut:
+            self.short = ConvBNLayer(
+                self.full_name(),
+                num_channels=num_channels,
+                num_filters=num_filters * 4,
+                filter_size=1,
+                stride=stride)
+
+        self.shortcut = shortcut
+
+        self._num_channels_out = num_filters * 4
+
+    def forward(self, inputs):
+        y = self.conv0(inputs)
+        conv1 = self.conv1(y)
+        conv2 = self.conv2(conv1)
+
+        if self.shortcut:
+            short = inputs
+        else:
+            short = self.short(inputs)
+
+        y = fluid.layers.elementwise_add(x=short, y=conv2)
+
+        layer_helper = LayerHelper(self.full_name(), act='relu')
+        return layer_helper.append_activation(y)
+
+
+class ResNet(fluid.dygraph.Layer):
+    def __init__(self, name_scope, layers=50, class_dim=102):
+        super(ResNet, self).__init__(name_scope)
+
+        self.layers = 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]
+
+        self.conv = ConvBNLayer(
+            self.full_name(),
+            num_channels=3,
+            num_filters=64,
+            filter_size=7,
+            stride=2,
+            act='relu')
+        self.pool2d_max = Pool2D(
+            self.full_name(),
+            pool_size=3,
+            pool_stride=2,
+            pool_padding=1,
+            pool_type='max')
+
+        self.bottleneck_block_list = []
+        num_channels = 64
+        for block in range(len(depth)):
+            shortcut = False
+            for i in range(depth[block]):
+                bottleneck_block = self.add_sublayer(
+                    'bb_%d_%d' % (block, i),
+                    BottleneckBlock(
+                        self.full_name(),
+                        num_channels=num_channels,
+                        num_filters=num_filters[block],
+                        stride=2 if i == 0 and block != 0 else 1,
+                        shortcut=shortcut))
+                num_channels = bottleneck_block._num_channels_out
+                self.bottleneck_block_list.append(bottleneck_block)
+                shortcut = True
+
+        self.pool2d_avg = Pool2D(
+            self.full_name(), pool_size=7, pool_type='avg', global_pooling=True)
+
+        import math
+        stdv = 1.0 / math.sqrt(2048 * 1.0)
+
+        self.out = FC(self.full_name(),
+                      size=class_dim,
+                      act='softmax',
+                      param_attr=fluid.param_attr.ParamAttr(
+                          initializer=fluid.initializer.Uniform(-stdv, stdv)))
+
+    def forward(self, inputs):
+        y = self.conv(inputs)
+        y = self.pool2d_max(y)
+        for bottleneck_block in self.bottleneck_block_list:
+            y = bottleneck_block(y)
+        y = self.pool2d_avg(y)
+        y = self.out(y)
+        return y
+
+
+def train_resnet():
+    with fluid.dygraph.guard():
+        resnet = ResNet("resnet")
+        optimizer = optimizer_setting()
+        train_reader = paddle.batch(
+            paddle.dataset.flowers.train(),
+            batch_size=batch_size)
+
+        for eop in range(epoch):
+            for batch_id, data in enumerate(train_reader()):
+                dy_x_data = np.array(
+                    [x[0].reshape(3, 224, 224) for x in data]).astype('float32')
+                if len(np.array([x[1] for x in data]).astype('int64')) != batch_size:
+                    continue
+                y_data = np.array([x[1] for x in data]).astype('int64').reshape(
+                    batch_size, 1)
+
+                img = to_variable(dy_x_data)
+                label = to_variable(y_data)
+                label._stop_gradient = True
+
+                out = resnet(img)
+                loss = fluid.layers.cross_entropy(input=out, label=label)
+                avg_loss = fluid.layers.mean(x=loss)
+
+                dy_out = avg_loss.numpy()
+                avg_loss.backward()
+
+                optimizer.minimize(avg_loss)
+                resnet.clear_gradients()
+
+                print("epoch id: %d, batch step: %d, loss: %f" % (eop, batch_id, dy_out))
+
+
+if __name__ == '__main__':
+    train_resnet()