[DO NOT MERGE]add dist resnet for paralle dygraph

dygraph/dist_resnet/README.md

@@ -0,0 +1,7 @@
+## Parallel Dygraph for the ResNet Training
+
+- Launch the training job
+
+    ``` bash
+    > python -m paddle.distributed.launch --gpus 8 dist_train.py
+    ```

dygraph/dist_resnet/dist_train.py

@@ -0,0 +1,274 @@
+# 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 time
+
+import paddle
+import paddle.fluid as fluid
+import paddle.fluid.dygraph as dygraph
+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
+from reader import train, val
+
+batch_size = 32
+epoch = 1
+
+total_images = 1281167
+
+
+def optimizer_setting():
+    optimizer = fluid.optimizer.Momentum(learning_rate=0.1, momentum=0.9)
+    return optimizer
+
+
+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.fc = 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, label):
+        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.fc(y)
+        acc1 = fluid.layers.accuracy(input=y, label=label, k=1)
+        acc5 = fluid.layers.accuracy(input=y, label=label, k=5)
+
+        return y, acc1, acc5
+
+
+def init_data(batch_size=32, img_shape=[3, 224, 224], label_range=9):
+    assert isinstance(img_shape, list)
+    np.random.seed(5)
+    input_shape = [batch_size] + img_shape
+    img = np.random.random(size=input_shape).astype(np.float32)
+    label = np.array([1 for _ in range(batch_size)]).reshape(
+        (-1, 1)).astype("int64")
+    #label = np.array(
+    #    [np.random.randint(0, label_range) for _ in range(batch_size)]).reshape(
+    #        (-1, 1)).astype("int64")
+    return img, label
+
+
+def train_resnet():
+    seed = 90
+    place = fluid.CUDAPlace(dygraph.parallel.Env().dev_id)
+    with fluid.dygraph.guard(place):
+        np.random.seed(seed)
+        import random
+        random.seed = seed
+
+        strategy = dygraph.parallel.prepare_context()
+        resnet = ResNet("dist_resnet", class_dim=1000)
+        resnet = dygraph.parallel.DataParallel(resnet, strategy)
+
+        optimizer = optimizer_setting()
+
+        train_reader = paddle.batch(
+            train(
+                data_dir="/imagenet/ImageNet_resize/",
+                pass_id_as_seed=0,
+                infinite=True),
+            batch_size=batch_size,
+            drop_last=True)
+        steps_per_epoch = int(total_images / strategy.nranks / batch_size)
+        print("steps per eoch: %d" % steps_per_epoch)
+
+        for eop in range(epoch):
+            for step_id, data in enumerate(train_reader()):
+                if step_id == steps_per_epoch:
+                    break
+                if len(np.array([x[1]
+                                 for x in data]).astype('int64')) != batch_size:
+                    continue
+
+                s_time = time.time()
+                #dy_x_data = np.array(
+                #    [x[0].reshape(3, 224, 224) for x in data]).astype('float32')
+                #y_data = np.array([x[1] for x in data]).astype('int64').reshape(
+                #    batch_size, 1)
+
+                dy_x_data, y_data = init_data()
+                print(np.sum(dy_x_data), np.sum(y_data))
+
+                img = to_variable(dy_x_data)
+                label = to_variable(y_data)
+                label._stop_gradient = True
+
+                out, acc1, acc5 = resnet(img, label)
+
+                loss = fluid.layers.cross_entropy(input=out, label=label)
+                avg_loss = fluid.layers.mean(x=loss)
+                dy_out = avg_loss.numpy()
+                avg_loss = resnet.scale_loss(avg_loss)
+                avg_loss.backward()
+                resnet.apply_collective_grads()
+                optimizer.minimize(avg_loss)
+                throughtput = batch_size / (time.time() - s_time)
+                print(
+                    "epoch id: %d, step: %d, loss: %f, acc1: %f, acc5: %f, throughtput: %f imgs/s "
+                    % (eop, step_id, dy_out, float(acc1.numpy()),
+                       float(acc5.numpy()), throughtput))
+                resnet.clear_gradients()
+
+
+if __name__ == '__main__':
+    train_resnet()