前几天讲解了PaddlePaddle框架使用的核心教程、超深层网络的搭建技巧、自定义网络层、参数特定初始化方法。

在PaddlePaddle中，有没有极简使用方法呢？

答案是肯定的，即直接在paddle.nn.Sequential()顺序容器中堆叠神经网络层。

import paddle


net = paddle.nn.Sequential(
    paddle.nn.Conv2D(in_channels=3, out_channels=16, kernel_size=3, stride=1, padding=1),
    paddle.nn.ReLU(),
    paddle.nn.MaxPool2D(kernel_size=2, stride=2),
    paddle.nn.Conv2D(in_channels=16, out_channels=32, kernel_size=1, stride=1, padding=1),
    paddle.nn.ReLU(),
    paddle.nn.MaxPool2D(kernel_size=2, stride=2),
    paddle.nn.Flatten(),
    paddle.nn.Linear(in_features=2048, out_features=1024),
    paddle.nn.ReLU(),
    paddle.nn.Dropout(0.5),
    paddle.nn.Linear(in_features=1024, out_features=10)
)

model = paddle.Model(net)

model.summary(input_size=(None, 3, 28, 28))

# 训练时候直接使用model.fit
model.fit(train_data=None, eval_data=None, batch_size=1, epochs=1)

输出模型结构信息如下：

---------------------------------------------------------------------------
 Layer (type)       Input Shape          Output Shape         Param #    
===========================================================================
   Conv2D-1       [[1, 3, 28, 28]]     [1, 16, 28, 28]          448      
    ReLU-1       [[1, 16, 28, 28]]     [1, 16, 28, 28]           0       
  MaxPool2D-1    [[1, 16, 28, 28]]     [1, 16, 14, 14]           0       
   Conv2D-2      [[1, 16, 14, 14]]     [1, 32, 16, 16]          544      
    ReLU-2       [[1, 32, 16, 16]]     [1, 32, 16, 16]           0       
  MaxPool2D-2    [[1, 32, 16, 16]]      [1, 32, 8, 8]            0       
   Flatten-1      [[1, 32, 8, 8]]         [1, 2048]              0       
   Linear-1         [[1, 2048]]           [1, 1024]          2,098,176   
    ReLU-3          [[1, 1024]]           [1, 1024]              0       
   Dropout-1        [[1, 1024]]           [1, 1024]              0       
   Linear-2         [[1, 1024]]            [1, 10]            10,250     
===========================================================================
Total params: 2,109,418
Trainable params: 2,109,418
Non-trainable params: 0
---------------------------------------------------------------------------
Input size (MB): 0.01
Forward/backward pass size (MB): 0.40
Params size (MB): 8.05
Estimated Total Size (MB): 8.45
---------------------------------------------------------------------------