pytorch入门ResNet

知识点1、ResNet 的关键结构
知识的2、定义重复使用的小函数
知识点3、函数结构微调
知识点4、实现bottleneck
知识点5、跟踪图像长宽的技巧
知识点1
ResNet的结构就像是 自动控制原理 中的前馈结构，这里称为bottleneck，用于减缓梯度消失的问题

import numpy as np
import torch
from torch import nn
import torch.nn.functional as F
from torch.autograd import Variable
from torchvision.datasets import CIFAR10

知识点2
把后面会多次重复使用的函数定义为更简洁的模式

def conv3x3(in_channels, out_channels, stride=1):
    return nn.Conv2d(in_channels, out_channels, 3, stride=stride, padding=1, bias=False)

知识点3
技巧1 学一下，通过不同的same_shape实现对结构的微调
知识点4
ResNet 最大的特点就是bottleneck的结构，在程序中只是用了F.relu(x+out, True)就可以实现了

class residual_block(nn.Module):
    def __init__(self, in_channel, out_channel, same_shape=True):
        super(residual_block, self).__init__()
        self.same_shape = same_shape    # 技巧1
        stride=1 if self.same_shape else 2
        self.conv1 = conv3x3(in_channel, out_channel, stride=stride)
        self.bn1 = nn.BatchNorm2d(out_channel)
        self.conv2 = conv3x3(out_channel, out_channel)
        self.bn2 = nn.BatchNorm2d(out_channel)
        if not self.same_shape:
            self.conv3 = nn.Conv2d(in_channel, out_channel, 1, stride=stride)

    def forward(self, x):
        out = self.conv1(x)
        out = F.relu(self.bn1(out), True)
        out = self.conv2(out)
        out = F.relu(self.bn2(out), True)
        if not self.same_shape:
            x = self.conv3(x)
        return F.relu(x+out, True)    # 点睛之笔

验证

test_net = residual_block(32, 32, False)
test_x = Variable(torch.zeros(1, 32, 96, 96))
print('input: {}'.format(test_x.shape))
test_y = test_net(test_x)
print('output: {}'.format(test_y.shape))

知识点5
定义一个verbose 当为True的时候，输出图像的长宽，以便分析

class resnet(nn.Module):
    def __init__(self, in_channel, num_class, verbose=False):
        super(resnet, self).__init__()
        self.verbose = verbose          # 技巧
        self.block1 = nn.Conv2d(in_channel, 64, 7, 2)
        self.block2 = nn.Sequential(
            nn.MaxPool2d(3,2),
            residual_block(64, 64),
            residual_block(64, 64)
        )
        self.block3 = nn.Sequential(
            residual_block(64, 128, False),
            residual_block(128, 128)
        )
        self.block4 = nn.Sequential(
            residual_block(128, 256, False),
            residual_block(256, 256)
        )
        self.block5 = nn.Sequential(
            residual_block(256, 512, False),
            residual_block(512, 512),
            nn.AvgPool2d(3)
        )
        self.classifier = nn.Linear(512, num_class)

    def forward(self, x):
        x = self.block1(x)
        if self.verbose:
            print('block 1 output: {}'.format(x.shape))
        x = self.block2(x)
        if self.verbose:
            print('block 2 output: {}'.format(x.shape))
        x = self.block3(x)
        if self.verbose:
            print('block 3 output: {}'.format(x.shape))
        x = self.block4(x)
        if self.verbose:
            print('block 4 output: {}'.format(x.shape))
        x = self.block5(x)
        if self.verbose:
            print('block 5 output: {}'.format(x.shape))
        x = x.view(x.shape[0], -1)
        x = self.classifier(x)
        return x

验证

test_net = resnet(3, 10, True)
test_x = Variable(torch.zeros(1, 3, 96, 96))
test_y = test_net(test_x)
print('output: {}'.format(test_y.shape))

基操，不多BB了

def data_tf(x):
    x = x.resize((96, 96), 2)
    x = np.array(x, dtype='float32') / 255
    x = (x - 0.5) / 0.5
    x = x.transpose((2, 0, 1))
    x = torch.from_numpy(x)
    return x


from torch.utils.data import DataLoader
from jc_utils import train
train_set = CIFAR10('./data', train=True, transform=data_tf)
train_data = DataLoader(train_set, batch_size=64, shuffle=True)
test_set = CIFAR10('./data', train=False, transform=data_tf)
test_data = DataLoader(test_set, batch_size=128, shuffle=False)


net = resnet(3, 10)
optimizer = torch.optim.SGD(net.parameters(), lr=0.01)
criterion = nn.CrossEntropyLoss()
train(net, train_data, test_data, 20, optimizer, criterion)