黑松山资源网 Design By www.paidiu.com
有时候我们在fintune时发现pytorch把许多层都集合在一个sequential里,但是我们希望能把中间层的结果引出来做下一步操作,于是我自己琢磨了一个方法,以vgg为例,有点僵硬哈!
首先pytorch自带的vgg16模型的网络结构如下:
VGG( (features): Sequential( (0): Conv2d (3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (1): ReLU(inplace) (2): Conv2d (64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (3): ReLU(inplace) (4): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), dilation=(1, 1)) (5): Conv2d (64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (6): ReLU(inplace) (7): Conv2d (128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (8): ReLU(inplace) (9): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), dilation=(1, 1)) (10): Conv2d (128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (11): ReLU(inplace) (12): Conv2d (256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (13): ReLU(inplace) (14): Conv2d (256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (15): ReLU(inplace) (16): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), dilation=(1, 1)) (17): Conv2d (256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (18): ReLU(inplace) (19): Conv2d (512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (20): ReLU(inplace) (21): Conv2d (512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (22): ReLU(inplace) (23): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), dilation=(1, 1)) (24): Conv2d (512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (25): ReLU(inplace) (26): Conv2d (512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (27): ReLU(inplace) (28): Conv2d (512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (29): ReLU(inplace) (30): MaxPool2d(kernel_size=(2, 2), stride=(2, 2), dilation=(1, 1)) ) (classifier): Sequential( (0): Linear(in_features=25088, out_features=4096) (1): ReLU(inplace) (2): Dropout(p=0.5) (3): Linear(in_features=4096, out_features=4096) (4): ReLU(inplace) (5): Dropout(p=0.5) (6): Linear(in_features=4096, out_features=1000) ) )
我们需要fintune vgg16的features部分,并且我希望把3,8, 15, 22, 29这五个作为输出进一步操作。我的想法是自己写一个vgg网络,这个网络参数与pytorch的网络一致但是保证我们需要的层输出在sequential外。于是我写的网络如下:
class our_vgg(nn.Module): def __init__(self): super(our_vgg, self).__init__() self.conv1 = nn.Sequential( # conv1 nn.Conv2d(3, 64, 3, padding=35), nn.ReLU(inplace=True), nn.Conv2d(64, 64, 3, padding=1), nn.ReLU(inplace=True), ) self.conv2 = nn.Sequential( # conv2 nn.MaxPool2d(2, stride=2, ceil_mode=True), # 1/2 nn.Conv2d(64, 128, 3, padding=1), nn.ReLU(inplace=True), nn.Conv2d(128, 128, 3, padding=1), nn.ReLU(inplace=True), ) self.conv3 = nn.Sequential( # conv3 nn.MaxPool2d(2, stride=2, ceil_mode=True), # 1/4 nn.Conv2d(128, 256, 3, padding=1), nn.ReLU(inplace=True), nn.Conv2d(256, 256, 3, padding=1), nn.ReLU(inplace=True), nn.Conv2d(256, 256, 3, padding=1), nn.ReLU(inplace=True), ) self.conv4 = nn.Sequential( # conv4 nn.MaxPool2d(2, stride=2, ceil_mode=True), # 1/8 nn.Conv2d(256, 512, 3, padding=1), nn.ReLU(inplace=True), nn.Conv2d(512, 512, 3, padding=1), nn.ReLU(inplace=True), nn.Conv2d(512, 512, 3, padding=1), nn.ReLU(inplace=True), ) self.conv5 = nn.Sequential( # conv5 nn.MaxPool2d(2, stride=2, ceil_mode=True), # 1/16 nn.Conv2d(512, 512, 3, padding=1), nn.ReLU(inplace=True), nn.Conv2d(512, 512, 3, padding=1), nn.ReLU(inplace=True), nn.Conv2d(512, 512, 3, padding=1), nn.ReLU(inplace=True), ) def forward(self, x): conv1 = self.conv1(x) conv2 = self.conv2(conv1) conv3 = self.conv3(conv2) conv4 = self.conv4(conv3) conv5 = self.conv5(conv4) return conv5
接着就是copy weights了:
def convert_vgg(vgg16):#vgg16是pytorch自带的 net = our_vgg()# 我写的vgg vgg_items = net.state_dict().items() vgg16_items = vgg16.items() pretrain_model = {} j = 0 for k, v in net.state_dict().iteritems():#按顺序依次填入 v = vgg16_items[j][1] k = vgg_items[j][0] pretrain_model[k] = v j += 1 return pretrain_model ## net是我们最后使用的网络,也是我们想要放置weights的网络 net = net() print ('load the weight from vgg') pretrained_dict = torch.load('vgg16.pth') pretrained_dict = convert_vgg(pretrained_dict) model_dict = net.state_dict() # 1. 把不属于我们需要的层剔除 pretrained_dict = {k: v for k, v in pretrained_dict.items() if k in model_dict} # 2. 把参数存入已经存在的model_dict model_dict.update(pretrained_dict) # 3. 加载更新后的model_dict net.load_state_dict(model_dict) print ('copy the weight sucessfully')
这样我就基本达成目标了,注意net也就是我们要使用的网络fintune部分需要和our_vgg一致。
以上这篇pytorch在fintune时将sequential中的层输出方法,以vgg为例就是小编分享给大家的全部内容了,希望能给大家一个参考,也希望大家多多支持。
黑松山资源网 Design By www.paidiu.com
广告合作:本站广告合作请联系QQ:858582 申请时备注:广告合作(否则不回)
免责声明:本站资源来自互联网收集,仅供用于学习和交流,请遵循相关法律法规,本站一切资源不代表本站立场,如有侵权、后门、不妥请联系本站删除!
免责声明:本站资源来自互联网收集,仅供用于学习和交流,请遵循相关法律法规,本站一切资源不代表本站立场,如有侵权、后门、不妥请联系本站删除!
黑松山资源网 Design By www.paidiu.com
暂无评论...
更新日志
2024年10月07日
2024年10月07日
- 群星《前途海量 电影原声专辑》[FLAC/分轨][227.78MB]
- 张信哲.1992-知道新曲与精丫巨石】【WAV+CUE】
- 王翠玲.1995-ANGEL【新艺宝】【WAV+CUE】
- 景冈山.1996-我的眼里只有你【大地唱片】【WAV+CUE】
- 群星《八戒 电影原声带》[320K/MP3][188.97MB]
- 群星《我的阿勒泰 影视原声带》[320K/MP3][139.47MB]
- 纪钧瀚《胎教古典音乐 钢琴与大提琴的沉浸时光》[320K/MP3][148.91MB]
- 刘雅丽.2001-丽花皇后·EMI精选王【EMI百代】【FLAC分轨】
- 齐秦.1994-黄金十年1981-1990CHINA.TOUR.LIVE精丫上华】【WAV+CUE】
- 群星.2008-本色·百代音乐人创作专辑【EMI百代】【WAV+CUE】
- 群星.2001-同步过冬AVCD【环球】【WAV+CUE】
- 群星.2020-同步过冬2020冀待晴空【环球】【WAV+CUE】
- 沈雁.1986-四季(2012梦田复刻版)【白云唱片】【WAV+CUE】
- 纪钧瀚《胎教古典音乐 钢琴与大提琴的沉浸时光》[FLAC/分轨][257.88MB]
- 《国语老歌 怀旧篇 3CD》[WAV/分轨][1.6GB]