博客
关于我
PSPNet-Model(pytorch版本)
阅读量:590 次
发布时间:2019-03-11

本文共 3385 字,大约阅读时间需要 11 分钟。

P S P N e t − M o d e l ( p y t o r c h 版 本 ) PSPNet-Model(pytorch版本) PSPNetModel(pytorch)

训练、验证代码逻辑




import torchfrom torch import nnfrom torch.nn import functional as Fimport extractorsimport warningswarnings.filterwarnings("ignore")
class PSPModule(nn.Module):    def __init__(self, features, out_features=1024, sizes=(1, 2, 3, 6)):        super().__init__()        self.stages = []        self.stages = nn.ModuleList([self._make_stage(features, size) for size in sizes])        self.bottleneck = nn.Conv2d(features * (len(sizes) + 1), out_features, kernel_size=1)        self.relu = nn.ReLU()    def _make_stage(self, features, size):        prior = nn.AdaptiveAvgPool2d(output_size=(size, size))        conv = nn.Conv2d(features, features, kernel_size=1, bias=False)        return nn.Sequential(prior, conv)    def forward(self, feats):        h, w = feats.size(2), feats.size(3)        print(feats.size())        priors = [F.upsample(input=stage(feats), size=(h, w), mode='bilinear') for stage in self.stages] + [feats]        bottle = self.bottleneck(torch.cat(priors, 1))        return self.relu(bottle)class PSPUpsample(nn.Module):    def __init__(self, in_channels, out_channels):        super().__init__()        self.conv = nn.Sequential(            nn.Conv2d(in_channels, out_channels, 3, padding=1),            nn.BatchNorm2d(out_channels),            nn.PReLU()        )    def forward(self, x):        h, w = 2 * x.size(2), 2 * x.size(3)        p = F.upsample(input=x, size=(h, w), mode='bilinear')        return self.conv(p)
class PSPNet(nn.Module):    def __init__(self, n_classes=18, sizes=(1, 2, 3, 6), psp_size=2048, deep_features_size=256, backend='resnet34',                 pretrained=False):        super().__init__()        self.feats = getattr(extractors, backend)(pretrained)        self.psp = PSPModule(psp_size, 1024, sizes)        self.drop_1 = nn.Dropout2d(p=0.3)        self.up_1 = PSPUpsample(1024, 256)        self.up_2 = PSPUpsample(256, 64)        self.up_3 = PSPUpsample(64, 64)        self.drop_2 = nn.Dropout2d(p=0.15)        self.final = nn.Sequential(            nn.Conv2d(64, n_classes, kernel_size=1),            nn.LogSoftmax()        )        self.classifier = nn.Sequential(            nn.Linear(deep_features_size, 256),            nn.ReLU(),            nn.Linear(256, n_classes)        )    def forward(self, x):        print('x:', x.size())        f, class_f = self.feats(x);print('f:', f.size());print('class_f:', class_f.size());        p = self.psp(f);print('p:', p.size())        p = self.drop_1(p);print('p:', p.size())        p = self.up_1(p);print('p:', p.size())        p = self.drop_2(p);print('p:', p.size())        p = self.up_2(p);print('p:', p.size())        p = self.drop_2(p);print('p:', p.size())        p = self.up_3(p);print('p:', p.size())        p = self.drop_2(p);print('p:', p.size())        auxiliary = F.adaptive_max_pool2d(input=class_f, output_size=(1, 1)).view(-1, class_f.size(1));print('auxiliary:', auxiliary.size())        res1 = self.final(p);print('res1:', res1.size())        res2 = self.classifier(auxiliary);print('res2:', res2.size())        return res1 , res2
# 随机生成输入数据rgb = torch.randn(1, 3, 512, 512)# 定义网络net = PSPNet(psp_size=512,n_classes=8,deep_features_size=256)# 前向传播out, out_cls = net(rgb)# 打印输出大小print('---out--'*5)print(out.shape)print('--out_cls---'*5)print(out_cls)print('-----'*5)

在这里插入图片描述

转载地址:http://medtz.baihongyu.com/

你可能感兴趣的文章
mysql 主从 lock_mysql 主从同步权限mysql 行锁的实现
查看>>
mysql 主从互备份_mysql互为主从实战设置详解及自动化备份(Centos7.2)
查看>>
mysql 主从关系切换
查看>>
MYSQL 主从同步文档的大坑
查看>>
mysql 主键重复则覆盖_数据库主键不能重复
查看>>
Mysql 事务知识点与优化建议
查看>>
Mysql 优化 or
查看>>
mysql 优化器 key_mysql – 选择*和查询优化器
查看>>
MySQL 优化:Explain 执行计划详解
查看>>
Mysql 会导致锁表的语法
查看>>
mysql 使用sql文件恢复数据库
查看>>
mysql 修改默认字符集为utf8
查看>>
Mysql 共享锁
查看>>
MySQL 内核深度优化
查看>>
mysql 内连接、自然连接、外连接的区别
查看>>
mysql 写入慢优化
查看>>
mysql 分组统计SQL语句
查看>>
Mysql 分页
查看>>
Mysql 分页语句 Limit原理
查看>>
MySql 创建函数 Error Code : 1418
查看>>