离线
TA的专栏
打卡等级：热心大叔
打卡总天数：205
打卡月天数：0
打卡总奖励：3099
最近打卡：2023-08-27 09:30:00
等级头衔
等級： 晓枫资讯-上等兵
在线时间 0 小时
积分成就
威望 0 点
贡献 388 次
主题 360 次 精华 0 篇 金钱 4237 枚
积分 773 枚 注册时间 2022-12-25
最后登录 2025-5-28
简单加载官方预训练模型
torchvision.models预定义了很多公开的模型结构
如果pretrained参数设置为False，那么仅仅设定模型结构；如果设置为True，那么会启动一个下载流程，下载预训练参数
如果只想调用模型，不想训练，那么设置model.eval()和model.requires_grad_(False)
想查看模型参数可以使用modules和named_modules，其中named_modules是一个长度为2的tuple，第一个变量是name，第二个变量是module本身。
# -*- coding: utf-8 -*- from torch import nn from torchvision import models # load model. If pretrained is True, there will be a downloading process model = models.vgg19(pretrained=True) model.eval() model.requires_grad_(False) # get model component features = model.features modules = features.modules() named_modules = features.named_modules() # print modules for module in modules: if isinstance(module, nn.Conv2d): weight = module.weight bias = module.bias print(module, weight.shape, bias.shape, weight.requires_grad, bias.requires_grad) elif isinstance(module, nn.ReLU): print(module) print() for named_module in named_modules: name = named_module[0] module = named_module[1] if isinstance(module, nn.Conv2d): weight = module.weight bias = module.bias print(name, module, weight.shape, bias.shape, weight.requires_grad, bias.requires_grad) elif isinstance(module, nn.ReLU): print(name, module) 复制代码 图片预处理
使用opencv和pil读图都可以使用transforms.ToTensor()把原本[H, W, 3]的数据转成[3, H, W]的tensor。但opencv要注意把数据改成RGB顺序。
vgg系列模型需要做normalization，建议配合torchvision.transforms来实现。
mini-batches of 3-channel RGB images of shape (3 x H x W), where H and W are expected to be at least 224. The images have to be loaded in to a range of [0, 1] and then normalized using mean = [0.485, 0.456, 0.406] and std = [0.229, 0.224, 0.225].
参考：https://pytorch.org/hub/pytorch_vision_vgg/
# -*- coding: utf-8 -*- from PIL import Image import cv2 import torch from torchvision import transforms # transforms for preprocess preprocess = transforms.Compose([ transforms.ToTensor(), transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]) ]) # load image using cv2 image_cv2 = cv2.imread('lena_std.bmp') image_cv2 = cv2.cvtColor(image_cv2, cv2.COLOR_BGR2RGB) image_cv2 = preprocess(image_cv2) # load image using pil image_pil = Image.open('lena_std.bmp') image_pil = preprocess(image_pil) # check whether image_cv2 and image_pil are same print(torch.all(image_cv2 == image_pil)) print(image_cv2.shape, image_pil.shape) 复制代码 提取单个特征图
如果只提取单层特征图，可以把模型截断，以节省算力和显存消耗。
下面索引之所以有+1是因为pytorch预训练模型里面第一个索引的module总是完整模块结构，第二个才开始子模块。
# -*- coding: utf-8 -*- from PIL import Image from torchvision import models from torchvision import transforms # load model. If pretrained is True, there will be a downloading process model = models.vgg19(pretrained=True) model = model.features[:16 + 1] # 16 = conv3_4 model.eval() model.requires_grad_(False) model.to('cuda') print(model) # load and preprocess image preprocess = transforms.Compose([ transforms.ToTensor(), transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]), transforms.Resize(size=(224, 224)) ]) image = Image.open('lena_std.bmp') image = preprocess(image) inputs = image.unsqueeze(0) # add batch dimension inputs = inputs.cuda() # forward output = model(inputs) print(output.shape) 复制代码 提取多个特征图
第一种方式：逐层运行model，如果碰到了需要保存的feature map就存下来。
第二种方式：使用register_forward_hook，使用这种方式需要用一个类把feature map以成员变量的形式缓存下来。
两种方式的运行效率差不多
第一种方式简单直观，但是只能处理类似VGG这种没有跨层连接的网络；第二种方式更加通用。
# -*- coding: utf-8 -*- from PIL import Image import torch from torchvision import models from torchvision import transforms # load model. If pretrained is True, there will be a downloading process model = models.vgg19(pretrained=True) model = model.features[:16 + 1] # 16 = conv3_4 model.eval() model.requires_grad_(False) model.to('cuda') # check module name for named_module in model.named_modules(): name = named_module[0] module = named_module[1] print('-------- %s --------' % name) print(module) print() # load and preprocess image preprocess = transforms.Compose([ transforms.ToTensor(), transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]), transforms.Resize(size=(224, 224)) ]) image = Image.open('lena_std.bmp') image = preprocess(image) inputs = image.unsqueeze(0) # add batch dimension inputs = inputs.cuda() # forward - 1 layers = [2, 7, 8, 9, 16] layers = sorted(set(layers)) feature_maps = {} feature = inputs for i in range(max(layers) + 1): feature = model[i](feature) if i in layers: feature_maps[i] = feature for key in feature_maps: print(key, feature_maps.get(key).shape) # forward - 2 class FeatureHook: def __init__(self, module): self.inputs = None self.output = None self.hook = module.register_forward_hook(self.get_features) def get_features(self, module, inputs, output): self.inputs = inputs self.output = output layer_names = ['2', '7', '8', '9', '16'] hook_modules = [] for named_module in model.named_modules(): name = named_module[0] module = named_module[1] if name in layer_names: hook_modules.append(module) hooks = [FeatureHook(module) for module in hook_modules] output = model(inputs) features = [hook.output for hook in hooks] for feature in features: print(feature.shape) # check correctness for i, layer in enumerate(layers): feature1 = feature_maps.get(layer) feature2 = features[i] print(torch.all(feature1 == feature2)) 复制代码 使用第二种方式（register_forward_hook），resnet特征图也可以顺利拿到。
而由于resnet的model已经不可以用model
的形式索引，所以无法使用第一种方式。# -*- coding: utf-8 -*- from PIL import Image from torchvision import models from torchvision import transforms # load model. If pretrained is True, there will be a downloading process model = models.resnet18(pretrained=True) model.eval() model.requires_grad_(False) model.to('cuda') # check module name for named_module in model.named_modules(): name = named_module[0] module = named_module[1] print('-------- %s --------' % name) print(module) print() # load and preprocess image preprocess = transforms.Compose([ transforms.ToTensor(), transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]), transforms.Resize(size=(224, 224)) ]) image = Image.open('lena_std.bmp') image = preprocess(image) inputs = image.unsqueeze(0) # add batch dimension inputs = inputs.cuda() class FeatureHook: def __init__(self, module): self.inputs = None self.output = None self.hook = module.register_forward_hook(self.get_features) def get_features(self, module, inputs, output): self.inputs = inputs self.output = output layer_names = [ 'conv1', 'layer1.0.relu', 'layer2.0.conv1' ] hook_modules = [] for named_module in model.named_modules(): name = named_module[0] module = named_module[1] if name in layer_names: hook_modules.append(module) hooks = [FeatureHook(module) for module in hook_modules] output = model(inputs) features = [hook.output for hook in hooks] for feature in features: print(feature.shape) 复制代码 print(list(model.children()) model = torch.nn.Sequential(*list(model.children())[:6]) 复制代码
晓枫资讯-科技资讯社区-免责声明
免责声明：以上内容为本网站转自其它媒体，相关信息仅为传递更多信息之目的，不代表本网观点，亦不代表本网站赞同其观点或证实其内容的真实性。
1 、注册用户在本社区发表、转载的任何作品仅代表其个人观点，不代表本社区认同其观点。
2 、管理员及版主有权在不事先通知或不经作者准许的情况下删除其在本社区所发表的文章。
3 、本社区的文章部分内容可能来源于网络，仅供大家学习与参考，如有侵权，举报反馈：
进行删除处理。
4 、本社区一切资源不代表本站立场，并不代表本站赞同其观点和对其真实性负责。
5 、以上声明内容的最终解释权归《晓枫资讯-科技资讯社区》所有。
发表于 2023-2-10 22:04:33