cannot import name model_urls from torchvision.models.resnet假设出问题的文件是pytracking/libs/tensorflow/backbone.py如果不是修改逻辑完全一样。把原来的代码python# 这是旧代码会报错 from torchvision.models.resnet import model_urls改为python# 新代码从新版 torchvision 中获取权重 URL import torchvision.models as models # 方式一直接获取权重枚举对象的 url 属性 try: # 针对 torchvision 0.14 的写法 model_urls { resnet18: models.ResNet18_Weights.IMAGENET1K_V1.url, resnet34: models.ResNet34_Weights.IMAGENET1K_V1.url, resnet50: models.ResNet50_Weights.IMAGENET1K_V1.url, resnet101: models.ResNet101_Weights.IMAGENET1K_V1.url, resnet152: models.ResNet152_Weights.IMAGENET1K_V1.url, } except AttributeError: # 如果还是不行退回到旧版写法兼容 from torchvision.models.resnet import model_urls cannot import name ‘model_urls‘ from ‘torchvision.models.resnet‘
cannot import name model_urls from torchvision.models.resnet假设出问题的文件是pytracking/libs/tensorflow/backbone.py如果不是修改逻辑完全一样。把原来的代码python# 这是旧代码会报错 from torchvision.models.resnet import model_urls改为python# 新代码从新版 torchvision 中获取权重 URL import torchvision.models as models # 方式一直接获取权重枚举对象的 url 属性 try: # 针对 torchvision 0.14 的写法 model_urls { resnet18: models.ResNet18_Weights.IMAGENET1K_V1.url, resnet34: models.ResNet34_Weights.IMAGENET1K_V1.url, resnet50: models.ResNet50_Weights.IMAGENET1K_V1.url, resnet101: models.ResNet101_Weights.IMAGENET1K_V1.url, resnet152: models.ResNet152_Weights.IMAGENET1K_V1.url, } except AttributeError: # 如果还是不行退回到旧版写法兼容 from torchvision.models.resnet import model_urls