PyTorch ModuleNotFoundError

Question

我已经在GoogleColab中为对象检测培训了Retinanet，现在我想在另一个python项目中加载它的.pt文件，但我一直遇到这个错误。有什么想法吗

Traceback (most recent call last):
  File "C:\Users\stefan_cepa995\Desktop\breast-mammography-app\app.py", line 522, in <module>
    model = torch.load(os.path.join(".", "models", "retinanet", "retinanet_gwd.pt"))
  File "C:\Users\stefan_cepa995\anaconda3\envs\tensorflow\lib\site-packages\torch\serialization.py", line 594, in load
    return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args)
  File "C:\Users\stefan_cepa995\anaconda3\envs\tensorflow\lib\site-packages\torch\serialization.py", line 853, in _load
    result = unpickler.load()
ModuleNotFoundError: No module named 'retinanet'

Answer 1

但请尝试以下步骤：

pip install .  user

通过pip list检查keras-retinanet是否是已安装的模块。 如果没有，则安装它

pip install keras-retinanet

Answer 2

正如@Rika在评论中提到的，解决方案是保存state_dict，然后使用load_state_dict()函数加载它