python - Pytorch：AttributeError：'function'对象没有属性'copy'

Question

我正在尝试加载state_dict我在 Google Colab GPU 上训练的模型，这是我加载模型的代码：

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

model = models.resnet50()
num_ftrs = model.fc.in_features
model.fc = nn.Linear(num_ftrs, n_classes)
model.load_state_dict(copy.deepcopy(torch.load("./models/model.pth",device)))
model = model.to(device)
model.eval()

这是错误：

state_dict = state_dict.copy()

AttributeError：“函数”对象没有属性“副本”

火炬：

>>> import torch
>>> print (torch.__version__)
1.4.0
>>> import torchvision
>>> print (torchvision.__version__)
0.5.0

请帮助我到处搜索无济于事

[完整错误详情][1] https://i.stack.imgur.com/s22DL.png

标签： pythondeep-learningpytorchtorch

Answer 1

我猜这是你做错了。你保存了函数

torch.save(model.state_dict, 'model_state.pth')

而不是 state_dict()

torch.save(model.state_dict(), 'model_state.pth')

否则，一切都应该按预期工作。（我在 Colab 上测试了以下代码）

替换model.state_dict()为model.state_dict重现错误

import copy
model = TheModelClass()
torch.save(model.state_dict(), 'model_state.pth')
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
model.load_state_dict(copy.deepcopy(torch.load("model_state.pth",device)))