【pytorch】cpu与gpu load时相互转化 torch.load(map_location=)

问题

gpu训练的模型，使用cpu测试时遇到：

RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location='cpu' to map your storages to the CPU.

解决

此时改为：

torch.load("0.9472_0048.weights",map_location='cpu')

就可以解决问题了。

方便查阅，整理：

假设我们只保存了模型的参数(model.state_dict())到文件名为modelparameters.pth, model = Net()

1. cpu -> cpu或者gpu -> gpu:

checkpoint = torch.load('modelparameters.pth')

model.load_state_dict(checkpoint)

2. cpu -> gpu 1

torch.load('modelparameters.pth', map_location=lambda storage, loc: storage.cuda(1))

3. gpu 1 -> gpu 0

torch.load('modelparameters.pth', map_location={'cuda:1':'cuda:0'})

4. gpu -> cpu

torch.load('modelparameters.pth', map_location=lambda storage, loc: storage)