please use torch.load with map_location=torch.device(‘cpu‘) to map your storages to the CPU.

please use torch.load with map_location=torch.device(‘cpu’) to map your storages to the CPU.

问题描述：

在GPU环境下，用pytorch框架训练好的网络模型，在CPU环境测试，报错如下：

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=torch.device('cpu') to map your storages to the CPU.

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=torch.device(‘cpu’) to map your storages to the CPU.

可定位到报错行的代码是模型加载部分：

torch.load('xxxNet.pth')

其中，‘xxxNet.pth’ 是用GPU训练好的模型权重文件。

---

解决方案：

在模型加载时，指定 map_location='cpu' 即可，代码如下：

torch.load('xxxNet.pth', map_location='cpu')