Tensors of the same index must be on the same device and the same dtype except `step` tensors that can be CPU and float32 notwithstanding

避免使用 torch.set_default_dtype(torch.float64)

 

可以尝试采用

model. Double
或者
model.to(torch. Double)

m = torch().to(device).to(torch.float64)

 

 

 

 

 

参考：

Tensors of the same index must be on the same device and the same dtype except `step` tensors that can be CPU and float32 notwithstanding - PyTorch Forums

PyTorch: Tensors of the same index must be on the same device and the same dtype except `step` tenso_runtimeerror: tensors of the same index must be on-CSDN博客