A common PyTorch convention is to save models using either a .pt or .pth file ... you will want to run inference on either GPU or CPU.
確定! 回上一頁