to numpy() and to torch()

def to_numpy(self) -> 'Batch':
    """Change all torch.Tensor to numpy.ndarray in-place."""
    for k, v in self.items():
        if isinstance(v, torch.Tensor):
            self[k] = v.detach().cpu().numpy()
    return self

def to_torch(self, dtype : torch.dtype = torch.float32, device: str = "cpu") -> 'Batch':
    """Change all numpy.ndarray to torch.Tensor in-place."""
    for k, v in self.items():
        self[k] = torch.as_tensor(v, dtype=dtype, device=device)
    return self


below is how to use them:

batch = batch.to_torch(dtype=torch.float32, device=self.args["device"])


from :
offlinerl /neorl

标签：Tensor,dtype,self,torch,device,numpy
来源： https://www.cnblogs.com/leifzhang/p/16198435.html