Reuse torch.from_numpy The original numpy The data goes to tensor type ... dtype=torch.long) h0 = encoder.initialize_hidden_state() output, ...
確定! 回上一頁