TL;DR: torch.flatten() ... For v0.4.1 and earlier, use t.reshape(-1) . With t.reshape(-1) : If the requested view is contiguous in memory this will equivalent to ...
確定! 回上一頁