[已解決] RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead.
問題描述
在使用 PyTorch 進行深度學習模型的建設時,我們免不了一次又一次地調整神經層與輸入輸出的形狀,這顯然是每位 AI 工程師必經的道路 —— 而在 PyTorch 的形狀變換 view()
方法中,顯然存在一個有趣的小陷阱:
RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead.Read More »[已解決] RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead.