大佬您好，训练sha时报错

[xuebaobizui1]

bregman_pytorch.py: 173:UserWarning: An output with one or more elements was resized since it had shape [1024], which does not match the required output shape [1, 1024].This behavior is deprecated, and in a future PyTorch release outputs will not be resized unless they have zero elements. You can explicitly reuse an out tensor t by resizing it, inplace, to zero elements with t.resize_(0). (Triggered internally at ../aten/src/ATen/native/Resize.cpp:23.)

我使用您的原版代码训练时只在开头报了一次这个错误，之后每个epoch都是正常显示的。
但是当我改动了一些网络结构之后再进行训练之后每个epoch都会报这个错误。我从训练日志看到 loss 确实在不断降低，我想知道这是否对于训练结果有影响。我现在还在训练中，没法进行test。当我训练完成后会对本问题的描述进行补充。
感谢！！！

[wfs123456]

应该不会有影响，，就是一个张量，一个size为1xn(二维的)，一个size为n(一维的)

…

---原始邮件--- 发件人: ***@***.***> 发送时间: 2023年3月22日(周三) 晚上6:16 收件人: ***@***.***>; 抄送: ***@***.***>; 主题: [wfs123456/CCTrans] 大佬您好，训练sha时报错 (Issue #27) bregman_pytorch.py: 173:UserWarning: An output with one or more elements was resized since it had shape [1024], which does not match the required output shape [1, 1024].This behavior is deprecated, and in a future PyTorch release outputs will not be resized unless they have zero elements. You can explicitly reuse an out tensor t by resizing it, inplace, to zero elements with t.resize_(0). (Triggered internally at ../aten/src/ATen/native/Resize.cpp:23.) 我使用您的原版代码训练时只在开头报了一次这个错误，之后每个epoch都是正常显示的。 但是当我改动了一些网络结构之后再进行训练之后每个epoch都会报这个错误。我从训练日志看到 loss 确实在不断降低，我想知道这是否对于训练结果有影响。我现在还在训练中，没法进行test。当我训练完成后会对本问题的描述进行补充。 感谢！！！ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>