Superman
Superman
the problem disappeared when restrictImageTypes, but in this way, I cannot watch tensor
👍hope you can fix soon
type = torch.tensor dtype = torch.float32 shape = (16, 3, 96, 96) I use that in debug pannel 
不是，我都是尽量装的最新，pytorch用的2.0.1，直接用作者的模型推理就没事，自己训练的就会出现这个问题