search

yananren2137 / huawei-ai-food-challenge

Apache License 2.0
11 stars 4 forks source link

RuntimeError: Invalid index in scatter at /pytorch/aten/src/TH/generic/THTensorEvenMoreMath.cpp:151 #2

Open whisperLiang opened 4 years ago

whisperLiang commented 4 years ago

我在第一个阶段训练的时候，模型到了cutmix的地方出现了这个错误，应该怎么解决呢？

yananren2137 commented 4 years ago

错误定位在cutmix里面的哪一句

whisperLiang commented 4 years ago

Traceback (most recent call last): File "train.py", line 193, in train_acc, train_loss = train(train_loader, model, criterion, optimizer, epoch, scheduler, mixup=using_cutmix) File "train.py", line 71, in train output, loss = tools.cutmix(input, target, model, criterion) File "/content/drive/Shared drives/Team-Coding-Like-Immortal/huawei-ai-food-challenge/utils/tools.py", line 95, in cutmix loss = criterion(output, target_a) lam + criterion(output, target_b) (1. - lam) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 532, in call result = self.forward(*input, **kwargs) File "/content/drive/Shared drives/Team-Coding-Like-Immortal/huawei-ai-food-challenge/utils/loss.py", line 31, in forward targets = torch.zeros(logprobs.size()).scatter(1, targets.unsqueeze(1).data.cpu(), 1) RuntimeError: Invalid index in scatter at /pytorch/aten/src/TH/generic/THTensorEvenMoreMath.cpp:151

whisperLiang commented 4 years ago

这是完整的错误提示.