感谢开源。训练过程遇到one of the variables needed for gradient computation has been modified by an inplace operation

[chunmengcai]

File "E:\work\code\Car_recognition-master\train.py", line 521, in train(hyp, opt, device, tb_writer, wandb) File "E:\work\code\Car_recognition-master\train.py", line 291, in train loss, loss_items = compute_loss(pred, targets.to(device), model) # loss scaled by batch_size File "E:\work\code\Car_recognition-master\utils\loss.py", line 163, in compute_loss lcls += BCEcls(ps[:, 13:], t) # BCE File "E:\soft1\conda\lib\site-packages\torch\nn\modules\module.py", line 1190, in _call_impl return forward_call(*input, **kwargs) File "E:\soft1\conda\lib\site-packages\torch\nn\modules\loss.py", line 720, in forward return F.binary_cross_entropy_with_logits(input, target, File "E:\soft1\conda\lib\site-packages\torch\nn\functional.py", line 3162, in binary_cross_entropy_with_logits return torch.binary_cross_entropy_with_logits(input, target, weight, pos_weight, reduction_enum) File "E:\soft1\conda\lib\site-packages\torch\fx\traceback.py", line 57, in format_stack return traceback.format_stack() (Triggered internally at C:\actions-runner_work\pytorch\pytorch\builder\windows\pytorch\torch\csrc\autograd\python_anomaly_mode.cpp:119.) Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass 0%| | 0/24 [00:00<?, ?it/s] Traceback (most recent call last): File "E:\work\code\Car_recognition-master\train.py", line 521, in train(hyp, opt, device, tb_writer, wandb) File "E:\work\code\Car_recognition-master\train.py", line 296, in train scaler.scale(loss).backward() File "E:\soft1\conda\lib\site-packages\torch_tensor.py", line 487, in backward torch.autograd.backward( File "E:\soft1\conda\lib\site-packages\torch\autograd__init__.py", line 197, in backward Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.FloatTensor [30, 2]], which is output 0 of AsStridedBackward0, is at version 4; expected version 0 instead. Hint: the backtrace further above shows the operation that failed to compute its gradient. The variable in question was changed in there or anywhere later. Good luck!

[Jane42070]

我也是这样，找了几天解决方法了，版本也换了很多，都没有解决，希望作者可以提醒一下，遇到这种情况是什么原因，解决方法

[we0091234]

加群问问，好像这个问题很熟悉

[CopyABCs]

请问这个问题各位解决了吗，可以说一下是怎么解决的吗

[1m82]

我也遇到了，请问解决了吗

[1m82]

@we0091234 @chunmengcai