how to add adaptive_max_pool2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True)'

[akashAD98]

🐛 Describe the bug

i have added attention mechanism to yolov5 repo, while doing the training I'm getting this issue , how can I solve this error?

error

i tried to train model using yolov5 command ,& i used C3CBAM attention mechanism , & I'm getting this error

!python train.py --img 640 --batch 16 --cfg /content/yolov5/models/yolov5s.yaml --epochs 250 --data coco128.yaml --weights yolov5s.pt --cache

Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
  0% 0/8 [00:00<?, ?it/s]
Traceback (most recent call last):
  File "train.py", line 637, in <module>
    main(opt)
  File "train.py", line 531, in main
    train(opt.hyp, opt, device, callbacks)
  File "train.py", line 320, in train
    scaler.scale(loss).backward()
  File "/usr/local/lib/python3.7/dist-packages/torch/_tensor.py", line 396, in backward
    torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs)
  File "/usr/local/lib/python3.7/dist-packages/torch/autograd/__init__.py", line 175, in backward
    allow_unreachable=True, accumulate_grad=True)  # Calls into the C++ engine to run the backward pass
RuntimeError: adaptive_max_pool2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True)'. You can turn off determinism just for this operation, or you can use the 'warn_only=True' option, if that's acceptable for your application. You can also file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation.
```

[soumith]

this is a bug tracker, not a Q&A forum. If you have questions, please use the PyTorch forums