RuntimeError: CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

[hgy123]

Hello！
when using the IPN-V2 network, I would like to obtain segmentation results for Largevessels and change the code from five class network to a two class network(the results of the five class network are normal). The following error occurred:
C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [832,0,0] Assertion t >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [480,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [481,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [929,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [930,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [33,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [800,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [804,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [805,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [704,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [705,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [706,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [903,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [904,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [329,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [330,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [129,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [130,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [131,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [232,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [233,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [234,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [104,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [105,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [207,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [208,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [683,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [684,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [685,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [780,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [781,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [1006,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [1007,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [555,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [556,0,0] Assertiont >= 0 && t < n_classesfailed. C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cuda\NLLLoss2d.cu:103: block: [0,0,0], thread: [6,0,0] Assertiont >= 0 && t < n_classesfailed. Traceback (most recent call last): File "D:\download\Code\3D Baselines\train.py", line 161, in <module> train_net(net=net,device=device) File "D:\download\Code\3D Baselines\train.py", line 68, in train_net loss = Loss_CE(pred, train_annotations)+ 0.6 * Loss_DSC(pred, train_annotations) File "C:\Users\hgyy\AppData\Roaming\Python\Python39\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "D:\download\Code\3D Baselines\utils.py", line 226, in forward target=make_one_hot(target.long(),shape) File "D:\download\Code\3D Baselines\utils.py", line 160, in make_one_hot result.scatter_(1, input.cpu(), 1) RuntimeError: CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile withTORCH_USE_CUDA_DSA to enable device-side assertions.
My environment: python=3.9.16 torch=2.0.0 torchvision=0.15.1
Where did I make a mistake?