1 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.
1 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.在代码中加入os.environ['CUDA_LAUNCH_BLOCKING'] = '1'可以将错误的具体位置显示出来。以上的问题大多是网络中的labe