RuntimeError: handle_0 INTERNAL ASSERT FAILED at “/opt/conda/conda-bld/pytorch_1695392036766/work/c1

今天我的pytorch报了一个神奇的错误

RuntimeError: handle_0 INTERNAL ASSERT FAILED at "/opt/conda/conda-bld/pytorch_1695392036766/work/c10/cuda/driver_api.cpp":15, please report a bug to PyTorch

查了半天没找到什么解决方法,然后我观察了一下显存

用了9G,因此有可能时反向传播的时候g了
RuntimeError: handle_0 INTERNAL ASSERT FAILED at “/opt/conda/conda-bld/pytorch_1695392036766/work/c1_第1张图片
调小batch了之后就能用了

总结:调小batch能解决这个问题

你可能感兴趣的:(深度学习,conda,pytorch,人工智能)