odd_errors(持续更新，欢迎补充，工程中遇到的错误)

zhe_csdn

于 2019-06-26 13:42:44 发布

阅读量1.4k

点赞数

CC 4.0 BY-SA版权

分类专栏： torch python

本文链接：https://blog.youkuaiyun.com/zhe_csdn/article/details/93733941

python 同时被 2 个专栏收录

36 篇文章

订阅专栏

torch

18 篇文章

订阅专栏

本文探讨了在使用PyTorch进行深度学习时常见的运行时错误，包括CUDA内存溢出、非法内存访问及状态字典加载失败等问题，并提供了相应的解决策略。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

                    
                        
                    
                    TypeError: ‘module’ object is not subscriptable
RuntimeError: merge_sort: failed to synchronize: an illegal memory access was encountered
RuntimeError: CUDA out of memory. Tried to allocate 322.00 MiB (GPU 0; 10.91 GiB total capacity; 8.62 GiB already allocated; 278.00 MiB free; 766.86 MiB cached)IndexError: invalid index of a 0-dim tensor. Use tensor.item() to convert a 0-dim tensor to a Python number .data[0]->.item()
TypeError: can’t convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.
predict.data.numpy()->predict.data.cpu().numpy()
RuntimeError: Error(s) in loading state_dict for CrowdCounter: Missing key(s) in state_dict: "CCN.de_pred.0.conv.bias
 net.load_state_dict(torch.load(model_path),strict=False) add strict=False
THCudaCheck FAIL file=/opt/conda/conda-bld/pytorch_1525796793591/work/torch/lib/THC/generic/THCStorage.c line=82 error=2 : out of memory ->cuda memory used off

                