RuntimeError: CUDA out of memory. Tried to allocate 326.00 MiB (GPU 0; 23.70 GiB total cap;tmux分屏操作

RuntimeError: CUDA out of memory. Tried to allocate 326.00 MiB (GPU 0; 23.70 GiB total capacity; 19.91 GiB already allocated; 140.56 MiB free; 21.20 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation

 RuntimeError: CUDA out of memory. Tried to allocate 326.00 MiB (GPU 0; 23.70 GiB total cap;tmux分屏操作_第1张图片

RuntimeError: CUDA out of memory. Tried to allocate 326.00 MiB (GPU 0; 23.70 GiB total cap;tmux分屏操作_第2张图片 RuntimeError: CUDA out of memory. Tried to allocate 326.00 MiB (GPU 0; 23.70 GiB total cap;tmux分屏操作_第3张图片

 vim ~/.bashrc

 export TORCH_CUDA_ALLOC_CONF="max_split_size_mb=512"

 

tmux分屏操作:

https://www.cnblogs.com/longbigbeard/p/9513491.html 

RuntimeError: CUDA out of memory. Tried to allocate 326.00 MiB (GPU 0; 23.70 GiB total cap;tmux分屏操作_第4张图片 

 

你可能感兴趣的:(人工智能)