【报错】服务器 cuda out of memory

RuntimeError: CUDA out of memory. Tried to allocate 32.00 MiB (GPU 0; 23.64 GiB total capacity; 21.55 GiB already allocated; 17.75 MiB free; 21.60 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

RuntimeError: CUDA out of memory. Tried to allocate 2.00 GiB (GPU 0; 23.64 GiB total capacity; 17.64 GiB already allocated; 118.25 MiB free; 19.15 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF翻译这段话

运行时错误：CUDA内存不足。尝试分配 2.00 GiB 的内存（GPU 0；总容量为 23.64 GiB；已分配 17.64 GiB；可用内存为 118.25 MiB；PyTorch 已保留 19.15 GiB 总内存）。如果保留的内存远大于分配的内存，试着设置 `max_split_size_mb` 以避免内存碎片化。有关内存管理和 `PYTORCH_CUDA_ALLOC_CONF` 的更多信息，请参阅文档。
 

batch size 改小 4→2

硬件跟不上，深度学习真玩不起

无卡启用  V100

有卡启用  V100