BrushNet
BrushNet copied to clipboard
CUDA out of memory.How much memory is needed for training?(训练时需要多大的显存?)
Here is the complete content
CUDA out of memory. Tried to allocate 1024.00 MiB (GPU 0; 23.59 GiB total capacity; 20.87 GiB already allocated; 624.12 MiB free; 21.77 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
我用的sd1.5训练,bs=1的时候占用显存17G左右。