RuntimeError: CUDA out of memory
#14
by
Raziel474
- opened
Is there a way to work around this issue?
RuntimeError: CUDA out of memory. Tried to allocate 1024.00 MiB (GPU 0; 6.00 GiB total capacity; 4.54 GiB already allocated; 0 bytes free; 4.66 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
Having the same error with an RTX 3060 GPU. I managed to get it working with version 3 of the model but only at 256x256 resolution. Even with 512x512, it ran out of memory.