StyleDrop-PyTorch icon indicating copy to clipboard operation
StyleDrop-PyTorch copied to clipboard

trouble running due to memory

Open ownimage opened this issue 2 years ago • 1 comments

I have a Nvidia 2060 Super 8GB ... is this enough? I see that the model download is 10GB and I am getting

torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 26.00 MiB (GPU 0; 8.00 GiB total capacity; 7.21 GiB already allocated; 0 bytes free; 7.31 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

Sorry for the noobie question , I just don't want to be chasing my tail trying to fix the impossible.

ownimage avatar Jul 15 '23 16:07 ownimage

Try to add this before training? import os os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "max_split_size_mb:128"

fypbiqi avatar Sep 14 '23 16:09 fypbiqi