Requirements?

[Natotela]

This looks so promising, I just don't wanna promise myself something I won't be able to afford :-}

[drimeF0]

like regular stable diffusion

[Natotela]

after creating conda env, also had to

pip install opencv-python diffusers transformers
conda install av -c conda-forge

and possibly pip install accelerate as well

[szriru]

nah, I think this needs freaking big VRAM.
I tried a video which is 1024x576, 30fps, 100frames, and it said it needs 50GB VRAM lmao

torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 13.73 GiB (GPU 0; 23.99 GiB total capacity; 22.73 GiB already allocated; 0 bytes free; 27.96 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