munonex comments

Repositories
Issues
Comments

Results 1 comments of


                                            munonex

manual cast: torch.bfloat16 when using fp8 combined flux.dev models causing vram issues with LoRAs

`python main.py --gpu-only` is a good workaround for my case. I encountered the same error and ComfyUI crashed ``` model weight dtype torch.float8_e4m3fn, manual cast: torch.bfloat16 model_type FLUX fish: Job...