Go to OpenArt main site
Upload workflow

FLUX GGUF Q8 12GB

5.0

0 reviews
17
9.7K
3.4K
3
Description

A very warm welcome to the Future and the GGUF era in ComfyUI on 12GB of VRAM.

  • - Please update ComfyUI. From August the 15th 2024 a new GUI is here. Polished and refined. Takes some using to, but worth it.
  • Move your old workflows if you wish from any directory to: ComfyUI/user/default/workflows
  • - in ComfyUI Manager look for and install “ComfyUI-GGUF” or if you wish you can manually install from: https://github.com/city96/ComfyUI-GGUF
  • - download the models and place the .gguf model files in your “ComfyUI/models/unet” folder [flux1-dev GGUF] from https://huggingface.co/city96/FLUX.1-dev-gguf [flux1-schnell GGUF] from https://huggingface.co/city96/FLUX.1-schnell-gguf
  • - Restart ComfyUI and upload the workflow.
  • - use the same Flux clips you already have and the same Flux vae.
  • Best results are with Euler – Beta - 25 to 35 seconds per generation using the Schnell with 4 steps.

Enjoy.

Discussion

(No comments yet)

Loading...

Author

19
19.9K
190
68.7K

No reviews yet

  • - latest (a month ago)

Primitive Nodes (2)

FluxGuidance (1)

UnetLoaderGGUF (1)

Custom Nodes (8)

ComfyUI

  • - CLIPTextEncode (2)

  • - VAEDecode (1)

  • - SaveImage (1)

  • - KSampler (1)

  • - EmptyLatentImage (1)

  • - DualCLIPLoader (1)

  • - VAELoader (1)

Checkpoints (0)

LoRAs (0)