Sign In

SPEED_Q8

41
537
8
Updated: Aug 22, 2024
base model
Type
Checkpoint Trained
Stats
267
Reviews
Published
Aug 19, 2024
Base Model
Flux.1 S
Hash
AutoV2
F27EEC723F
License:

I chose the best from each kind

  1. the best in small models (Q2_K)

  2. in the middle (Q4_K_M)

  3. the most close to original model is (Q8)

    its up to you

I will be happy to make any quantization request for this merged version
DONE

  • For optimal results, we recommend trying this advanced workflow:

https://civitai.com/models/658101/flux-advance

basic

https://civitai.com/models/652981/gguf-workflow-simple

just download this and install missing nodes from manager

  • for t5 gguf

https://civitai.com/models/668417/t5gguf

  • what is the best of (4th gguf quantization)?

Key Features:

Merges the strengths of Flux1-dev and Flux1-schnell

big thanks for https://huggingface.co/city96 who start GGUF journy

if you face this error during loading gguf loader

 newbyteorder was removed from the ndarray class in NumPy 2.0.

pip install numpy==1.26.4

Works on lower-end GPUs (tested on 12GB GPU with t5 fp16)

High-quality output comparable to more resource-intensive models