These are the FluxBooru models created by someone else and converted to GGUF Q_8, reach out to me if you'd like to see other quantized versions made available. You may also visit the huggingface discussion for more insight as to how they were converted including the scripts.

• I didn't create these models, I just made it compatible with Unet Loader node

• Thank you goes to terminusresearch and ptx0 for their respective versions

• Recommend steps is 20 to 25 according to the author

• Recommended CFG is 5 to 6 according to the author however I'm observing results also look fine at cfg 3.5 for all versions

• Recommended Workflow in ComfyUI is

  • Load UNET GGUF > Power Lora Loader > ModelSamplingFlux > KSampler Advanced Efficient > VAE Decode > Save Image

ptx0's FluxBooru original models (Providing v0.2 and v0.3):

https://civarchive.com/models/859032

terminusresearch's FluxBooru CFG 3.5 model:
https://huggingface.co/terminusresearch/flux-booru-CFG3.5/tree/main/transformer

Discussion regarding converting the Diffuser's based format to GGUF: https://huggingface.co/terminusresearch/flux-booru-CFG3.5/discussions/1