Flux3 Klein Distilled GGUF Conversion.
Loader nodes: https://github.com/molbal/ComfyUI-GGUF / https://registry.comfy.org/publishers/molbal/nodes/comfyui-gguf-reboot
Description
FAQ
Comments (6)
Got a massive speed boost with Qwen2.1 after updating the GGUF loader. Is that expected?
btw, could you add support for MOSS-Image gguf to the loader?
https://huggingface.co/OpenMOSS-Team/MOSS-Image-SFT
https://huggingface.co/Comfy-Org/MOSS-Image
It’s a promising new model that seems to be gaining traction.
Yes, last week I pushed an update that improved performance especially with _K quant type GGUFs :)
I'm adding MOSS soon too!
(I need perhaps 1-2 more days with MOFF but I'm on it)
@molbal tyvm man, and just saw post on reddit, Moss GGUF can get speedup like qwen2..1
@dolores332 here it's on hugging face, I made also a turbo gguf. https://huggingface.co/molbal/MOSS-Image-GGUF
I'll upload to civitai tomorrow
@molbal Noice, thx for the turbo merge.