other versions

#1
by michel-io - opened

works great but slow on limited vram. can you make fp8 versions? and fp16 versions for older cards?

https://huggingface.co/mradermacher/Qwen3-4b-Z-Image-Turbo-AbliteratedV1-GGUF --static

https://huggingface.co/mradermacher/Qwen3-4b-Z-Image-Turbo-AbliteratedV1-i1-GGUF --imatrix
@mradermacher beat me making quants!

works great but slow on limited vram. can you make fp8 versions? and fp16 versions for older cards?
Thank you! I will upload some GGUF format Q2-Q8. thank you!

BennyDaBall changed discussion status to closed

Sign up or log in to comment