I would love to see a NVFP4 and/or FP8 quant. I've seen you provided a NVFP4 version for the Mistral Small 4. Are the necessary changes in the main branch of llm-compressor? Then I would try it on my own. ;-)
Β· Sign up or log in to comment