Is this actually usable and tested, or is it like your Devstral Small 2 quant?
#1
by positiveelevation - opened
I wanna know if your model is just electronic waste or if it can actually be used.
c.ref. https://huggingface.co/Firworks/Devstral-Small-2-24B-Instruct-2512-nvfp4/discussions/2
There's a note on that model page saying I couldn't get it running. It's been there the whole time that quant has been up. I put a note on the model cards if I have trouble getting a model running or if I had to take special steps to run it.
GLM-4.5-Air-Derestricted-nvfp4 does work with the docker command that's on my model card. I just re-tested it again with an RTX Pro 6000 Blackwell.