mradermacher

/

model_requests

like 191

I guess that's a no?

This model requires conversion before it can be quanted.

hey, it requires manual handling so might take a bit of time, unless someone can provide a link for llamacpp compatible version so I can queue it like a normal model

https://huggingface.co/sahilchachra/Leanstral-1.5-119B-A6B-BF16 may work as it has been converted to Huggingface format.

fix the config to comply with standard please:
sahilchachra/Leanstral-1.5-119B-A6B-BF16: no architectures entry ()

@sahilchachra can you do this please?

I have updated the config and added - "architectures": ["Mistral3ForConditionalGeneration"].

I don't have access to the GPU which I was using to work on this model's conversion/quantisation. Let me see if I can arrange one and get this done in a few days.

Thanks!