inferencerlabs

/

Kimi-K2.5-MLX-3.6bit

like 15

Hi

Can you make a REAP version of this , to reduce the ram size requirements

I have a 96 gb mac m2

can you make it like this verson here , They manage to get the weight down to 80 gigabytes.

https://huggingface.co/scaryrawr/GLM-4.7-REAP-50-mlx-3Bit/tree/main

Would be amazing if you could do this on Kimi K 2.5