Nice work. Best 32B model(quantized to 4.7bpw) so far, more people should try.

#1
by imoc - opened

In my testing(logic/math/code/instrction follow/knowledge), most of time it is better than these model:
Qwen2.5-32B-AGI-4.7bpw-exl2
Qwen2.5-32B-Instruct-4.5bpw-exl2
Qwen2.5-32B-Instruct-abliterated-v2-exl2-4.7bpw
Rombos-LLM-V2.5-Qwen-32b-exl2-4.7bpw
Qwen2.5-Coder-32B-Instruct-EXL2-4.0bpw
Rombos-Coder-V2.5-Qwen-32b-exl2-4.7bpw

GJ :) Later I may update the quantized model.

Wow! Thank you so much! Which language did you use it in? I’d love for many people to try it out!

Sign up or log in to comment