Draft model with ik_llama.cpp

#1
by Chico70 - opened

I just want to thank you for this great draft model! I have test it with ik_llama.cpp and it works great!!!

Hi, could you please build the draft model for Qwen/Qwen3-Next-80B-A3B-Instruct?
Thank you very much!

Hi, could you please build the draft model for Qwen/Qwen3-Next-80B-A3B-Instruct?
Thank you very much!

Sadly it won't work as the effectiveness of using a draft model relies on the active parameters being much more for the main model (and Qwen3-Next-80B-A3B-Instruct only has 3B active).

Sign up or log in to comment