Running 1776 with draft model
#1
by
ernestr
- opened
Thanks for the quant @matatonic ! I'm currently running it with Llama3.2 as a draft model. It's working well but I'm curious if using a draft model degrades the thinking.
ernestr
changed discussion status to
closed
It doesn't degrade the thinking, I use the 1B 3.2 as draft also, worst case it doesn't make it faster.