Please, extend this idea to a larger model!

#4
by CamiloMM - opened

It's clear the results here are impressive for something that could be ran on a consumer GPU from a decade ago. It's also clear this was a tech demo / proof of concept! I hope you guys put some compute on making a larger version of this! It's a great approach.

DeepSeek org

@CamiloMM Thank you very much for your feedback! We are currently planning this~

Sign up or log in to comment