Post by Tim Dettmers on X
Tim Dettmers@Tim_Dettmers
XBeating DeepSeek-V3 with a 405B Llama base is not easy -- solid post-training goes a long way. The nice thing is that it is fully open-source, so anyone can use this recipe for their base models.
287 likes3 repliesPosted Jan 30, 2025