←back to thread

Mercury: Ultra-fast language models based on diffusion

(arxiv.org)
566 points PaulHoule | 1 comments | 07 Jul 25 12:31 UTC | HN request time: 0.481s | source
Show context
mynti ◴[07 Jul 25 12:41 UTC] No.44489785[source]▶
is there a kind of nanogpt for diffusion language models? i would love to understand them better
replies(1): >>44490381 #
1. nvtop ◴[07 Jul 25 13:51 UTC] No.44490381[source]▶
This video has a live coding part which implements a masked diffusion generation process: https://www.youtube.com/watch?v=oot4O9wMohw