10:31pm PDT - July 7th, 2023

RT @roadjiang: We learn a vector quantizer to map an image, or non-linguisitic modalities, to the token space of a frozen LLM. This transl…

0 retweets   4 likes  - Kangying (@timcanby)