Mirasol3B

From Humanoid Robots Wiki
Jump to: navigation, search

Mirasol3B is an autoregressive multimodal model for time-aligned video and audio.

Mirasol3B
NameMirasol3B
Full NameMirasol3B: A Multimodal Autoregressive Model for Time-Aligned and Contextual Modalities
ArxivLink
TwitterTwitter
Publication DateFebruary 2024
AuthorsAJ Piergiovanni, Isaac Noble, Dahun Kim, Michael Ryoo, Victor Gomes, Anelia Angelova