Skip to content

mlfoundations/dcvlm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 

Repository files navigation

DataComp-VLM

DataComp-VLM is a competition to curate the best training dataset to (pre)train Vision-Language Models. We release a standardized pool of up to 6T multimodal tokens, and participants can experiment with filtering, mixing, and formatting samples to create the best data. The resulting models are evaluated on a suite of upto 52 multimodal tasks.

Release Timeline

Everything is coming soon! 🤩 We expect to finalize the release in the next ~2/3 weeks.

About

No description, website, or topics provided.

Resources

Stars

44 stars

Watchers

6 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors