HN
Hacker News
Top
New
Best
Ask
Show
Jobs
Comment by janmo | Hacker News Reader
Full thread
janmo
·
We need the same with text to train Large Language Models
View on HN
simonw
·
That's pretty much what Red Pajama is:
https://simonwillison.net/2023/Apr/17/redpajama-data/
circuit10
·
https://pile.eleuther.ai/
angrais
·
It contains captions. That's the alt text. LAION is used for train LLMs.
Reply on news.ycombinator.com