Closed syspider closed 3 months ago
Thank you for the great job! I am just a bit curious about the computing resources used for LLaVA-NeXT.
The largest 34B variant finishes training in ~1 day with 32 A100s according to the blog post
Thank you for the great job! I am just a bit curious about the computing resources used for LLaVA-NeXT.