search

luogen1996 / LaVIN

[NeurIPS 2023] Official implementations of "Cheap and Quick: Efficient Vision-Language Instruction Tuning for Large Language Models"
502 stars 36 forks source link

LaVin-7B model on task Multimodal ChatBot #5

Closed Dicomsky closed 1 year ago

Dicomsky commented 1 year ago

Would you mind releasing LaVin-7B model on task Multimodal ChatBot? I'm interested in your work but I have 4 3090Ti and no enough resourse to fully train a LaVin-7B model, and lavin-13B is too large.

luogen1996 commented 1 year ago

We haven't experimented with the multimodal chatbot on the 7B model. However, we are trying to accelerate the training and save GPU memory, looking forward our updates.

Dicomsky commented 1 year ago

Thank you for your reply!

18445864529 commented 11 months ago

Any update on this? the GPU requirement 55G for 13B chatbot is not affordable on many GPU types. Would be glad if you could release a 7B chatbot checkpoint or reduce the memory usage to under 40G, for the chatbot experiment. Thanks.