microsoft / LLaVA-Med

Large Language-and-Vision Assistant for Biomedicine, built towards multimodal GPT-4 level capabilities.
Other
1.59k stars 202 forks source link

Are you going to release checkpoints? #4

Open lesterlitch opened 1 year ago

lesterlitch commented 1 year ago

You write in this repo that checkpoints are available for research use. From where can we download these please?

hellocym commented 1 year ago

Hi :) I found a repo called LLaVA-Med-preview and found a checkpoint inside it. I managed to run an inference on VQA-Rad with it, but I couldn't reproduce the result (0 epochs on VQA-Rad, with an accuracy of 0.61) in the paper. :( I think it could be something wrong in my evaluation code, or I used inappropriate parameters during inference. Hope this may help you :)

ChunyuanLI commented 1 year ago

LLaVA-Med is finally released after 5-month intensive discussions to go through Microsoft release process. Thanks for your patience.