The llava-next-video-34b DPO model is not performing well, whereas the 7B-dpo model works fine.
I've reviewed related issues and tried _changing the conv mode to mistraldirect, but the responses still seem off. I also updated IMAGE_TOKEN_INDEX to 64002, but the performance issues persist.
The llava-next-video-34b DPO model is not performing well, whereas the 7B-dpo model works fine.
I've reviewed related issues and tried _changing the conv mode to mistraldirect, but the responses still seem off. I also updated
IMAGE_TOKEN_INDEX
to 64002, but the performance issues persist.Could you advise on how to resolve this?