search

TinyLLaVA / TinyLLaVA_Factory

A Framework of Small-scale Large Multimodal Models
https://arxiv.org/abs/2402.14289
Apache License 2.0
658 stars 68 forks source link

[Feature Request] qwen2.5 and llava onevision #133

Open sailfish009 opened 1 week ago

sailfish009 commented 1 week ago

Hi, we would love to see you add the latest models to take advantage of the modular structure. For example, it would be nice to be able to use models like llava-onevision-qwen2.5-7b.

https://github.com/QwenLM/Qwen2.5

ZhangXJ199 commented 5 days ago

Thank you for your suggestion, we will update it as soon as possible.

yi-ming-qian commented 1 day ago

Yes. Looking forward to the incorporations of qwen series.

ZhangXJ199 commented 1 day ago

We have added qwen2.5-0.5B and qwen2.5-1.5B to the Model Performance leaderboard. The training script can refer to qwen2.

yi-ming-qian commented 1 day ago

Thanks. Have you considered to incorporate qwen 7b models?

ZhangXJ199 commented 1 day ago

The qwen2.5-7B has higher requirements on the machine. We may try it later.