taggui
taggui copied to clipboard
jhc13
[Model Request] Qwen-VL
I would love to see support for Qwen-VL added! The weights for this version are available, see below.
- https://github.com/QwenLM/Qwen-VL
- https://huggingface.co/Qwen/Qwen-VL/tree/main
- https://arxiv.org/abs/2308.12966
- https://github.com/QwenLM/Qwen-VL/blob/master/LICENSE
Is there a specific reason you want this model added? It seems to offer no advantages over newer models like InternLM-XComposer2.
Qwen seems to be a bit better with realism images than Anime images as Cog is vs Intern with my usage of them.
I found qwen to be better at realistic images than the other models, but lacking in identifying certain art. This seems to be common for most models where they excel at certain images and fail at other types. No model at the moment excels at everything.
Did you try Qwen-VL and not Qwen-VL-Plus or Qwen-VL-Max? Qwen-VL is an old model and should be pretty weak.
It was qwen-Vl that I was testing. I spend a few hours with each model before I suggest it here. If it sucks then I don’t recommend it. If it’s promising for most images, some art styles, then I make a request.