the qwen QVQ-72B-Preview new model, is a vision model published today.
see Qwen/QVQ-72B-Preview · Hugging Face
4 Likes
also hope support soon
@NB2025 and @selenium39 can you provide the business/scholastic benefits this specific model would provide you ? I have opened a roadmap request but this does not guarantee we would adopt this model.
-Coby
1 Like
I would add it as an alternative for the Llama 3.2 90b vision model.
With multiple models, the precision of the image recognition can be enhanced.
I just started evaluating the Llama 3.2 vision, but it does not always do what I would like to use it for.
If the QvQ does it easier, I would choose it. Another use case is a fallback mechanism to resolve false positive and false negative recognitions.
1 Like