What it does
Qwen Choice reads option logits from a local vision-language model to classify one image without generating an answer.
Benchmarks & research
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
Qwen Choice reads option logits from a local vision-language model to classify one image without generating an answer.
Start by gathering sample screenshots of customer messages, such as billing or technical issues. Write down one direct question and a short list of options for each. Your team can then sort incoming pictures into these categories without sending private data to an outside service.
A developer can adapt the project code to run the model directly on your own computer. Your setup will need a dedicated graphics card with compatible software. Keep in mind that this method only picks from preset choices; it will not write custom replies.