Few shot prompting

[FaisalAldarees]

Hello guys 👋

I'm exploring few-shot prompting for the Qwen QVQ model and was wondering.

• Does the model support few-shot learning, meaning can it accept multiple images as input for context?

• If not, are there any workarounds, such as using a custom inference setup that makes multiple LLM calls (one for each shot) and consolidates the results?

Would love to hear any insights or experiences.

Thanks in advance! 🙌

[AbdulrahmanAlGhofaily]

Facing the same issue here. A few-shots including multiple image would be a great addition, where the model learns what it should look for with image examples!