Qwen-Applications/STAR

STAR: Similarity-guided Teacher-Assisted Refinement for Super-Tiny Function Calling Models

28
/ 100
Experimental

This project helps create very small, efficient AI models that can understand and execute specific commands or "function calls" from users. It takes a large, capable AI model and a collection of example interactions, then distills that knowledge into a much smaller, faster model. The end result is a compact AI that can accurately interpret user requests to perform actions, ideal for deployment on devices with limited resources or for scenarios where quick, specialized responses are needed. This is for AI developers and researchers who need to deploy function-calling AI at a tiny scale.

Use this if you need to build a highly compact, cost-effective AI model that can accurately interpret and respond to user requests by calling specific functions or tools, while maintaining performance close to much larger models.

Not ideal if you are looking for a general-purpose large language model for broad conversational tasks rather than specialized function calling.

AI-model-compression function-calling on-device-AI efficient-AI language-model-distillation
No Package No Dependents
Maintenance 10 / 25
Adoption 7 / 25
Maturity 11 / 25
Community 0 / 25

How are scores calculated?

Stars

39

Forks

Language

Python

License

Apache-2.0

Last pushed

Feb 12, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/Qwen-Applications/STAR"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.