NJUxlj/Travel-Agent-based-on-Qwen2-RLHF

A travel agent based on Qwen2.5, fine-tuned by SFT + DPO/PPO/GRPO using traveling question-answer dataset, a mindmap can be output using the response. A RAG system is build upon the tuned qwen2, using Prompt-Template + Tool-Use + Chroma embedding database + LangChain

31
/ 100
Emerging

Implements multiple RAG strategies (naive, Self-RAG with adaptive retrieval tokens, and MemWalker with hierarchical memory trees and backtracking) selectable via dispatcher, plus a multi-agent tool system integrating Google Search, weather, hotel booking, and flight APIs with LLM-driven function calling. Supports end-to-end RLHF training across SFT, DPO, PPO, and GRPO with LiteLLM/Hugging Face trainers, and generates visual mindmaps from travel itineraries via Gradio frontend.

No License No Package No Dependents
Maintenance 6 / 25
Adoption 8 / 25
Maturity 8 / 25
Community 9 / 25

How are scores calculated?

Stars

61

Forks

5

Language

Python

License

Last pushed

Nov 14, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/rag/NJUxlj/Travel-Agent-based-on-Qwen2-RLHF"

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