FareedKhan-dev/multi-agent-training-grpo
A multi-agent system trained with GRPO for reliable long-horizon task planning and execution.
Uses QLoRA and PEFT for parameter-efficient fine-tuning of the policy model, with group-based trajectory evaluation and relative advantage computation to suppress hallucinations in multi-agent planning. Trains on DeepMath-103K and Natural Questions datasets with GPT-4o-based reward modeling, implementing a complete GRPO loop that compares multiple rollouts per query to reinforce successful planning strategies over group averages.
Stars
22
Forks
6
Language
Jupyter Notebook
License
MIT
Category
Last pushed
Feb 09, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/rag/FareedKhan-dev/multi-agent-training-grpo"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
deepsense-ai/ragbits
Building blocks for rapid development of GenAI applications
infiniflow/ragflow
RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses...
GiovanniPasq/agentic-rag-for-dummies
A modular Agentic RAG built with LangGraph — learn Retrieval-Augmented Generation Agents in minutes.
truefoundry/cognita
RAG (Retrieval Augmented Generation) Framework for building modular, open source applications...
NVIDIA/context-aware-rag
Context-Aware RAG library for Knowledge Graph ingestion and retrieval functions.