opendilab/LightRFT
LightRFT: Light, Efficient, Omni-modal & Reward-model Driven Reinforcement Fine-Tuning Framework
This framework helps AI practitioners improve the performance and behavior of Large Language Models (LLMs) and Vision-Language Models (VLMs). You feed in a pre-trained language or vision-language model along with human feedback or a reward model, and it outputs a fine-tuned model that better aligns with desired outcomes, like generating more accurate text or understanding multimodal data. It's designed for machine learning engineers and researchers working with advanced AI models.
208 stars. Available on PyPI.
Use this if you need an efficient and scalable way to fine-tune your LLMs or VLMs using reinforcement learning from human feedback, especially for multimodal tasks.
Not ideal if you are looking for a simple, out-of-the-box solution for basic model training without deep customization or advanced optimization.
Stars
208
Forks
10
Language
Python
License
Apache-2.0
Category
Last pushed
Mar 05, 2026
Commits (30d)
0
Dependencies
21
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/opendilab/LightRFT"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Compare
Related models
agentscope-ai/Trinity-RFT
Trinity-RFT is a general-purpose, flexible and scalable framework designed for reinforcement...
OpenRLHF/OpenRLHF
An Easy-to-use, Scalable and High-performance Agentic RL Framework based on Ray (PPO & DAPO &...
zjunlp/EasyEdit
[ACL 2024] An Easy-to-use Knowledge Editing Framework for LLMs.
huggingface/alignment-handbook
Robust recipes to align language models with human and AI preferences
hyunwoongko/nanoRLHF
nanoRLHF: from-scratch journey into how LLMs and RLHF really work.