minChatGPT
Model alignment
This project demonstrates the effectiveness of reinforcement learning from human feedback (RLHF) in improving small language models like GPT-2.
A minimum example of aligning language models with RLHF similar to ChatGPT
214 stars
5 watching
28 forks
Language: Python
last commit: about 3 years agoRelated projects:
| Repository | Description | Stars |
|---|---|---|
 rlhf-v/rlhf-v | Aligns large language models' behavior through fine-grained correctional human feedback to improve trustworthiness and accuracy. | 245 |
 pku-yuangroup/languagebind | Extending pretraining models to handle multiple modalities by aligning language and video representations | 751 |
 llava-rlhf/llava-rlhf | Aligns large multimodal models with factually enhanced reward functions to improve performance and mitigate hacking in reinforcement learning | 328 |
 openai/lm-human-preferences | Training methods and tools for fine-tuning language models using human preferences | 1,240 |
 pku-alignment/align-anything | Aligns large multimodal models with human intentions and values using various algorithms and fine-tuning methods. | 270 |
 jshilong/gpt4roi | Training and deploying large language models on computer vision tasks using region-of-interest inputs | 517 |
 tristandeleu/pytorch-maml-rl | Replication of Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks in PyTorch for reinforcement learning tasks | 830 |
 brightmart/xlnet_zh | Trains a large Chinese language model on massive data and provides a pre-trained model for downstream tasks | 230 |
 xrsrke/instructgoose | A framework for training language models using human feedback and reinforcement learning | 171 |
 minqi/learning-to-communicate-pytorch | This project implements a PyTorch-based framework for learning discrete communication protocols in multi-agent reinforcement learning environments. | 349 |
 ymcui/macbert | Improves pre-trained Chinese language models by incorporating a correction task to alleviate inconsistency issues with downstream tasks | 646 |
 sjtu-marl/malib | A framework for parallel population-based reinforcement learning | 507 |
 guopengf/auto-fedrl | A reinforcement learning-based framework for optimizing hyperparameters in distributed machine learning environments. | 15 |
 wangrongsheng/ivygpt | Develops large language models to support medical diagnoses and provide helpful suggestions | 59 |
 x-plug/cvalues | Evaluates and aligns the values of Chinese large language models with safety and responsibility standards | 481 |