XU-YIJIE / grpo-flatLinks
Train your grpo with zero dataset and low resources, 8bit/4bit/lora/qlora supported, multi-gpu supported ...
☆71Updated last month
Alternatives and similar repositories for grpo-flat
Users that are interested in grpo-flat are comparing it to the libraries listed below
Sorting:
- llm & rl☆134Updated last week
- 使用单个24G显卡,从0开始训练LLM☆54Updated last week
- ☆76Updated 8 months ago
- RLHF experiments on a single A100 40G GPU. Support PPO, GRPO, REINFORCE, RAFT, RLOO, ReMax, DeepSeek R1-Zero reproducing.☆62Updated 3 months ago
- ☆110Updated 11 months ago
- ☆79Updated 4 months ago
- ☆42Updated 3 months ago
- A visuailzation tool to make deep understaning and easier debugging for RLHF training.☆203Updated 3 months ago
- Code for a New Loss for Mitigating the Bias of Learning Difficulties in Generative Language Models☆62Updated 3 months ago
- verl: Volcano Engine Reinforcement Learning for LLMs☆39Updated last week
- The related works and background techniques about Openai o1☆221Updated 4 months ago
- 通义千问的DPO训练☆48Updated 8 months ago
- ☆43Updated 9 months ago
- 本项目用于大模型数学解题能力方面的数据集合成,模型训练及评测,相关文章记录。☆87Updated 8 months ago
- 怎么训练一个LLM分词器☆149Updated last year
- ☆140Updated 4 months ago
- From Llama to Deepseek, grpo/mtp implemented. With pt/sft/lora/qlora included☆27Updated last month
- Reinforcement Learning in LLM and NLP.☆36Updated 3 weeks ago
- ☆193Updated last week
- NTK scaled version of ALiBi position encoding in Transformer.☆68Updated last year
- 基于DPO算法微调语言大模型,简单好上手。☆39Updated 11 months ago
- ☆141Updated last year
- deepspeed+trainer简单高效实现多卡微调大模型☆125Updated 2 years ago
- Trinity-RFT is a general-purpose, flexible and scalable framework designed for reinforcement fine-tuning (RFT) of large language models (…☆100Updated this week
- Official code for the paper, "Stop Summation: Min-Form Credit Assignment Is All Process Reward Model Needs for Reasoning"☆120Updated this week
- pytorch分布式训练☆66Updated last year
- This is the reading list for the survey "A Survey on the Optimization of LLM-based Agents ". We will keep adding papers and improving the…☆99Updated 3 weeks ago
- ☆83Updated last month
- ☆142Updated 11 months ago
- Deepseek R1 zero tiny version own reproduce on two A100s.☆67Updated 4 months ago