LLaMA-TRL: Fine-tuning LLaMA with PPO and LoRA

Notus is a collection of fine-tuned LLMs using SFT, DPO, SFT+DPO, and/or any other RLHF techniques, while always keeping a data-first approach

An implementation of GRPO for Unsloth's VLMs training

基于DPO算法微调语言大模型，简单好上手。

Various training, inference and validation code and results related to Open LLM's that were pretrained (full or partially) on the Dutch language.

realize the reinforcement learning training for gpt2 llama bloom and so on llm model

simpleR1: A Simple Framework for Training R1-like Models

使用trl、peft、transformers等库，实现对huggingface上模型的微调。

This project demonstrates the process of fine-tuning the Qwen2.5-3B-Instruct model using GRPO (Generalized Reward Policy Optimization) on the GSM8K dataset.

Supervised Fine tuning using TRL library

Different post-training techniques for LLMs, including: SFT, DPO and Online RL

This project demonstrates the process of fine-tuning the Qwen2.5-3B-Instruct model using GRPO (Generalized Reward Policy Optimization) on the GSM8K dataset.

Notebooks to create an instruction following version of Microsoft's Phi 2 LLM with Supervised Fine Tuning and Direct Preference Optimization (DPO)

Direct Preference Optimization of ChatGPT2 using TRL Library

ODM: TRL to BAL rules automated translation

Reinforcement Fine-Tuning LLMs With GRPO

The overall aim of this project is to create a term rewriting system that could be useful in everyday programming, and to represent data in a way that roughly correspond to the definition of a term in formal logic. Terms should be familiar to any programmer because they are basically constants, variables, and function symbols.

Notebooks to create an instruction following version of Microsoft's Phi 1.5 LLM with Supervised Fine Tuning and Direct Preference Optimization (DPO)

Minimal code to train reasoning model with reinforcement learning.

Aligning FLAN-T5 with Reinforcement Learning from Human Feedback (RLHF) for Neutral, Grammatically Correct News Summaries