Proof-of-concept leveraging DPO loss to fine-tune a ResNet to classify images from CIFAR10 dataset.

DPO using human Votes, Model Combat is an application that compares responses from different AI models (ChatGPT, Hanooman, and Cohere) based on user inputs. Users can vote on which model gives a better response or remark on the responses. The results are saved in Google Sheets for further analysis.

A travel agent based on qwen2 and DPO

replaced pandas-ta calls with numpy/numba functions to speed up calculating ema, tema, rsi, mfi, adx, dpo

An open source collection of tools meant to simplify the life of data protection officers (DPOs) of large entities

Re-usable & scalable RLHF training pipeline with Dagster and Modal.

dpo finetuning

EasyRLHF aims to provide an easy and minimal interface to train aligned language models, using off-the-shelf solutions and datasets

Towards Robust Alignment of Language Models: Distributionally Robustifying Direct Preference Optimization

Code for "Bridging and Modeling Correlations in Pairwise Data for Direct Preference Optimization"

Introducing Filtered Direct Preference Optimization (fDPO) that enhances language model alignment with human preferences by discarding lower-quality samples compared to those generated by the learning model

A open-source framework designed to adapt pre-trained Language Models (LLMs), such as Llama, Mistral, and Mixtral, to a wide array of domains and languages.

Learning to route instances for Human vs AI Feedback

Data and models for the paper "Configurable Safety Tuning of Language Models with Synthetic Preference Data"

Examples for using the SiLLM framework for training and running Large Language Models (LLMs) on Apple Silicon

dpo算法实现

🌾 OAT: Online AlignmenT for LLMs

[NeurIPS 2024] Official code of $\beta$-DPO: Direct Preference Optimization with Dynamic $\beta$

基于DPO算法微调语言大模型，简单好上手。

[ACL 2024] Self-Training with Direct Preference Optimization Improves Chain-of-Thought Reasoning