muupan

SPPO

The official implementation of Self-Play Preference Optimization (SPPO)

CPO_SIMPO

This repository contains the joint use of CPO and SimPO method for better reference-free preference learning methods.

EasyLM

Large language models (LLMs) made easy, EasyLM is a one stop solution for pre-training, finetuning, evaluating and serving LLMs in JAX/Flax.

Online-RLHF

A recipe for online RLHF.

CameraController

📷 Control USB Cameras from an app

alpaca_eval

An automatic evaluator for instruction-following language models. Human-validated, high-quality, cheap, and fast.

audiocraft

Audiocraft is a library for audio processing and generation with deep learning. It features the state-of-the-art EnCodec audio compressor / tokenizer, along with MusicGen, a simple and controllable music generation LM with textual and melodic conditioning.

Gymnasium

An API standard for single-agent reinforcement learning environments, with popular reference environments and related utilities (formerly Gym)

NeMo-Aligner

Scalable toolkit for efficient model alignment

alignment-handbook

Robust recipes to align language models with human and AI preferences

mathematics_dataset

This dataset code generates mathematical question and answer pairs, from a range of question types at roughly school-level difficulty.

Vim

:star: Vim for Visual Studio Code

open-instruct

llm-japanese-dataset

LLM構築用の日本語チャットデータセット

safe-rlhf

Safe RLHF: Constrained Value Alignment via Safe Reinforcement Learning from Human Feedback

LLaMA-Factory

A WebUI for Efficient Fine-Tuning of 100+ LLMs (ACL 2024)

flash-attention

Fast and memory-efficient exact attention

vscode-journal

Lightweight journal and simple notes support for Visual Studio Code

evals

instruction_ja

Japanese instruction data (日本語指示データ)

unsloth

Finetune Llama 3.1, Mistral, Phi & Gemma LLMs 2-5x faster with 80% less memory

awesome-japanese-nlp-resources

A curated list of resources dedicated to Python libraries, LLMs, dictionaries, and corpora of NLP for Japanese

japanese-llm-ranking

direct-preference-optimization

Reference implementation for DPO (Direct Preference Optimization)

Karabiner-Elements

Karabiner-Elements is a powerful utility for keyboard customization on macOS Sierra (10.12) or later.

hh-rlhf

Human preference data for "Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback"

llm-numbers

Numbers every LLM developer should know

YouTube-Blocker

A Chrome Extension that blocks non-educational YouTube videos

big-list-of-naughty-strings

The Big List of Naughty Strings is a list of strings which have a high probability of causing issues when used as user-input data.

vimGPT

Browse the web with GPT-4V and Vimium