The simplest, fastest repository for training/finetuning medium-sized GPTs.

Development repository for the Triton-Linalg conversion

Robust recipes to align language models with human and AI preferences

Synthetic data generation for tabular data

A collection of large question answering datasets

Implementation of the training framework proposed in Self-Rewarding Language Model, from MetaAI

The official implementation of Self-Play Fine-Tuning (SPIN)

The RedPajama-Data repository contains code for preparing large datasets for training large language models.

FlagScale is a large model toolkit based on open-sourced projects.

Reference implementation for DPO (Direct Preference Optimization)

MAD: The first work to explore Multi-Agent Debate with Large Language Models :D

AISystem 主要是指AI系统，包括AI芯片、AI编译器、AI推理和训练框架等AI全栈底层技术

Effective Data Augmentation With Diffusion Models

An interactive NVIDIA-GPU process viewer and beyond, the one-stop solution for GPU process management.

website

CoreDNS is a DNS server that chains plugins

Unified Efficient Fine-Tuning of 100+ LLMs (ACL 2024)

real Transformer TeraFLOPS on various GPUs

Inference code for Llama models

Scripts for fine-tuning Meta Llama with composable FSDP & PEFT methods to cover single/multi-node GPUs. Supports default & custom datasets for applications such as summarization and Q&A. Supporting…

中文LLaMA-2 & Alpaca-2大模型二期项目 + 64K超长上下文模型 (Chinese LLaMA-2 & Alpaca-2 LLMs with 64K long context models)

中文LLaMA&Alpaca大语言模型+本地CPU/GPU训练部署 (Chinese LLaMA & Alpaca LLMs)

中文语言理解测评基准 Chinese Language Understanding Evaluation Benchmark: datasets, baselines, pre-trained models, corpus and leaderboard

⏳🛡 Pythonic universal errors-and-erasures Reed-Solomon codec to protect your data from errors and bitrot. Includes a future-proof zero-dependencies pure-python implementation 🔮 and an optional spee…

An open-source library for Python 3 providing tools for analysis and simulation of analog and digital communication systems.

A Convolutional Encoder and Viterbi Decoder in C++.

KAPAO is an efficient single-stage human pose estimation model that detects keypoints and poses as objects and fuses the detections to predict human poses.

Gaze estimation using MPIIGaze and MPIIFaceGaze