An Efficient and User-Friendly Scaling Library for Reinforcement Learning with Large Language Models

Freeing data processing from scripting madness by providing a set of platform-agnostic customizable pipeline processing blocks.

SWE-bench: Can Language Models Resolve Real-world Github Issues?

The LLM Evaluation Framework

ResiDual: Transformer with Dual Residual Connections, https://arxiv.org/abs/2304.14802

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

🐳 Efficient Triton implementations for "Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention"

Efficient Mixture of Experts for LLM Paper List

Awesome LLM Books: Curated list of books on Large Language Models

High-speed Large Language Model Serving for Local Deployment

GPU operators for sparse tensor operations

Benchmarking Benchmark Leakage in Large Language Models

A resource repository for machine unlearning in large language models

LLM Unlearning

An Extensible Deep Learning Library

2025年12月更新，目前国内可用Docker镜像源汇总，DockerHub国内镜像加速列表，🚀DockerHub镜像加速器

FlashMLA: Efficient Multi-head Latent Attention Kernels

Minimal reproduction of DeepSeek R1-Zero

Fully open data curation for reasoning models

交易模块

MiniCPM-V 4.5: A GPT-4o Level MLLM for Single Image, Multi Image and High-FPS Video Understanding on Your Phone

🚀 Efficient implementations of state-of-the-art linear attention models

A Telegram bot to recommend arXiv papers

🚀🚀 「大模型」2小时完全从0训练26M的小参数GPT！🌏 Train a 26M-parameter GPT from scratch in just 2h!

Allegro is a powerful text-to-video model that generates high-quality videos up to 6 seconds at 15 FPS and 720p resolution from simple text input.

O1 Replication Journey

RLHF implementation details of OAI's 2019 codebase