A high-throughput and memory-efficient inference and serving engine for LLMs

Enjoy the magic of Diffusion models!

No fortress, purely open ground. OpenManus is Coming.

Building DeepSeek R1 from Scratch

A bidirectional pipeline parallelism algorithm for computation-communication overlap in V3/R1 training.

Muon optimizer: +>30% sample efficiency with <3% wallclock overhead

Fully open reproduction of DeepSeek-R1

Solve Visual Understanding with Reinforced VLMs

pytorch distribute tutorials

A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)

本项目旨在分享大模型相关技术原理以及实战经验（大模型工程化、大模型应用落地）

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

Best practice for training LLaMA models in Megatron-LM

Stable-Hair: Real-World Hair Transfer via Diffusion Model (AAAI 2025)

《代码随想录》LeetCode 刷题攻略：200道经典题目刷题顺序，共60w字的详细图解，视频难点剖析，50余张思维导图，支持C++，Java，Python，Go，JavaScript等多语言版本，从此算法学习不再迷茫！🔥🔥 来看看，你会发现相见恨晚！🚀

Transformer Explained Visually: Learn How LLM Transformer Models Work with Interactive Visualization

Transformer seq2seq model, program that can build a language translator from parallel corpus

Transformer: PyTorch Implementation of "Attention Is All You Need"

The hardware design for AgiBot X1.

OneFlow is a deep learning framework designed to be user-friendly, scalable and efficient.

The official repo of Pai-Megatron-Patch for LLM & VLM large scale training developed by Alibaba Cloud.

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

Hunyuan-DiT : A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding

An easy-to-use LLMs quantization package with user-friendly apis, based on GPTQ algorithm.

Ongoing research training transformer models at scale

Inference code for Llama models

主要记录大语言大模型（LLMs） 算法（应用）工程师相关的知识及面试题