DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

DeepEP: an efficient expert-parallel communication library

Differentiable fast wavelet transforms in PyTorch with GPU support.

Numerical integration in arbitrary dimensions on the GPU using PyTorch / TF / JAX

Minimal reproduction of DeepSeek R1-Zero

FlashInfer: Kernel Library for LLM Serving

Fully open reproduction of DeepSeek-R1

My learning notes/codes for ML SYS.

Official PyTorch implementation of "FlatQuant: Flatness Matters for LLM Quantization"

[NeurIPS 2024 Oral🔥] DuQuant: Distributing Outliers via Dual Transformation Makes Stronger Quantized LLMs.

Acode - powerful text/code editor for android

📚A curated list of Awesome LLM/VLM Inference Papers with codes: WINT8/4, FlashAttention, PagedAttention, Parallelism, MLA, etc.

Materials for learning SGLang

DashInfer is a native LLM inference engine aiming to deliver industry-leading performance atop various hardware architectures, including CUDA, x86 and ARMv9.

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

LeetCode 101：力扣刷题指南

Tile primitives for speedy kernels

Fastest kernels written from scratch

SGLang is a fast serving framework for large language models and vision language models.

This project is about convolution operator optimization on GPU, include GEMM based (Implicit GEMM) convolution.

Several optimization methods of half-precision general matrix multiplication (HGEMM) using tensor core with WMMA API and MMA PTX instruction.

Hackable and optimized Transformers building blocks, supporting a composable construction.

A Easy-to-understand TensorOp Matmul Tutorial

Yinghan's Code Sample

Fast CUDA matrix multiplication from scratch

Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.