Fast low-bit matmul kernels in Triton

This repository delivers end-to-end, code-first tutorials covering every layer of production-grade GenAI agents, guiding you from spark to scale with proven patterns and reusable blueprints for re…

Implementing a ChatGPT-like LLM in PyTorch from scratch, step by step

In-depth tutorials on LLMs, RAGs and real-world AI agent applications.

[Lumina Embodied AI] 具身智能技术指南 Embodied-AI-Guide

Machine Learning Engineering Open Book

CUDA Core Compute Libraries

🚀 Awesome System for Machine Learning ⚡️ AI System Papers and Industry Practice. ⚡️ System for Machine Learning, LLM (Large Language Model), GenAI (Generative AI). 🍻 OSDI, NSDI, SIGCOMM, SoCC, MLSy…

hpc 教程，包含集合通信(mpi、nccl)、cuda 编程、向量化 SIMD、RDMA 通信等

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

Simple samples for TensorRT programming

A curated list of Large Language Model resources, covering model training, serving, fine-tuning, and building LLM applications.

GPU Kernels

CUDA Python: Performance meets Productivity

NVIDIA curated collection of educational resources related to general purpose GPU programming.

A next generation Python CMake adaptor and Python API for plugins

AITemplate is a Python framework which renders neural network into high performance CUDA/HIP C++ code. Specialized for FP16 TensorCore (NVIDIA GPU) and MatrixCore (AMD GPU) inference.

A Datacenter Scale Distributed Inference Serving Framework

RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine based on deep document understanding.

Material for gpu-mode lectures

making the official triton tutorials actually comprehensible

Kernels Written for 100 days of CUDA Challenge

Repo to accompany my mastering LLM engineering course

Question and Answer based on Anything.

A curated collection of resources, tutorials, and best practices for learning and mastering NVIDIA CUTLASS

xDiT: A Scalable Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism

AIInfra（AI 基础设施）指AI系统从底层芯片等硬件，到上层软件栈支持AI大模型训练和推理。

how to optimize some algorithm in cuda.