Train speculative decoding models effortlessly and port them smoothly to SGLang serving.

Supercharge Your LLM with the Fastest KV Cache Layer

Nano vLLM

A light llama-like llm inference framework based on the triton kernel.

Collection of kernels written in Triton language

Building DeepSeek R1 from Scratch

[ACL 2024 Demo] Official GitHub repo for UltraEval: An open source framework for evaluating foundation models.

General technology for enabling AI capabilities w/ LLMs and MLLMs

Composable camera rigs

A curated list for Efficient Large Language Models

小彭老师特意从零开始研发的一款教学用，基于 C++17 回调函数的异步 HTTP 服务器

A self-learning tutorail for CUDA High Performance Programing.

TTS

A generative speech model for daily dialogue.

Tile primitives for speedy kernels

📚LeetCUDA: Modern CUDA Learn Notes with PyTorch for Beginners🐑, 200+ CUDA Kernels, Tensor Cores, HGEMM, FA-2 MMA.🎉

A minimal GPU design in Verilog to learn how GPUs work from the ground up

Material for gpu-mode lectures

LLM training in simple, raw C/CUDA

Distribute and run LLMs with a single file.

Header only c++ network library, based on asio,support tcp,udp,http,websocket,rpc,ssl,icmp,serial_port,socks5.

Fast Multimodal LLM on Mobile Devices

Open-Sora: Democratizing Efficient Video Production for All

Implementation of FlashAttention in PyTorch

An improved server based on MapleSolaxia (v83 MapleStory private server)

model convert extension for stable-diffusion-webui. supports convert fp16/bf16 no-ema/ema-only safetensors