Expert Specialization MoE Solution based on CUTLASS

[EuroSys'25] Mist: Efficient Distributed Training of Large Language Models via Memory-Parallelism Co-Optimization

SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse–Linear Attention

Tritonbench is a collection of PyTorch custom operators with example inputs to measure their performance.

LongLive: Real-time Interactive Long Video Generation

[ICML2025, NeurIPS2025 Spotlight] Sparse VideoGen 1 & 2: Accelerating Video Diffusion Transformers with Sparse Attention

Distributed MoE in a Single Kernel [NeurIPS '25]

[NeurIPS 2025] ClusterFusion: Expanding Operator Fusion Scope for LLM Inference via Cluster-Level Collective Primitive

🔥 LLM-powered GPU kernel synthesis: Train models to convert PyTorch ops into optimized Triton kernels via SFT+RL. Multi-turn compilation feedback, cross-platform NVIDIA/AMD, Kernelbook + KernelBench

🥢像老乡鸡🐔那样做饭。主要部分于2024年完工，非老乡鸡官方仓库。文字来自《老乡鸡菜品溯源报告》，并做归纳、编辑与整理。CookLikeHOC.

NVSHMEM‑Tutorial: Build a DeepEP‑like GPU Buffer

Run your own AI cluster at home with everyday devices 📱💻 🖥️⌚

gpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI

CUDA Templates and Python DSLs for High-Performance Linear Algebra

Text-audio foundation model from Boson AI

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

A curated list of recent papers on efficient video attention for video diffusion models, including sparsification, quantization, and caching, etc.

open-source coding LLM for software engineering tasks

诺亚盘古大模型研发背后的真正的心酸与黑暗的故事。

Storing long contexts in tiny caches with self-study

ArcticInference: vLLM plugin for high-throughput, low-latency inference

kernels, of the mega variety

ByteCheckpoint: An Unified Checkpointing Library for LFMs

https://wavespeed.ai/ Context parallel attention that accelerates DiT model inference with dynamic caching