thu-ml / SpargeAttnView external linksLinks
[ICML2025] SpargeAttention: A training-free sparse attention that accelerates any model inference.
☆927Dec 31, 2025Updated last month
Alternatives and similar repositories for SpargeAttn
Users that are interested in SpargeAttn are comparing it to the libraries listed below
Sorting:
- [ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-t…☆3,159Jan 17, 2026Updated 3 weeks ago
- A sparse attention kernel supporting mix sparse patterns☆453Jan 18, 2026Updated 3 weeks ago
- https://wavespeed.ai/ Context parallel attention that accelerates DiT model inference with dynamic caching☆422Jul 5, 2025Updated 7 months ago
- [ICML2025, NeurIPS2025 Spotlight] Sparse VideoGen 1 & 2: Accelerating Video Diffusion Transformers with Sparse Attention☆627Feb 3, 2026Updated last week
- [ICML 2025] XAttention: Block Sparse Attention with Antidiagonal Scoring☆269Jul 6, 2025Updated 7 months ago
- A unified inference and post-training framework for accelerated video generation.☆3,059Updated this week
- Distributed Compiler based on Triton for Parallel Systems☆1,350Updated this week
- 🐳 Efficient Triton implementations for "Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention"☆964Feb 5, 2026Updated last week
- [ICLR2025 Spotlight] SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models☆3,673Updated this week
- [NeurIPS 2025] Radial Attention: O(nlogn) Sparse Attention with Energy Decay for Long Video Generation☆581Nov 11, 2025Updated 3 months ago
- ☆190Jan 14, 2025Updated last year
- xDiT: A Scalable Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism☆2,527Feb 5, 2026Updated last week
- Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model☆1,254Jun 8, 2025Updated 8 months ago
- A Distributed Attention Towards Linear Scalability for Ultra-Long Context, Heterogeneous Data Training☆631Feb 6, 2026Updated last week
- [ICLR 2025] DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Heads☆524Feb 10, 2025Updated last year
- Helpful tools and examples for working with flex-attention☆1,127Updated this week
- [ICLR 2025] FasterCache: Training-Free Video Diffusion Model Acceleration with High Quality☆259Dec 27, 2024Updated last year
- VideoSys: An easy and efficient system for video generation☆2,017Aug 27, 2025Updated 5 months ago
- 🚀 Efficient implementations of state-of-the-art linear attention models☆4,379Updated this week
- Model Compression Toolbox for Large Language Models and Diffusion Models☆753Aug 14, 2025Updated 6 months ago
- SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse–Linear Attention☆264Jan 17, 2026Updated 3 weeks ago
- A Survey of Efficient Attention Methods: Hardware-efficient, Sparse, Compact, and Linear Attention☆278Dec 1, 2025Updated 2 months ago
- 🤖FFPA: Extend FlashAttention-2 with Split-D, ~O(1) SRAM complexity for large headdim, 1.8x~3x↑🎉 vs SDPA EA.☆250Feb 5, 2026Updated last week
- [MLSys'25] QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving; [MLSys'25] LServe: Efficient Long-sequence LLM Se…☆812Mar 6, 2025Updated 11 months ago
- FlashInfer: Kernel Library for LLM Serving☆4,935Updated this week
- Quantized Attention on GPU☆44Nov 22, 2024Updated last year
- [ICLR'25] ViDiT-Q: Efficient and Accurate Quantization of Diffusion Transformers for Image and Video Generation☆149Mar 21, 2025Updated 10 months ago
- ☆65Apr 26, 2025Updated 9 months ago
- Tile primitives for speedy kernels☆3,139Updated this week
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference☆160Oct 13, 2025Updated 4 months ago
- Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels☆5,094Feb 7, 2026Updated last week
- Mirage Persistent Kernel: Compiling LLMs into a MegaKernel☆2,120Jan 29, 2026Updated 2 weeks ago
- MoBA: Mixture of Block Attention for Long-Context LLMs☆2,044Apr 3, 2025Updated 10 months ago
- Official PyTorch Implementation of "Optimal Stepsize for Diffusion Sampling".☆195Apr 13, 2025Updated 10 months ago
- FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.☆1,011Sep 4, 2024Updated last year
- [NeurIPS'24 Spotlight, ICLR'25, ICML'25] To speed up Long-context LLMs' inference, approximate and dynamic sparse calculate the attention…☆1,183Sep 30, 2025Updated 4 months ago
- [ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference☆372Jul 10, 2025Updated 7 months ago
- Code for Draft Attention☆99May 22, 2025Updated 8 months ago
- End-to-end recipes for optimizing diffusion models with torchao and diffusers (inference and FP8 training).☆393Jan 8, 2026Updated last month