zhuzilin / flash-attention-with-sinkView external linksLinks
☆38Aug 7, 2025Updated 6 months ago
Alternatives and similar repositories for flash-attention-with-sink
Users that are interested in flash-attention-with-sink are comparing it to the libraries listed below
Sorting:
- High performance RMSNorm Implement by using SM Core Storage(Registers and Shared Memory)☆26Jan 22, 2026Updated 3 weeks ago
- My tests and experiments with some popular dl frameworks.☆17Sep 11, 2025Updated 5 months ago
- DeeperGEMM: crazy optimized version☆74May 5, 2025Updated 9 months ago
- Official repository for the paper Local Linear Attention: An Optimal Interpolation of Linear and Softmax Attention For Test-Time Regressi…☆23Oct 1, 2025Updated 4 months ago
- Persistent dense gemm for Hopper in `CuTeDSL`☆15Aug 9, 2025Updated 6 months ago
- An experimental communicating attention kernel based on DeepEP.☆35Jul 29, 2025Updated 6 months ago
- ☆42Nov 1, 2025Updated 3 months ago
- Fast and memory-efficient exact attention☆18Jan 23, 2026Updated 3 weeks ago
- Nex Venus Communication Library☆72Nov 17, 2025Updated 3 months ago
- Aims to implement dual-port and multi-qp solutions in deepEP ibrc transport☆73May 9, 2025Updated 9 months ago
- It is an LLM-based AI agent, which can write correct and efficient gpu kernels automatically.☆63Updated this week
- ☆130Aug 18, 2025Updated 6 months ago
- Implement Flash Attention using Cute.☆100Dec 17, 2024Updated last year
- Benchmark tests supporting the TiledCUDA library.☆18Nov 19, 2024Updated last year
- DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling☆22Feb 9, 2026Updated last week
- Tile-based language built for AI computation across all scales☆123Updated this week
- ☆86Feb 10, 2026Updated last week
- ☆118May 19, 2025Updated 8 months ago
- NSA Triton Kernels written with GPT5 and Opus 4.1☆70Aug 12, 2025Updated 6 months ago
- Wave: Python Domain-Specific Language for High Performance Machine Learning☆44Updated this week
- Debug print operator for cudagraph debugging☆14Aug 2, 2024Updated last year
- TiledLower is a Dataflow Analysis and Codegen Framework written in Rust.☆14Nov 23, 2024Updated last year
- Github repo for ICLR-2025 paper, Fine-tuning Large Language Models with Sparse Matrices☆24Feb 2, 2026Updated 2 weeks ago
- ☆23Updated this week
- STREAMer: Benchmarking remote volatile and non-volatile memory bandwidth☆17Aug 21, 2023Updated 2 years ago
- DeepSeek-V3.2-Exp DSA Warmup Lightning Indexer training operator based on tilelang☆43Nov 19, 2025Updated 2 months ago
- Handwritten GEMM using Intel AMX (Advanced Matrix Extension)☆17Jan 11, 2025Updated last year
- A Triton JIT runtime and ffi provider in C++☆31Jan 26, 2026Updated 3 weeks ago
- ☆32Jul 2, 2025Updated 7 months ago
- Open deep learning compiler stack for cpu, gpu and specialized accelerators☆19Feb 9, 2026Updated last week
- ☆41Oct 15, 2025Updated 4 months ago
- Accelerate LLM preference tuning via prefix sharing with a single line of code☆51Jul 4, 2025Updated 7 months ago
- ☆95Apr 2, 2025Updated 10 months ago
- ☆47Jul 16, 2025Updated 7 months ago
- An open source SDR SDRAM controller based on the AXI4 bus and verified by FPGA and tapeout. It can support memory particles of different …☆22May 12, 2025Updated 9 months ago
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆79Aug 12, 2024Updated last year
- Tilus is a tile-level kernel programming language with explicit control over shared memory and registers.☆443Feb 4, 2026Updated last week
- LLMA = LLM + Arithmetic coder, which use LLM to do insane text data compression. LLMA=大模型+算术编码，它能使用LLM对文本数据进行暴力的压缩，达到极高的压缩率。☆22Nov 24, 2024Updated last year
- ☆60Updated this week