FlashKDA: high-performance Kimi Delta Attention kernels
☆1,253Sep 1, 2026Updated 2 weeks ago
Alternatives and similar repositories for FlashKDA
Users that are interested in FlashKDA are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- CUDA kernels for linear attention variants, written in CuTe DSL and CUTLASS C++.☆546Sep 8, 2026Updated last week
- high-performance linear attention kernel library built on TileLang☆702Aug 26, 2026Updated 3 weeks ago
- A kernel library written in tilelang☆1,770Apr 23, 2026Updated 4 months ago
- mKernel: fast multi-node, multi-GPU fused kernels☆272Updated this week
- A Quirky Assortment of CuTe Kernels☆1,146Updated this week
- AI Agents on DigitalOcean Gradient AI Platform • AdBuild production-ready AI agents using customizable tools or access multiple LLMs through a single endpoint. Create custom knowledge bases or connect external data.
- Accelerating MoE with IO and Tile-aware Optimizations☆767Aug 29, 2026Updated 2 weeks ago
- NVIDIA cuTile learn☆168Dec 9, 2025Updated 9 months ago
- ☆37Aug 7, 2025Updated last year
- High-performance LLM operator library built on TileLang.☆187Updated this week
- DeeperGEMM: crazy optimized version☆85May 5, 2025Updated last year
- [MLSys 26] 🥇 Solution for Gated Delta Net Track of MLSys 26 Flash infer competition☆36May 22, 2026Updated 3 months ago
- ☆453Aug 26, 2026Updated 3 weeks ago
- Triton kernels and PyTorch ops for Block Attention Residuals (AttnRes)☆89May 29, 2026Updated 3 months ago
- Distributed Compiler and Optimized Parallel Kernels☆1,546Aug 12, 2026Updated last month
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- ☆254Nov 19, 2025Updated 9 months ago
- Learning TileLang with 10 puzzles!☆379Aug 11, 2026Updated last month
- ☆51Aug 17, 2026Updated last month
- Tile-Based Runtime for Ultra-Low-Latency LLM Inference☆1,777Aug 13, 2026Updated last month
- TokenSpeed is a speed-of-light LLM inference engine.☆2,140Updated this week
- 🚀 Efficient implementations for emerging model architectures☆5,755Updated this week
- Tilus is a tile-level kernel programming language with explicit control over shared memory and registers.☆492Sep 7, 2026Updated last week
- ☆32Jul 2, 2025Updated last year
- ☆229Aug 26, 2026Updated 3 weeks ago
- Managed Kubernetes at scale on DigitalOcean • AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- Compact and Agent-Native MoE Training System☆350Updated this week
- Skills for writing tilelang and debugging with CUDA toolkits.☆137Updated this week
- PTX ISA 9.1 documentation converted to searchable markdown. Includes Claude Code skill for CUDA development.☆228Dec 24, 2025Updated 8 months ago
- ☆285Sep 5, 2026Updated last week
- FLA but cuTile☆27Apr 17, 2026Updated 5 months ago
- ☆1,608Nov 17, 2025Updated 10 months ago
- FA4-based Relative Attention Kernel developed by TML and Colfax☆18Sep 11, 2026Updated last week
- Kernel Library for Large Headdim Attention (64~1024, BF16/FP8/FP4), 1.5x~15x↑ vs PyTorch SDPA.☆333Updated this week
- ☆422Updated this week
- End-to-end encrypted cloud storage - Proton Drive • AdSpecial offer: 40% Off Yearly / 80% Off First Month. Protect your most important files, photos, and documents from prying eyes.
- Official repository for Parallax (Parameterized Local Linear Attention)☆69Jul 30, 2026Updated last month
- High Performance LLM Inference Operator Library☆1,159Sep 7, 2026Updated last week
- a size profiler for cuda binary☆71Jan 15, 2026Updated 8 months ago
- Distributed MoE in a Single Kernel [NeurIPS '25]☆291May 5, 2026Updated 4 months ago
- NVSHMEM‑Tutorial: Build a DeepEP‑like GPU Buffer☆201Feb 11, 2026Updated 7 months ago
- Agent-friendly GPU profile-query CLI☆122Aug 21, 2026Updated 3 weeks ago
- An experimental communicating attention kernel based on DeepEP.☆34Jul 29, 2025Updated last year