GitHub Trending

MoonshotAI/FlashKDA — FlashKDA: high-performance Kimi Delta Attention kernels

29 sec read2 views0 listens

1,062 stars | 91 forks | Cuda

FlashKDA: high-performance Kimi Delta Attention kernels

What it does

FlashKDA is a high-performance CUDA kernel for Kimi Delta Attention, optimized for SM90 and above GPUs with CUDA 12.9+. It integrates seamlessly with flash-linear-attention to enhance performance in AI applications.

Why it matters: 🚀 Dive into the future of GPU-accelerated AI with FlashKDA, a cutting-edge CUDA kernel for Kimi Delta Attention. #AI #CUDA

View on GitHub

Trending today with 91 new stars

Want to create content about this repo? Use Nemati AI tools to generate articles, tutorials, and social posts.

deepseek-ai/DeepEP — DeepEP: an efficient expert-parallel communication library

9,312 stars | 1,180 forks | Cuda DeepEP: an efficient expert-parallel communication library What it does DeepEP is a communication library designed for Mixture-of-Experts models, offering high-throughput and low-latency GPU kernels that enhance both ...

Ali Nemati

GitHub TrendingApr 2529 sec read

deepseek-ai/DeepGEMM — DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

6,903 stars | 871 forks | Cuda DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling What it does DeepGEMM is a high-performance CUDA library for tensor core kernels, offering efficient FP8 GEMM operations and more. It simplifies G...

Ali Nemati

GitHub TrendingApr 1329 sec read

jamiepine/voicebox — The open-source voice synthesis studio

15,797 stars | 1,879 forks | TypeScript The open-source voice synthesis studio What it does Voicebox is an open-source voice synthesis studio that allows users to clone voices, generate speech in multiple languages, and apply various effects locally ...

Ali Nemati

GitHub Trending1 hour ago29 sec read

vudovn/ag-kit

8,034 stars | 1,505 forks | TypeScript What it does AG KIT is a TypeScript toolkit for creating and managing AI agents within Google Antigravity, offering features like rules discovery, specialist routing, persistent memory, and safety hooks to ensur...

Ali Nemati

GitHub Trending1 hour ago28 sec read

maderix/ANE — Training neural networks on Apple Neural Engine via reverse-engineered private A

7,197 stars | 956 forks | Objective-C Training neural networks on Apple Neural Engine via reverse-engineered private APIs What it does This repository demonstrates training neural networks directly on Apple's Neural Engine using reverse-engineered pr...

Ali Nemati

MoonshotAI/FlashKDA — FlashKDA: high-performance Kimi Delta Attention kernels

What it does

Related Articles

deepseek-ai/DeepEP — DeepEP: an efficient expert-parallel communication library

deepseek-ai/DeepGEMM — DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

jamiepine/voicebox — The open-source voice synthesis studio

vudovn/ag-kit

maderix/ANE — Training neural networks on Apple Neural Engine via reverse-engineered private A