flash-attention-with-sink

🐙 Implements Flash Attention with sink for gpt-oss-20b; includes test.py. WIP backward pass, varlen support, and community sync to return softmax_lse only.

flash

View on GitHub

2 Stars

0 Forks

2 Watchers

Python Language

53.9 SrcLog Score

Cost to Build

$17.3K

Market Value

$11.5K

How is this calculated?

Growth over time

2 data points · 2026-04-12 → 2026-04-19

Stars Forks Watchers

💬

How do you feel about this project?

Ask AI about flash-attention-with-sink

Question copied to clipboard

What is the kelvindelrosario/flash-attention-with-sink GitHub project? Description: "🐙 Implements Flash Attention with sink for gpt-oss-20b; includes test.py. WIP backward pass, varlen support, and community sync to return softmax_lse only.". Written in Python. Explain what it does, its main use cases, key features, and who would benefit from using it.

Question is copied to clipboard — paste it after the AI opens.

How to clone flash-attention-with-sink

Clone via HTTPS

git clone https://github.com/kelvindelrosario/flash-attention-with-sink.git

Clone via SSH

[email protected]:kelvindelrosario/flash-attention-with-sink.git

Download ZIP

Download master.zip

Found an issue?

Report bugs or request features on the flash-attention-with-sink issue tracker:

Open GitHub Issues

Similar to flash-attention-with-sink

video.js open-source-flash zeroclipboard mediaelement Starling-Framework red5-server LayaAir-v1 SwiftyCam pmbootstrap IronOS freshplayerplugin livego spiffs flash Citrus-Engine EasyFlash video-js-swf DragonBonesAS MornUI Simple-Camera swf2js format-udf mediaelement-plugins SFUD dynamicaudio.js Aura.Session BOSSA AiLight SPIMemory janus-webrtc-gateway-docker