What is flash-attention built with?
flash-attention is a Python repo. It uses AI through apache-tvm-ffi, NVIDIA, quack-kernels, torch-c-dlpack-ext, triton-windows, and 3 more. It has 139,521 written lines, with 80 test files.
Read on November 16, 2026, at commit 328dd7b. Read it as markdown
The card says what it's built with. Its whole map opens it up: every part of the code and how it connects to the rest.
Backend
AI 8
- torch86 files
apache-tvm-ffi
NVIDIA
- einops25 files
- triton10 files
- quack-kernels
- torch-c-dlpack-ext
- triton-windows
Other repos with a similar stack
Ordered by how much of its stack they share with flash-attention, then by stars.
- NVIDIATensorRT-LLM15k
TensorRT LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and supports state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT LLM also contains components to create Python and C++ runtimes that orchestrate the inference execution in a performant way.
7 of 12 in commonCompare →
Frequently asked questions
What language is flash-attention written in?
Does flash-attention use AI?
Does flash-attention have tests?
Which repos have a stack like flash-attention's?
See inside your code, today.
Paste your repo and see it drawn, part by part: a medium one in under a minute.
Free, no account.Private repo? Sign in with GitHub
Is it your repo? Add its badge, keep it up to date or take it down →

