https://github.com/simonri/flash-attention-3-h100
FlashAttention-3 pre-built wheels for H100
https://github.com/simonri/flash-attention-3-h100
Last synced: 12 days ago
JSON representation
FlashAttention-3 pre-built wheels for H100
- Host: GitHub
- URL: https://github.com/simonri/flash-attention-3-h100
- Owner: simonri
- Created: 2025-09-27T22:35:59.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2025-11-16T16:30:40.000Z (8 months ago)
- Last Synced: 2026-05-30T22:06:46.914Z (about 1 month ago)
- Language: Python
- Homepage:
- Size: 21.5 KB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# FlashAttention-3 pre-built wheels for H100
This repository hosts prebuilt Python wheels for [flash-attn v3](https://github.com/Dao-AILab/flash-attention),
compiled for NVIDIA H100 GPUs.
## Available wheels
- **flash_attn_3-3.0.0b1-cp39-abi3-linux_x86_64.whl**
- Python **3.9+** (abi3)
- Linux x86_64
- Built with CUDA support for **H100 (sm_90)**
## Installation
Download the wheel from the [Releases](../../releases) page, then install with:
```bash
pip install flash_attn_3-3.0.0b1-cp39-abi3-linux_x86_64.whl
```