Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,16 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: mit
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
---
|
| 4 |
+
# SageAttention 2++ Pre-compiled Wheel
|
| 5 |
+
|
| 6 |
+
🚀 **Ultra-fast attention mechanism with 2-3x speedup over FlashAttention2**
|
| 7 |
+
|
| 8 |
+
Pre-compiled Python wheel for high-performance GPU inference, optimized for RTX 4090 and CUDA 12.8+.
|
| 9 |
+
|
| 10 |
+
## 🚀 Quick Installation
|
| 11 |
+
|
| 12 |
+
### Method 1: Direct Pip Install (Recommended)
|
| 13 |
+
```bash
|
| 14 |
+
wget https://huggingface.co/ModelsLab/Sage_2_plus_plus_build/resolve/main/sageattention-2.2.0-cp311-cp311-linux_x86_64.whl
|
| 15 |
+
|
| 16 |
+
pip install sageattention-2.2.0-cp311-cp311-linux_x86_64.whl
|