Update README.md
Browse files
README.md
CHANGED
|
@@ -10,7 +10,7 @@ tags:
|
|
| 10 |
lyraXVERSE is currently the **fastest XVERSE-13b** available. The inference speed of lyraXVERSE has achieved up to **3900+ tokens/s** on A100, up to **2.7x** acceleration upon the torch version.
|
| 11 |
|
| 12 |
Among its main features are:
|
| 13 |
-
- device: Nvidia GPU with Amperer architecture or Volta architecture (A100 or higher, V100).
|
| 14 |
- batch_size: compiled with dynamic batch size, maximum depends on device.
|
| 15 |
- MEMOPT mode: significantly optimized VRAM usage and increased speed
|
| 16 |
|
|
|
|
| 10 |
lyraXVERSE is currently the **fastest XVERSE-13b** available. The inference speed of lyraXVERSE has achieved up to **3900+ tokens/s** on A100, up to **2.7x** acceleration upon the torch version.
|
| 11 |
|
| 12 |
Among its main features are:
|
| 13 |
+
- device: Nvidia GPU with Amperer architecture or Volta architecture (A10, A100 or higher, V100).
|
| 14 |
- batch_size: compiled with dynamic batch size, maximum depends on device.
|
| 15 |
- MEMOPT mode: significantly optimized VRAM usage and increased speed
|
| 16 |
|