Update README.md
Browse files
README.md
CHANGED
|
@@ -21,6 +21,7 @@ We use the LLaMA.13B model for measurement, but this optimized inference is appl
|
|
| 21 |
|
| 22 |
## Speed
|
| 23 |
|
|
|
|
| 24 |
* test on A100 40G
|
| 25 |
* fp16 precision
|
| 26 |
|
|
|
|
| 21 |
|
| 22 |
## Speed
|
| 23 |
|
| 24 |
+
* Evaluated at tokens/s
|
| 25 |
* test on A100 40G
|
| 26 |
* fp16 precision
|
| 27 |
|