Update README.md (#3)
Browse files- Update README.md (285f2701de45a4234238eac2e5e57c4611f2979b)
Co-authored-by: Dhia hf <[email protected]>
README.md
CHANGED
|
@@ -91,7 +91,10 @@ print(response)
|
|
| 91 |
<br>
|
| 92 |
|
| 93 |
## Benchmarks
|
| 94 |
-
We report in the following table our internal pipeline benchmarks
|
|
|
|
|
|
|
|
|
|
| 95 |
|
| 96 |
<table border="1" style="width: 100%; text-align: center; border-collapse: collapse;">
|
| 97 |
<colgroup>
|
|
|
|
| 91 |
<br>
|
| 92 |
|
| 93 |
## Benchmarks
|
| 94 |
+
We report in the following table our internal pipeline benchmarks.
|
| 95 |
+
- We use [lm-evaluation harness](https://github.com/EleutherAI/lm-evaluation-harness).
|
| 96 |
+
- We report **raw scores** obtained by applying chat template **without fewshot_as_multiturn** (unlike Llama3.1).
|
| 97 |
+
- We use same batch-size across all models.
|
| 98 |
|
| 99 |
<table border="1" style="width: 100%; text-align: center; border-collapse: collapse;">
|
| 100 |
<colgroup>
|