Update README.md
Browse files
README.md
CHANGED
|
@@ -126,3 +126,7 @@ Detailed results can be found [here](https://huggingface.co/datasets/open-llm-le
|
|
| 126 |
|Winogrande (5-shot) |81.06|
|
| 127 |
|GSM8k (5-shot) |45.03|
|
| 128 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 126 |
|Winogrande (5-shot) |81.06|
|
| 127 |
|GSM8k (5-shot) |45.03|
|
| 128 |
|
| 129 |
+
### Results
|
| 130 |
+
- small quality loss can be observed comparing to base model, as described in the DUS paper
|
| 131 |
+
- this merge has best evaluation results, os it will be finetuned to 'recover' from the merge
|
| 132 |
+
- v03 > v01 > v02 - based on average evaluation scores, removing 1/4 of total layers seems to be the correct way to scale DUS
|