Update README.md
Browse files
README.md
CHANGED
|
@@ -104,7 +104,7 @@ This is a new kind of model optimization. It is based on a new method for the an
|
|
| 104 |
### Model improvement with layer duplication:
|
| 105 |
| | Average | IFEval | BBH | MATH Lvl 5 | GPQA | MUSR | MMLU-PRO |
|
| 106 |
|-----------------|--------:|-------:|-----:|-----------:|-----:|------:|---------:|
|
| 107 |
-
| RYS Improvement | 2.61 | -2.05 | 2.51 | 8.16 | 2.58 | 17.72 | 0.31 |
|
| 108 |
|
| 109 |
|
| 110 |
This model is based on MaziyarPanahi/calme-2.1-qwen2-72b, which in turn was tuned from Qwen2-72B. As this method is orthogonal to fine-tuning, the further finetune from MaziyarPanahi now has the top position:
|
|
|
|
| 104 |
### Model improvement with layer duplication:
|
| 105 |
| | Average | IFEval | BBH | MATH Lvl 5 | GPQA | MUSR | MMLU-PRO |
|
| 106 |
|-----------------|--------:|-------:|-----:|-----------:|-----:|------:|---------:|
|
| 107 |
+
| RYS Improvement | 2.61% | -2.05% | 2.51% | 8.16% | 2.58% | 17.72% | 0.31% |
|
| 108 |
|
| 109 |
|
| 110 |
This model is based on MaziyarPanahi/calme-2.1-qwen2-72b, which in turn was tuned from Qwen2-72B. As this method is orthogonal to fine-tuning, the further finetune from MaziyarPanahi now has the top position:
|