Update README.md

README.md

@@ -25,7 +25,8 @@ language:
 - **Model Developers:** Neural Magic
 
 Quantized version of [Meta-Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-405B-Instruct).
-It achieves an average score of 77.75 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 78.67.
+<!-- It achieves an average score of 77.75 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 78.67. -->
+It achieves an average recovery of 99.44% on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1) compared to the unquantized model.
 
 ### Model Optimizations
 
@@ -170,6 +171,7 @@ lm_eval \
   --tasks openllm \
   --batch_size auto
 ```
+Certain benchmarks for the full precision model are still being acquired. Average recovery is calculated only with metrics that both models have been evaluated on.
 
 ### Accuracy
 
@@ -188,71 +190,71 @@ lm_eval \
   <tr>
    <td>MMLU (5-shot)
    </td>
-   <td>82.21
+   <td>*
    </td>
-   <td>82.24
+   <td>86.06
    </td>
-   <td>100.0%
+   <td>*
    </td>
   </tr>
   <tr>
    <td>ARC Challenge (25-shot)
    </td>
-   <td>70.65
+   <td>73.38
    </td>
-   <td>69.03
+   <td>72.87
    </td>
-   <td>97.71%
+   <td>99.30%
    </td>
   </tr>
   <tr>
    <td>GSM-8K (5-shot, strict-match)
    </td>
-   <td>87.95
+   <td>95.07
    </td>
-   <td>86.50
+   <td>94.39
    </td>
-   <td>98.35%
+   <td>99.28%
    </td>
   </tr>
   <tr>
    <td>Hellaswag (10-shot)
    </td>
-   <td>86.33
+   <td>*
    </td>
-   <td>85.67
+   <td>*
    </td>
-   <td>99.24%
+   <td>*
    </td>
   </tr>
   <tr>
    <td>Winogrande (5-shot)
    </td>
-   <td>85.00
+   <td>87.21
    </td>
-   <td>85.79
+   <td>86.98
    </td>
-   <td>100.9%
+   <td>99.74%
    </td>
   </tr>
   <tr>
    <td>TruthfulQA (0-shot)
    </td>
-   <td>59.90
+   <td>*
    </td>
-   <td>57.24
+   <td>64.9
    </td>
-   <td>95.56%
+   <td>*
    </td>
   </tr>
   <tr>
    <td><strong>Average</strong>
    </td>
-   <td><strong>78.67</strong>
+   <td><strong>*</strong>
    </td>
-   <td><strong>77.75</strong>
+   <td><strong>*</strong>
    </td>
-   <td><strong>98.82%</strong>
+   <td><strong>99.44%</strong>
    </td>
   </tr>
 </table>