add evals for bf16

README.md

@@ -22,7 +22,7 @@ tags:
 - **Model Developers:** Neural Magic
 
 Quantized version of [Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1).
-It achieves an average score of <TODO> on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves <TODO>.
+It achieves an average score of <TODO> on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 73.44.
 
 ### Model Optimizations
 
@@ -172,12 +172,12 @@ lm_eval \
 
 | Metric                                   | mistralai/Mixtral-8x7B-Instruct-v0.1             | neuralmagic/Mixtral-8x7B-Instruct-v0.1-FP8 |
 |-----------------------------------------|:---------------------------------:|:-------------------------------------------:|
-| ARC-Challenge (Acc-Norm, 25-shot)       |       <TODO>                     |                 <TODO>                       |
-| GSM8K (Strict-Match, 5-shot)            |         <TODO>                    |                 <TODO>                        |
-| HellaSwag (Acc-Norm, 10-shot)           |         <TODO>                    |                 <TODO>                       |
-| MMLU (Acc, 5-shot)                      |         <TODO>                    |                 <TODO>                       |
-| TruthfulQA (MC2, 0-shot)                |          <TODO>                   |                 <TODO>                       |
-| Winogrande (Acc, 5-shot)                |          <TODO>                   |                 <TODO>                       |
-| **Average Score**                       | **<TODO>**                        | **<TODO>**                                   |
+| ARC-Challenge (Acc-Norm, 25-shot)       |       70.48                     |                 <TODO>                       |
+| GSM8K (Strict-Match, 5-shot)            |       65.50                    |                 <TODO>                        |
+| HellaSwag (Acc-Norm, 10-shot)           |         87.33                    |                 <TODO>                       |
+| MMLU (Acc, 5-shot)                      |         70.30                    |                 <TODO>                       |
+| TruthfulQA (MC2, 0-shot)                |          64.81                   |                 <TODO>                       |
+| Winogrande (Acc, 5-shot)                |          82.24                   |                 <TODO>                       |
+| **Average Score**                       | **73.44**                        | **<TODO>**                                   |
 | **Recovery (%)**                            | **100.00**                       | **<TODO>**                                   |