Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,115 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
### Accuracy
|
| 7 |
+
<table>
|
| 8 |
+
<thead>
|
| 9 |
+
<tr>
|
| 10 |
+
<th>Category</th>
|
| 11 |
+
<th>Metric</th>
|
| 12 |
+
<th>meta-llama/Llama-3.3-70B-Instruct</th>
|
| 13 |
+
<th>nm-testing/Llama-3.3-70B-Instruct-FP8-block</th>
|
| 14 |
+
<th>Recovery (%)</th>
|
| 15 |
+
</tr>
|
| 16 |
+
</thead>
|
| 17 |
+
<tbody>
|
| 18 |
+
<!-- OpenLLM Leaderboard V1 -->
|
| 19 |
+
<tr>
|
| 20 |
+
<td rowspan="7"><b>OpenLLM V1</b></td>
|
| 21 |
+
<td>ARC-Challenge (Acc-Norm, 25-shot)</td>
|
| 22 |
+
<td>72.53</td>
|
| 23 |
+
<td>72.61</td>
|
| 24 |
+
<td>100.12</td>
|
| 25 |
+
</tr>
|
| 26 |
+
<tr>
|
| 27 |
+
<td>GSM8K (Strict-Match, 5-shot)</td>
|
| 28 |
+
<td>76.35</td>
|
| 29 |
+
<td>73.16</td>
|
| 30 |
+
<td>95.83</td>
|
| 31 |
+
</tr>
|
| 32 |
+
<tr>
|
| 33 |
+
<td>HellaSwag (Acc-Norm, 10-shot)</td>
|
| 34 |
+
<td>86.65</td>
|
| 35 |
+
<td>86.56</td>
|
| 36 |
+
<td>99.90</td>
|
| 37 |
+
</tr>
|
| 38 |
+
<tr>
|
| 39 |
+
<td>MMLU (Acc, 5-shot)</td>
|
| 40 |
+
<td>82.51</td>
|
| 41 |
+
<td>82.38</td>
|
| 42 |
+
<td>99.84</td>
|
| 43 |
+
</tr>
|
| 44 |
+
<tr>
|
| 45 |
+
<td>TruthfulQA (MC2, 0-shot)</td>
|
| 46 |
+
<td>62.83</td>
|
| 47 |
+
<td>62.64</td>
|
| 48 |
+
<td>99.69</td>
|
| 49 |
+
</tr>
|
| 50 |
+
<tr>
|
| 51 |
+
<td>Winogrande (Acc, 5-shot)</td>
|
| 52 |
+
<td>83.50</td>
|
| 53 |
+
<td>83.27</td>
|
| 54 |
+
<td>99.72</td>
|
| 55 |
+
</tr>
|
| 56 |
+
<tr>
|
| 57 |
+
<td><b>Average Score</b></td>
|
| 58 |
+
<td><b>77.39</b></td>
|
| 59 |
+
<td><b>76.77</b></td>
|
| 60 |
+
<td><b>99.20</b></td>
|
| 61 |
+
</tr>
|
| 62 |
+
<!-- OpenLLM Leaderboard V2 -->
|
| 63 |
+
<tr>
|
| 64 |
+
<td rowspan="7"><b>OpenLLM V2</b></td>
|
| 65 |
+
<td>IFEval (Inst Level Strict Acc, 0-shot)</td>
|
| 66 |
+
<td>92.57</td>
|
| 67 |
+
<td>92.57</td>
|
| 68 |
+
<td>100.00</td>
|
| 69 |
+
</tr>
|
| 70 |
+
<tr>
|
| 71 |
+
<td>BBH (Acc-Norm, 3-shot)</td>
|
| 72 |
+
<td>69.03</td>
|
| 73 |
+
<td>68.98</td>
|
| 74 |
+
<td>99.92</td>
|
| 75 |
+
</tr>
|
| 76 |
+
<tr>
|
| 77 |
+
<td>Math-Hard (Exact-Match, 4-shot)</td>
|
| 78 |
+
<td>49.24</td>
|
| 79 |
+
<td>49.47</td>
|
| 80 |
+
<td>100.46</td>
|
| 81 |
+
</tr>
|
| 82 |
+
<tr>
|
| 83 |
+
<td>GPQA (Acc-Norm, 0-shot)</td>
|
| 84 |
+
<td>32.63</td>
|
| 85 |
+
<td>32.63</td>
|
| 86 |
+
<td>100.00</td>
|
| 87 |
+
</tr>
|
| 88 |
+
<tr>
|
| 89 |
+
<td>MUSR (Acc-Norm, 0-shot)</td>
|
| 90 |
+
<td>44.31</td>
|
| 91 |
+
<td>43.92</td>
|
| 92 |
+
<td>99.10</td>
|
| 93 |
+
</tr>
|
| 94 |
+
<tr>
|
| 95 |
+
<td>MMLU-Pro (Acc, 5-shot)</td>
|
| 96 |
+
<td>53.55</td>
|
| 97 |
+
<td>53.56</td>
|
| 98 |
+
<td>100.02</td>
|
| 99 |
+
</tr>
|
| 100 |
+
<tr>
|
| 101 |
+
<td><b>Average Score</b></td>
|
| 102 |
+
<td><b>56.89</b></td>
|
| 103 |
+
<td><b>56.85</b></td>
|
| 104 |
+
<td><b>99.93</b></td>
|
| 105 |
+
</tr>
|
| 106 |
+
<!-- Coding -->
|
| 107 |
+
<tr>
|
| 108 |
+
<td rowspan="1"><b>Coding</b></td>
|
| 109 |
+
<td>HumanEval Pass@1</td>
|
| 110 |
+
<td>N/A</td>
|
| 111 |
+
<td>N/A</td>
|
| 112 |
+
<td>N/A</td>
|
| 113 |
+
</tr>
|
| 114 |
+
</tbody>
|
| 115 |
+
</table>
|