docs: Add Benchamark table

README.md

@@ -23,6 +23,85 @@ You can also test mask filling performance using `[MASK]` token like this.
 [BOS] [MASK] 먹었어? [EOS]
 ```
 
+## Benchmark
+
+<table>
+ <tr>
+  <th>Dataset</th>
+  
+  <td>KLUE NLI dev</th>
+  <td>NSMC test</td>
+  <td>QuestionPair test</td>
+  <td colspan="2">KLUE TC dev</td>
+  <td colspan="3">KLUE STS dev</td>
+  <td colspan="3">KorSTS dev</td>
+  <td colspan="2">HateSpeech dev</td>
+ </tr>
+ <tr>
+  <th>Metric</th>
+  
+  <!-- KLUE NLI -->
+  <td>Acc</th>
+  
+  <!-- NSMC -->
+  <td>Acc</td>
+  
+  <!-- QuestionPair -->
+  <td>Acc</td>
+  
+  <!-- KLUE TC -->
+  <td>Acc</td>
+  <td>F1</td>
+  
+  <!-- KLUE STS -->
+  <td>F1</td>
+  <td>Pearson</td>
+  <td>Spearman</td>
+  
+  <!-- KorSTS -->
+  <td>F1</td>
+  <td>Pearson</td>
+  <td>Spearman</td>
+  
+  <!-- HateSpeech -->
+  <td>Bias Acc</td>
+  <td>Hate Acc</td>
+ </tr>
+ 
+ <tr>
+  <th>Score</th>
+  
+  <!-- KLUE NLI -->
+  <td>0.5253</th>
+  
+  <!-- NSMC -->
+  <td>0.8425</td>
+  
+  <!-- QuestionPair -->
+  <td>0.8945</td>
+  
+  <!-- KLUE TC -->
+  <td>0.8047</td>
+  <td>0.7988</td>
+  
+  <!-- KLUE STS -->
+  <td>0.7411</td>
+  <td>0.7471</td>
+  <td>0.7399</td>
+  
+  <!-- KorSTS -->
+  <td>0.7725</td>
+  <td>0.6503</td>
+  <td>0.6191</td>
+  
+  <!-- HateSpeech -->
+  <td>0.7537</td>
+  <td>0.5605</td>
+ </tr>
+</table>
+
+- The performance was measured using [the notebooks here](https://github.com/cosmoquester/transformers-bart-finetune) with colab.
+
 ## Used Datasets
 
 ### [모두의 말뭉치](https://corpus.korean.go.kr/)