Update README.md

README.md

@@ -462,26 +462,27 @@ You can finetune this model on your own dataset.
 
 #### Information Retrieval
 
-* Datasets: `miriad-eval-1kq-31kd` and `clean-val-only`
+* Datasets: `clean-val-only`
 * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
 
-| Metric              | miriad-eval-1kq-31kd | clean-val-only |
-|:--------------------|:---------------------|:---------------|
-| cosine_accuracy@1   | 0.6595               | 0.8471         |
-| cosine_accuracy@3   | 0.8518               | 0.9328         |
-| cosine_accuracy@5   | 0.8865               | 0.9544         |
-| cosine_accuracy@10  | 0.9221               | 0.9772         |
-| cosine_precision@1  | 0.6595               | 0.8471         |
-| cosine_precision@3  | 0.2839               | 0.3109         |
-| cosine_precision@5  | 0.1773               | 0.1909         |
-| cosine_precision@10 | 0.0922               | 0.0977         |
-| cosine_recall@1     | 0.6595               | 0.8471         |
-| cosine_recall@3     | 0.8518               | 0.9328         |
-| cosine_recall@5     | 0.8865               | 0.9544         |
-| cosine_recall@10    | 0.9221               | 0.9772         |
-| **cosine_ndcg@10**  | **0.8011**           | **0.9151**     |
-| cosine_mrr@10       | 0.7612               | 0.8949         |
-| cosine_map@100      | 0.7639               | 0.8957         |
+| Metric              | Base model (clean)   | Fine-tuned (clean) |
+|:--------------------|:---------------------|:-------------------|
+| cosine_accuracy@1   | 0.4425               | 0.8471             |
+| cosine_accuracy@3   | 0.6182               | 0.9328             |
+| cosine_accuracy@5   | 0.6779               | 0.9544             |
+| cosine_accuracy@10  | 0.7668               | 0.9772             |
+| cosine_precision@1  | 0.4425               | 0.8471             |
+| cosine_precision@3  | 0.2061               | 0.3109             |
+| cosine_precision@5  | 0.1356               | 0.1909             |
+| cosine_precision@10 | 0.0767               | 0.0977             |
+| cosine_recall@1     | 0.4425               | 0.8471             |
+| cosine_recall@3     | 0.6182               | 0.9328             |
+| cosine_recall@5     | 0.6779               | 0.9544             |
+| cosine_recall@10    | 0.7668               | 0.9772             |
+| **cosine_ndcg@10**  | **0.5994**           | **0.9151**         |
+| cosine_mrr@10       | 0.5465               | 0.8949             |
+| cosine_map@100      | 0.5533               | 0.8957             |
+
 
 <!--
 ## Bias, Risks and Limitations