Update README.md

README.md

@@ -18,8 +18,9 @@ configs:
   - split: train
     path: data/train-*
 ---
-__Note__: [Oct. 3, 2024] this dataset is likely not fully aligned in terms of translation. It is being checked. 
+__Note__: [Oct. 3, 2024] this dataset is likely not fully aligned in terms of translation or the score are not compatible with Arabic translation. 
+It is being checked and so far, one clear mistranslation was spotted. 
+
+
 This is the Arabic (translated) queries of the MS-MARCO dataset that have scores in [this dataset](https://huggingface.co/datasets/lightonai/ms-marco-en-bge). 
-This dataset is used with akhooli/ar_mmarco_scores and akhooli/ar_mmarco_docs to train ColBERT knowledge distillation models (information retrieval). 
-Note: if you downloaded or tested this dataset (and the two related) before October 2, 2024 make sure to refresh your data. 
-There were alignment issues due to different versions of queries, documents and scores.
+This dataset is used with akhooli/ar_mmarco_scores and akhooli/ar_mmarco_docs to train ColBERT knowledge distillation models (information retrieval).