Update README.md

README.md

@@ -7,7 +7,7 @@ language:
 ---
 ## Description  
 
-The dataset is sourced from the French Senate's records, available at [this link](https://data.senat.fr/la-base-comptes-rendus/). After segmenting the text into sentences, 160,000 sentences were extracted.  
+The dataset is sourced from the French Senate's records, available at [this link](https://data.senat.fr/la-base-comptes-rendus/). After segmenting the text into sentences, 200,000 sentences were extracted(160,000 for train, 40,000 for test).  
 
 OCR errors are generated using the following script, designed to simulate errors found in certain documents from the French National Library caused by physical constraints (e.g., page curvature).