Update README.md

README.md

@@ -13,7 +13,7 @@ widget:
 
 
 This Bangla Question Answering model has been using Roberta Model architecture, which is currently trained on a small set of human-annotated data. 
-However, a larger corpus model is in progress. To train this model, the Bangla QA data has been converted into the SQuAD v2 format as well as preprocessed it. 
+However, a larger corpus model is in progress. For training this model, the Bangla QA data has been converted into the SQuAD v2 format as well as preprocessed. 
 The dataset contains 2504 question-answer pairs.
 
 Github Link:![Link](https://github.com/saiful9379/Bangla_Roberta_Question_and_Answer)