Update README.md

README.md

@@ -1,4 +1,6 @@
-A monolingual T5 model for Persian trained on oscar (https://oscar-corpus.com/) dataset with self-supervised method. It's similar to English T5 but just for Persian. You need to fine-tune it on your specific task. 
+A monolingual T5 model for Persian trained on OSCAR 21.09 (https://oscar-corpus.com/) corpus with self-supervised method. 35 Gig deduplicated version of Persian data was used for pre-training the model.
+
+It's similar to English T5 but just for Persian. You need to fine-tune it on your specific task. 
 
 Steps: 725000