Update README.md

README.md

@@ -9,4 +9,4 @@ library_name: transformers
 pipeline_tag: question-answering
 ---
 * The model was fine-tuned from `FacebookAI/roberta-large` on `rajpurkar/squad` for 3 epochs.
-* The model achieved an exact-match score of 89.5 and an f-1 score of 94.7.
+* The model achieved an exact-match score of 88.2 and an f-1 score of 94.3.