Update README.md

README.md

@@ -27,7 +27,7 @@ You can use the model directly with a pipeline for extractive question answering
 
 ## Training data
 
-Training data comes from three sources: [cmrc2018](https://github.com/ymcui/cmrc2018), [webqa](https://spaces.ac.cn/archives/4338), and [laisi](https://www.kesci.com/home/competition/5d142d8cbb14e6002c04e14a/content/0). We only use train set of the three datasets.
+Training data comes from three sources: [cmrc2018](https://github.com/ymcui/cmrc2018), [webqa](https://spaces.ac.cn/archives/4338), and [laisi](https://www.kesci.com/home/competition/5d142d8cbb14e6002c04e14a/content/0). We only use the train set of three datasets.
 
 ## Training procedure