Update README.md

README.md

@@ -15,8 +15,8 @@ Since most of the original urls are expired, we decided to directly provide the
 All the data can be download by the following link: <https://pan.baidu.com/s/1F5aKsurZkjZie09GsseOlw?pwd=vstf>
 
 The files with the suffix of '.tgz' need first to be uncompressed to a file with the suffix of '.tar' using the command line ```pigz -d baidu_images*.tgz ```.
-Enven though a part of the images are damaged or lossed because of some reasons, you can still access to the most part of TaiSu's data.
-Each image and its captions can be matched by the id, for example, 'img1baiducomitu1848496827104259151'.
+Even though a part of the images is damaged or lost because of some reasons, you can still access the most part of TaiSu's data. Each image and its captions can be matched by the id, for example, 'img1baiducomitu1848496827104259151'.
+
 
 ## Pretrained models ##
  Models trained on the web data of TaiSu and on the complete data of TaiSu are now availbale.