Update README.md

README.md

@@ -9,9 +9,9 @@ This is an implementation of CRNN (CNN+LSTM+CTC) for chinese text recognition.
 1. Download the [Synthetic Chinese Dataset](https://pan.baidu.com/s/1dFda6R3)(contributed by https://github.com/senlinuc/caffe_ocr)
 
 A glance of the dataset: 
-* `almost 3.6 million synthetic chinese text images.
-* `5,990 different categories in total.
-* `Each image has a length of 10 characters. 
+* almost 3.6 million synthetic chinese text images.
+* 5,990 different categories in total.
+* each image has a length of 10 characters. 
    
 2. Create train.txt and text.txt with the format like this:  
 ```