Update README.md

README.md

@@ -69,7 +69,7 @@ The model's performance during training:
 | 6000 | 0.196800      | 0.270521        | 0.687662 |
 
 ## Note
-Prior to training, specific character replacements were performed to reduce the tokenizer vocabulary by replacing digraphs with single characters. The replacements are as follows:
+To optimize training and reduce tokenizer vocabulary size, prior to training the following digraphs in the training data were replaced with single characters:
 ```
 гъ -> ɣ
 дж -> j