Update README.md

README.md

@@ -68,4 +68,4 @@ print(tokenizer.batch_decode(output_sequences, skip_special_tokens=True))
 ```
 ### Limitations
 
-The model has been trained on text that has been artificially corrupted to look like OCR errors. These errors may not be similar for all OCR models and hence the model not do a good job at producing fully correct text. 
+The model has been trained on text that has been artificially corrupted to look like OCR errors. These errors may not be similar for all OCR models and hence the model may not do a good job at producing fully correct text.