Update README.md

README.md

@@ -18,7 +18,7 @@ fine-tuned on [**Kazakh Speech Corpus 2**](https://huggingface.co/datasets/issai
 After training on the Train partition, it achieved **9.16% WER** on the Test partition.  
 
 **For longer audio** (35+ seconds), you can divide them into 30-second chunks, transcribe each chunk separately, and then merge the results.  
-Example implementation of a transcriber that can handle both short (up to 30 sec) and long (over 30 sec) audio files:  
+Example implementation of a transcriber that can handle both short and long audio files:  
 ```python
 import librosa
 import numpy as np