Explainations of demo.md

demo.md

@@ -1,5 +1,7 @@
 # Evaluate ASR models
 
+This is a breakdown of the steps to evaluate ASR models on a small subset of the Librispeech dataset based on the script in the `evaluate_asr.py` file.
+
 ## 0. Import the necessary libraries
 
 ```python
@@ -14,20 +16,6 @@ from collections import defaultdict
 import json
 ```
 
-Requirements:
-
-```txt
-transformers==4.44.0
-torch
-jiwer
-soundfile
-numpy
-datasets
-gradio
-librosa
-evaluate
-```
-
 
 ## 1. Pick a speech dataset (English)from the Hugging Face hub and create a small subset of this dataset (100 rows) by streaming the data