Update README.md

README.md

@@ -17,7 +17,21 @@ configs:
   - split: train
     path: data/train-*
 ---
-Dataset to Train TTS model: 
-No.of rows : 1.6k 
-Audio Language: Hinglish
-Transcript: Hindi (Devnagari)
+## Dataset to train any TTS in Hindi language:
+
+This dataset contains Hindi text with Hinglish audio.
+
+### Details
+
+Each base class for each unique class in each module file has a question and answer mp3 where one voice reads the question and another voice reads the answer. Both mp3s are stored in the parquet **dbytes** column and the associated source code **file_path** identifier.
+- Rows: 1164
+- Size: 1.6 GB
+- Data type: wav
+- Format: narrated Hinglish audio into Hindi script.
+### Schema
+```
+{
+    "audio": "wav"
+    "text": "string"
+}
+```