Add dataset card with subset configurations

README.md

@@ -9,23 +9,32 @@ dataset_info:
     - name: label
       dtype: int64
     - name: audio
+      dtype:
+        audio:
+          sampling_rate: 16000
+    - name: md5_audio
       dtype: string
+    - name: duration
+      dtype: float32
   splits:
     - name: train
-      num_bytes: 6148000
+      num_bytes: 1537000000
       num_examples: 3074
     - name: test
-      num_bytes: 1538000
+      num_bytes: 384500000
       num_examples: 769
-  download_size: 6533100
-  dataset_size: 7686000
+  download_size: 1633275000
+  dataset_size: 1921500000
 task_categories:
-  - text-classification
-  - text-generation
+  - automatic-speech-recognition
+  - text-to-speech
 language:
   - en
 tags:
   - dataset
+  - audio
+  - speech
+  - asr
 pretty_name: Mucs
 size_categories:
   - 1K<n<10K
@@ -36,7 +45,7 @@ size_categories:
 ## Dataset Description
 
 This dataset contains 3,843 samples organized across multiple splits.
-The dataset includes text data.
+The dataset includes audio data.
 
 ### Splits
 
@@ -45,7 +54,7 @@ The dataset includes text data.
 
 ## Dataset Creation
 
-This dataset was created using StreamableDatasetManager on 2025-10-16T12:28:48.085816.
+This dataset was created using StreamableDatasetManager on 2025-10-16T12:45:15.159031.
 
 ### Data Fields
 
@@ -53,10 +62,13 @@ The dataset includes the following columns:
 - **id**: Int64 data
 - **text**: String data
 - **label**: Int64 data
-- **audio**: String data
+- **audio**: Audio data (16kHz sampling rate)
+- **md5_audio**: String data
+- **duration**: Float32 data
 
 ## Technical Details
 
 - **Total samples**: 3,843
 - **Shard length**: 1000
-- **Number of splits**: 2
+- **Number of splits**: 2
+- **Audio format**: 16kHz sampling rate