Update README.md

README.md

@@ -118,9 +118,8 @@ This is the recommended way to use the dataset, as it avoids downloading all ~5.
 
 ```python
 from datasets import load_dataset
-import json
 
-# The library automatically finds all .tar.gz shards
+# The library automatically finds all .tar.gz shards in the repo
 ds = load_dataset(
     "freococo/media_queen_entertaiment_voices",
     split="train",
@@ -129,16 +128,17 @@ ds = load_dataset(
 
 # Iterate through the first 5 samples
 for sample in ds.take(5):
-    # The transcript is now a top-level feature for easy access!
+    # The transcript is a top-level feature for easy access
     print(f"🎙️ Transcript: {sample['txt']}")
 
-    # The audio can be accessed directly as a NumPy array
-    audio_array = sample['audio']['array']
-    sampling_rate = sample['audio']['sampling_rate']
+    # The audio data is in the 'mp3' column
+    audio_data = sample['mp3']
+    audio_array = audio_data['array']
+    sampling_rate = audio_data['sampling_rate']
     print(f"🎧 Audio loaded with shape {audio_array.shape} and rate {sampling_rate} Hz")
 
-    # The full metadata is still available in the JSON string
-    metadata = json.loads(sample['json'])
+    # The 'json' column is ALREADY a Python dictionary
+    metadata = sample['json']
     print(f"📺 Channel: {metadata.get('channel')}")
     print(f"🎥 Video URL: {metadata.get('video_url')}")
     print("---")