Update dataset card schema for clip-level format

README.md

@@ -6,14 +6,13 @@ configs:
     path: data/train-*
 dataset_info:
   features:
+  - name: clip_id
+    dtype: string
   - name: audio
     dtype: audio
-  - name: annotation
+  - name: segments
     dtype: string
   splits:
   - name: train
-    num_bytes: 28899427409
-    num_examples: 1414
-  download_size: 30607562312
-  dataset_size: 28899427409
+    num_examples: 12
 ---