Add dataset metadata

dataset_info.json

@@ -0,0 +1,19 @@
+{
+  "total_samples": 5,
+  "dataset_name": "speaker-diarization-natural-dataset",
+  "created_at": "2025-10-25T10:30:00.679932",
+  "image_columns": [],
+  "audio_columns": [
+    "audio"
+  ],
+  "subsets": [],
+  "splits": {
+    "default": {
+      "train": {
+        "num_samples": 5,
+        "num_shards": 1,
+        "shard_length": 5
+      }
+    }
+  }
+}