Add dataset README documentation

README.md

@@ -2,22 +2,37 @@
 license: apache-2.0
 task_categories:
 - audio-classification
-- other
 language:
 - en
 tags:
 - audio
 - speech
 - deepfake-detection
-- audio-authentication
-- speech-synthesis
-- temporal-localization
 size_categories:
-- 100K<n<1M
+- 10K<n<100K
 pretty_name: "SINE: Speech INfilling Edit Dataset"
+dataset_info:
+  features:
+  - name: audio
+    dtype: audio
+  - name: filename
+    dtype: string
+  - name: category
+    dtype: string
+  - name: timestamp
+    dtype: string
+  - name: label
+    dtype: class_label
+    names:
+      '0': manipulated
+      '1': authentic
+  - name: manipulation_type
+    dtype: string
 configs:
 - config_name: default
-  data_files: "data/*.tar.gz"
+  data_files:
+  - split: train
+    path: "preview/metadata.csv"
 ---
 
 # SINE Dataset