Complete dataset configs

README.md

@@ -40,7 +40,7 @@ configs:
   - split: test
     path: data/partitions/yt_seg_titles.test.json
 dataset_info:
-  config_name: audio
+- config_name: audio
   features:
   - name: audio
     dtype:
@@ -92,6 +92,58 @@ dataset_info:
     num_examples: 1448
   download_size: 71470669858
   dataset_size: 73269108427
+- config_name: text
+  features:
+  - name: audio_path
+    dtype: string
+  - name: text_ref
+    list: string
+  - name: text_wt
+    list: string
+  - name: text_wl
+    list: string
+  - name: target_binary_ref
+    dtype: string
+  - name: target_binary_wt
+    dtype: string
+  - name: target_binary_wl
+    dtype: string
+  - name: target_text_ref
+    dtype: string
+  - name: target_text_ts_ref
+    dtype: string
+  - name: target_ts
+    dtype: string
+  - name: chapter_titles
+    list: string
+  - name: chapter_timestamps
+    list: float64
+  - name: channel_id
+    dtype: string
+  - name: video_id
+    dtype: string
+  - name: speaker_category
+    dtype: string
+  - name: dominant_speaker_proportion
+    dtype: float64
+  - name: num_speakers
+    dtype: int64
+  - name: duration
+    dtype: float64
+- config_name: titles
+  features:
+  - name: text_section_ref
+    dtype: string
+  - name: text_section_prev_titles_ref
+    dtype: string
+  - name: target_title
+    dtype: string
+  - name: channel_id
+    dtype: string
+  - name: video_id
+    dtype: string
+  - name: chapter_idx
+    dtype: int64
 ---
 
 # From Text Segmentation to Smart Chaptering: A Novel Benchmark for Structuring Video Transcriptions