Add transcriptions: 3,933 segments transcribed with OpenAI Whisper API, hallucinations removed (0.93%)

README.md

@@ -16,6 +16,44 @@ tags:
 - fine-tuning
 size_categories:
 - 1K<n<10K
+dataset_info:
+  features:
+  - name: id
+    dtype: string
+  - name: audio
+    dtype:
+      audio:
+        sampling_rate: 16000
+  - name: text
+    dtype: string
+  - name: duration
+    dtype: float32
+  - name: category
+    dtype: string
+  - name: quality
+    dtype: string
+  - name: source
+    dtype: string
+  - name: speaker_role
+    dtype: string
+  - name: domain
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 1397393290.44
+    num_examples: 3720
+  - name: validation
+    num_bytes: 81247861.0
+    num_examples: 213
+  download_size: 1433246274
+  dataset_size: 1478641151.44
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+  - split: validation
+    path: data/validation-*
 ---
 
 # French Education Speech - Transcribed Dataset