horrid-qvc/ArabicVoicesPseudoClean_v1

--- dataset_info: features: - name: audio_filepath dtype: string - name: text dtype: string - name: speaker dtype: string - name: duration dtype: float32 - name: session_id dtype: string - name: start dtype: float32 - name: end dtype: float32 - name: word_count dtype: int32 - name: audio dtype: audio: decode: false - name: gemini_transcript dtype: string splits: - name: train num_bytes: 6746970044 num_examples: 21464 download_size: 6742363175 dataset_size: 6746970044 configs: - config_name: default data_files: - split: train path: data/train-* ---