Dataset Viewer
Auto-converted to Parquet Duplicate
The dataset viewer is not available for this split.
Parquet error: Scan size limit exceeded: attempted to read 338029525 bytes, limit is 300000000 bytes Make sure that 1. the Parquet files contain a page index to enable random access without loading entire row groups2. otherwise use smaller row-group sizes when serializing the Parquet files
Error code:   TooBigContentError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

Bengali TTS — YouTube Pipeline (test)

YouTube-sourced Bengali speech chunks transcribed with Gemini 2.5 Flash, produced by the YT scraping + VAD chunking + Gemini transcription pipeline.

Columns

Column Description
uuid Unique identifier (<video_id>_<chunk#>)
speaker Speaker / channel name
video_id YouTube video ID
chunk_file Chunk filename
audio_file <speaker>_<video_id>_<chunk_file>
duration Duration in seconds
transcription Gemini transcription text
audio Audio (WAV, 16-bit PCM, mono, 22050 Hz)
gemini_transcription Gemini transcription text
whisper_transcription Whisper ASR output (NULL — not run yet)
wer Word Error Rate (NULL — not computed yet)
cer Character Error Rate (NULL — not computed yet)
has_english English code-switching flag (NULL — not computed yet)
decision KEEP / EXCLUDE
Downloads last month
8