---
language:
- ne
license: unknown  # set this to match your actual sources
task_categories:
- automatic-speech-recognition
---

# Nepali Speech Dataset (YouTube-sourced)

63 labeled speech segments, split by **channel** (not by individual video) so the same speaker/recording can't appear in more than one split.

## Splits

- **train**: 63 segments
- **validation**: 0 segments
- **test**: 0 segments

## Transcript columns — read this before training

Each segment carries three transcript variants. They are NOT interchangeable:

- `text_original` — the YouTube caption text (if any) that overlapped this segment's time range, in whatever spelling/style convention that caption used.
- `text_whisper` — this project's fine-tuned Whisper model's own transcription of the segment's audio.
- `text` / `text_final` — the transcript actually selected as this segment's label (see `transcript_source` for which one won). **This is a mix of both conventions across the dataset** — some segments carry caption-style text, others carry Whisper-style text — so treat `text_final` as a single working label, not a normalized ground truth, and check `transcript_source`/`label_tier` per segment if your use case needs one consistent convention.

## Label trust tiers (`label_tier`)

`transcript_source` is collapsed into a coarser trust label:

- **gold** (`youtube_manual_caption`) — human-authored caption.
- **silver** (`youtube_auto_caption`) — YouTube's own ASR caption.
- **bronze** (`custom_whisper`) — this project's Whisper model's own output.

Silver and bronze are both machine-generated transcripts, and the Whisper model used for bronze was fine-tuned on a corpus with its own systematic error patterns — treat bronze as a separate, lower-confidence pool rather than blending it silently with gold for anything where transcript accuracy matters most.

- gold: 28 segments (44.4%)
- bronze: 35 segments (55.6%)

## Quality tiers (`quality_tier`)

`A`/`B`/`C` reflect a combination of audio cleanliness (clipping, VAD speech coverage, estimated SNR), forced-alignment confidence, and language-ID confidence. **These thresholds are heuristic starting points, not independently calibrated values** — if precision matters for your use case, spot-check a sample against human judgment before trusting a tier at face value.

`music_likelihood` and `overlapping_speech` are best-effort heuristic flags (not trained classifiers) that cap a segment at tier B rather than rejecting it outright — a false positive should cost a tier, not the clip.

## Content mix

- explainer content: 100.0%

KL divergence from the target content mix: 13.1223 (0 = matches target exactly).