Afrikaans learner: Goldfish (unscripted)
This 2.49-minute, 33-sentence conversation between a native English speaker learning Afrikaans revolves around goldfish. Gemini supplied fairly good information, but also used two words from the contaminated Afrikaans dataset that it only seems able to spell, not say, and it switches to a Dutch accent frequently.
Every package is hand-curated by a native speaker, and includes:
- original audio.wav
- raw Whisper X and Faster Whisper output
- human edits - transcripts and English translation in:
- .txt
- .json
- .csv
- ML package with accurate timestamping and perfect
single-sentence cuts side-by-side with human-curated transcript snippets.
- detailed, human-curated manifest
- diff report
- speaker profile
- license