bunsen-asr
Speech to text on bunsen, by profile (English, multilingual, translated to English) or by model family: Parakeet TDT (FermionResearch's Phonon-2 by default), Qwen3-ASR (Phonon-1), and Whisper; and who spoke when, through NVIDIA's Sortformer diarizer.
Activity
- Latest release
- 6d ago
- Total releases
- 3
- Cadence
- ~3 days
- Last 12 months
- 3
Reach
- Downloads
- 35
Details
- License
- MIT OR Apache-2.0
- First release
- Oct 17, 2026
Releases