NeMo
English
audio
speech
forced-alignment
reference-text-alignment
speech-processing
mfa
whisperx
python
metric:mae
Eval Results (legacy)
Instructions to use vertox-ai/speech-text-data-aligners with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use vertox-ai/speech-text-data-aligners with NeMo:
# tag did not correspond to a valid NeMo domain.
- Notebooks
- Google Colab
- Kaggle
Commit History
-
docs(aligners): finalize three-backend selection
- Add the strict 36-row MFA empirical result and post-final comparison evidence - Compare MFA, WhisperX, and NeMo across all 36 rows with pairwise review queues - Select MFA as the operational final backend after closing the full-corpus evidence gap - Publish the reviewed MFA 36-row annotation as the canonical final result - Preserve historical Stage 5 evidence and document remaining manual-gold limitations09db84f
-
fix(nemo): close empirical alignment validation
- Correct NeMo native word-CTM manifest handling - Close real GPU smoke, five-row, and full-36 evidence - Record the final three-aligner selection and final 36-row annotation - Preserve historical Stage 5 evidence and bounded architecture588788b