Acknowledgements

Waxrose uses third-party models and software. Waxrose is not affiliated with or endorsed by the authors of the models and software listed below.

Models

Downloaded from Hugging Face on first use; not modified by Waxrose. The models are provided by their authors "as is", without warranties (CC BY 4.0, section 5).

Parakeet Ultra

Core ML conversion by Fluid Inference (FluidInference/parakeet-ultra-coreml) of "parakeet-ultra" by moondream, a post-training of NVIDIA's "parakeet-tdt-0.6b-v3". Modified work (post-training; Core ML conversion and compilation).

License: Creative Commons Attribution 4.0 International (CC BY 4.0)

Speaker diarization (Segmentation, FBank, Embedding, PLDA)

Modified Core ML conversions by Fluid Inference (FluidInference/speaker-diarization-coreml) of pyannote's "speaker-diarization-community-1" pipeline, with WeSpeaker embeddings and PLDA parameters from Brno University of Technology / BUT Speech@FIT. Changes by Fluid Inference: Core ML conversion, fixed input shapes, mixed precision, separate FBank frontend.

License: CC BY 4.0

[citations from the README of FluidInference/speaker-diarization-coreml and the pyannote model card]

Software

FluidAudio 0.18.2

Copyright Fluid Inference. Apache License 2.0. Includes fastcluster (BSD 2-Clause) and VBx (Apache License 2.0). github.com/FluidInference/FluidAudio

[full license texts, as in Acknowledgements.txt]

Typefaces on this website

Trademarks

Apple, Apple Intelligence, Mac and macOS are trademarks of Apple Inc. OpenAI is a trademark of OpenAI. NVIDIA and Parakeet are trademarks or names of NVIDIA Corporation, mentioned for attribution only.