speaker-verification
audio
nemotron
tyger
experimental
ChristophSchuhmann's picture
Publish audited Robby source and completed Nemotron speaker-prediction experiments
0efd0f1 verified
Raw History Blame Contribute Delete
1.27 kB
Robby integration uses TYGER EARS, wire framing and playback concepts from
nineninesix-ai/tyger-s2s, source revision
97d011dcd4c8d2a117e54f00bf5b2d457cefa160, under Apache License 2.0.
https://github.com/nineninesix-ai/tyger-s2s
The Robby SVG and bounded rig.js are the separate MIT-licensed addon in
christophschuhmann/cartoon-face-generator, revision
210ca55ca464ef1be03017eb784b29aaa0a67ef8.
https://github.com/christophschuhmann/cartoon-face-generator/tree/main/robby
See static/LICENSE-MIT.txt. The original Python 101-keypoint rig is unchanged.
Third-party model and dataset licenses are separate from this integration code:
TYGER inference bundle: OpenMDW 1.1; SpeechBrain ECAPA: its own model card/license;
MediaPipe code: Apache 2.0; public voice profiles: CC BY 4.0 with source attribution.
No encoder, teacher, TTS or private microphone weights/data are distributed here.
The desk-robot procedural geometry and software 3D renderer are new integration
code. Face recipes/curve controls reuse the MIT robot addon above, with its
license retained. The user-supplied design PDFs, images and dense 3MF/FBX mesh
archives were inspected as references, but are not distributed or relicensed.
Nominal dimensions are a design proposal, not manufacturing certification.