johnnyrobotai/johnnylingo-jingo-wav2vec2-xlsr-53-espeak-ct2

Johnny Lingo Pronunciation CT2 Model

This repository hosts the external CTranslate2 acoustic model artifact used by johnnylingo-pron. The Python wheel does not bundle this model.

• Intended repo id: johnnyrobotai/johnnylingo-pron-wav2vec2-xlsr-53-espeak-ct2
• SDK engine version: 0.2.1
• Default calibration: v1.1-no-tutor-2026.07
• Calibration provisional: True
• Source model: facebook/wav2vec2-xlsr-53-espeak-cv-ft
• Source model license: Apache-2.0
• Archive: wav2vec2-xlsr-53-espeak-ct2-v0.2.1.zip
• Archive size: 273800331 bytes
• Archive sha256: c81c17bce9f311c0aed217312ce4c49327b27d09b6acc9004ad582cee8d77a2a
• Generated at: 2026-07-06T00:36:55.461626+00:00

Files In The Archive

• model.bin: 317779386 bytes, sha256 911f57e20cde2b47eb2996cc53608b108074360280b156b07616eb5effa2975e
• config.json: 3 bytes, sha256 ca3d163bab055381827226140568f3bef7eaac187cebd76878e0b63e9e442356
• vocabulary.json: 5037 bytes, sha256 945dffdc7990fa8c553bdf2f52d4f11f87b1eeb24758dcdad0873bc9208c36db

Important Caveats

• This is a pronunciation-learning support model, not a clinical or academic assessment instrument.
• Default thresholds are channel-sensitive and not human-label validated.
• The default calibration remains provisional: true.
• Host apps must verify the archive checksums before enabling scoring.

Install

pron-engine-model ensure \
  --model-dir /models/wav2vec2-xlsr-53-espeak-ct2 \
  --hf-repo johnnyrobotai/johnnylingo-pron-wav2vec2-xlsr-53-espeak-ct2 \
  --hf-filename wav2vec2-xlsr-53-espeak-ct2-v0.2.1.zip