CaptionBuddy models
Redistributed copies of open speech models used by CaptionBuddy, the WorshipBuddy live-captioning app. The app downloads these files and checks each one against a sha256 recorded in its model catalog. Every file keeps its original licence.
| Path | Source | Licence | Changes |
|---|---|---|---|
indic-conformer-600m/ |
ai4bharat/indic-conformer-600m-multilingual | MIT | none (the ONNX export from assets/; the CTC head and TorchScript preprocessor are left out) |
ggml-medium-ml-thennal-q5_0.bin |
thennal/whisper-medium-ml | Apache-2.0 | converted to whisper.cpp ggml, quantized to q5_0 |
voxlingua107-lid/ |
speechbrain/lang-id-voxlingua107-ecapa @ 0253049 | Apache-2.0 | exported to ONNX whole (features → classifier), STFT rewritten as a convolution; export.py reproduces it |
ggml-turbo-ml-betty.bin |
BettySara/whisper-large-v3-malayalam-FT | Apache-2.0 | converted to whisper.cpp ggml |
See LICENSE.md for the licence texts.