- pip install ds-ctcdecoder
- get deepspeech scorer (deepspeech-0.9.3-models.scorer) from here
- get models from here
{
"stt": {
"module": "neon-stt-plugin-scribosermo",
"neon-stt-plugin-scribosermo": {
"model": "qnetp15/model_quantized.tflite",
"ds_scorer": "deepspeech-0.9.3-models.scorer"
}
}
}