Instructions to use ms180/librispeech_100h_e_branchformer with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- ESPnet
How to use ms180/librispeech_100h_e_branchformer with ESPnet:
unknown model type (must be text-to-speech or automatic-speech-recognition)
- Notebooks
- Google Colab
- Kaggle
File size: 312 Bytes
cb6ab54 | 1 2 3 4 5 6 7 8 9 10 | """Inference output helpers for LibriSpeech ASR recipes."""
def build_output(data, model_output, idx):
"""Build a dict of outputs for SCP writing."""
utt_id = data.get("utt_id", str(idx))
hyp = model_output[0][0]
ref = data.get("text", "")
return {"utt_id": utt_id, "hyp": hyp, "ref": ref}
|