Icefall Streaming Zipformer for Estonian
This model is ONNX version of an Icefall Zipformer streaming ASR model for Estonian. It is trained on ~1300 h of Estonian transcribed speech (https://cs.taltech.ee/staff/tanel.alumae/data/est-pub-asr-data/), ~3000 h of automatically transcribed Estonian TV broadcasts (transcribed using a strong offline ASR model) and ~500 h of English podcasts and YouTube data, transcrined using Whisper-turbo.
Use sherpa-onnx to do ASR:
E.g., under Linux, using card 3 as input device:
sherpa-onnx-alsa --encoder=encoder.onnx --decoder=decoder.onnx --joiner=joiner.onnx --tokens=tokens.txt --decoding-method=modified_beam_search plughw:3,0
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support