Instructions to use nvidia/parakeet-unified-en-0.6b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use nvidia/parakeet-unified-en-0.6b with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("nvidia/parakeet-unified-en-0.6b") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
onnx-web example
Hi @eschmidbauer ,
That's really nice! Thank you for the onnx and web support!
BTW, did you test the model's performance in your use cases?
BTW, did you test the model's performance in your use cases?
its just fantastic, thanks for this incredible release, its handling the bad pronunciations better.
Hi @eschmidbauer ,
That's really nice! Thank you for the onnx and web support!
BTW, did you test the model's performance in your use cases?
Yes, amazing speed and accuracy especially with currency, dates, numbers.
Great to hear this, @eschmidbauer .
I believe the Unified model is one of the best choices for VAD-based decoding -- it is robust to any segment sizes as small as 0.16s.
@eschmidbauer can you share the source code for the onnx-web example ?
thanks for the webUI
released the source code here @D3vShoaib
https://github.com/eschmidbauer/fireredvad.com