#Live ASR Engine
I have implemented (not from scratch) LiveASREngine using whisper using the following codebase:
https://github.com/oliverguhr/wav2vec2-live
Installation requirements for whisper (developing anaconda environment):
- conda create --name tips -c nvidia -c pytorch -c huggingface -c defaults -c conda-forge python autopep8 pytorch torchvision torchaudio
- pytorch-cuda=11.7 transformers pyaudio webrtcvad rx halo onnx onnxruntime pyctcdecode
- pip install pysoundfile
- conda install -c conda-forge ffmpeg