Original code is provided by the team from ByteDance [pdf, code]
- Install dependencies:
pip install -r requirements.txt
- Different versions of packages may result in errors when running.
-
You may want to change the designated GPU for transcription in
inference.py
-
If you have multiple GPUs to run, you can either:
- Uncomment lines 58-62 in
piano_transcription_inference/inference.py
to run parallel on multiple GPUs for faster transcription - Run with the original package directly:
pip install piano_transcription_inference
- Uncomment lines 58-62 in
-
To start transcribing with pre-trained model:
python transcribe.py