Project idea: use Deep Learning to ear the difference between instruments
Developed with Python 3.7 and should be compatible with Python 3.6+
Developed on macOS and Ubuntu. Not sure about compatibility with other systems
Install the requirements with the usual:
pip3 install -r requirements.txt
To adapt the configuration file to your environment, you need to create a local.py
file
in the config
directory, next to the default.py
config file.
You can overwrite every item of the default config, maybe paths, log config,...
You should be able to run the code with:
./src/main.py
You may have to run export PYTHONPATH=$(pwd)
beforehand for the imports to work.
You can also run the realtime recognition with:
./src/realtime.py
At the end of the project we aldo pushed the notebook with tests, MLP and CNN, named Tests.ipynb
You can view its content easily with Tests.html
, or open it in jupyter notebook.