A poject made for school, in the Cassiopée project of Télécom Sudparis.
- Requires python 3.10
- For python dependencies, please run :
pip install requirements.txt
If you can run make files, then you can do most of the work bby just calling the make script.
It should generate a brand new model and train it according to the databases available.
To avoid having it compute for 20 hours, please ask for a pre-computed xy folder. It is too heavy to store on github, but should make the script much faster.
- main application can be run by using
bash app.sh - to train the model, you can use
python3.10 codebase/musicHandler.py - a display of general functionalities of librosa can be run by using :
python3.10 codebase/__init__.pypython3.10 codebase/librosaTest.py
- Currently, is used only to test the jams reading capabilities using code found on the official github page of the database.
- classifier (WIP) can be run using
python3.9 codebase/classify.py. This code is meant to be integrated within the final product. As of now, reduces the entire database to a numpy matrix.