Music Genre Classifier
Requires Python to run
pip install -r requirements.txt
When running the Jupyter networks, remember to set the Kernel properly.
The project does require fma_metadata and fma_small to be downloaded from the following links:
FMA Metadata ~342MB in Zipped format. 1.36GB unzipped. https://os.unil.cloud.switch.ch/fma/fma_metadata.zip FMA Small ~7.2GB Data https://os.unil.cloud.switch.ch/fma/fma_small.zip
The step_final_classify in notebooks can run standalone as it runs all the prior steps automatically as needed.
All python files or notebooks in the Notebooks folder can run standalone as well.
There are several secondary notebooks in the notebooks_secondary.
For the Yamnet Classifier please do the following:
To run Yamnet Model follow these steps,
-
access cscigpu server -> cd /research2/
-
create virtual environment using conda:
• conda create -n genrewave python=3.9 -y
• conda activate genrewave
• conda install -c conda-forge tensorflow=2.12 numpy=1.23.5 tensorflow-hub librosa pandas scikit-learn matplotlib joblib
• If encounter an error with missing tensorflow eg, “ModuleNotFoundError: No module named 'tensorflow’” run this command,
- conda install -c conda-forge tensorflow=2.12
FMA Metadata ~342MB in Zipped format. 1.36GB unzipped. https://os.unil.cloud.switch.ch/fma/fma_metadata.zip FMA Small ~7.2GB Data https://os.unil.cloud.switch.ch/fma/fma_small.zip