Instruction.txt

1. Download the dataset GTZAN/FMA from below links
    http://opihi.cs.uvic.ca/sound/genres.tar.gz
    https://github.com/mdeff/fma
    https://os.unil.cloud.switch.ch/fma/fma_small.zip
    https://os.unil.cloud.switch.ch/fma/fma_medium.zip
    https://os.unil.cloud.switch.ch/fma/fma_large.zip
2. Set the correct paths for different in config.py according to your directory structure.
3. Run preprocess_songs.py to divide the data set to divide into training/test and validation.
4. Run train_model.py to train the train_model
5. Run test_model.py to test the model
6. Run Generate_softmax.py to get the softmax layer output for all the songs.
7. Run create_tsne_embeddings.py to create the .json file of embeddings for demo.
8. For t-SNE/UMAP/PCA visualization open demo.html once data.json is generated.
9. For playing music in visualization server set up is required in the songs directory.