An attempt at implementing a GAN similar to the one explained in https://www.researchgate.net/publication/340640304_A_Symbol-level_Melody_Completion_Based_on_a_Convolutional_Neural_Network_with_Generative_Adversarial_Learning .
All credit to the authors of the paper:
- Kosuke Nakamura
- Takashi Nose
- Yuya Chiba
- Akinori Ito
The data set used for training consists of MAESTRO midi files: https://magenta.tensorflow.org/datasets/maestro.
To create the environment, run conda env create -f environment.yml
.
To track training, tensorboard --logdir lightning_logs
.
To launch training, python run_training.py -v version_name
.