This project is the generating music using self attention gan.
Out result is here.
Before training the model, we convert midi file to grayscale midi image.
training the gan without thresh hold / training the gan applying 0.3 thresh / training the gan applying 0.5 thresh
GTZAN Dataset - Music Genre Classification
Link : https://www.kaggle.com/andradaolteanu/gtzan-dataset-music-genre-classification
Classical Music MIDI
Link : https://www.kaggle.com/soumikrakshit/classical-music-midi
Magenta Datasets
Link : https://magenta.tensorflow.org/datasets/
Music MIDI Collection
Link : https://www.kaggle.com/programgeek01/anime-music-midi
youtube link: https://youtu.be/HZKP8TjhIk8
youtube link: https://youtu.be/apQXf4pcT2I
youtube link: https://youtu.be/QxlSpPadJnE
youtube link: https://youtu.be/5XvYgP_0rw0
youtube link: https://youtu.be/nfh7_8glCIM
youtube link: https://youtu.be/dNytCDL6Ss4