Melody Generation with Transformers - Generative Music AI
Valerio Velardo - The Sound of AI via YouTube
Overview
Implement a transformer architecture in TensorFlow to generate melodies end-to-end in this comprehensive 1-hour 12-minute tutorial. Explore melody encoding, dataset preprocessing, model building, training, and inference. Dive into the code overview, examine the melody dataset, and understand the MelodyPreprocessor. Learn about the Transformer model, including its Encoder, EncoderLayer, Decoder, and DecoderLayer components. Discover how to create a MelodyGenerator and write a train script. Run the training process and experiment with the code to generate unique melodies using AI.
Syllabus
Intro
What we'll build
Code overview
Melody dataset
MelodyPreprocessor
Transformer
Encoder
EncoderLayer
Decoder
DecoderLayer
MelodyGenerator
Train script
Runing the train script
Experimenting with the code
Taught by
Valerio Velardo - The Sound of AI