praca-magisterska/project/readme.md

MUSIC GENERATION USING DEEP LEARNING

AUTHOR: CEZARY PUKOWNIK

Files:

• midi.py - code for data extraction, and midi convertion
• train.py - code for model definition, and training session
• generate.py - code for model loading, predicting ang saving to midi_dir
• settings.py - file where deafult settings are stored
• readme.md - this file

Directories:

• data/midi - directory where input midi are stored
• data/models - directory where trained models are stored
• data/output - directory where generated music is stored
• data/samples - directory where extracted data from midi is stored
• data/samples.npz - deprecated

How to use:

1. Use midi.py to export data from midi files

./midi.py [midi_folder_path] [output_path]

2. Use train.py to train a model (this can take a while)

./train.py [input_training_data] [model_save_path] [epochs]

3. Use generate.py to generate music from trained models

./generate.py [trained_model_path] [output_path] [treshold]