Refactor sound_generator.py to load models on GPU if available and streamline audio generation process
f8bd524
Curinha
commited on