Yes, you can train/fine-tune models on your own voice with Coqui