Realistic MIDI Playback
Project Details
Drawing parallels from text-to-speech transcription, this model uses a transformer architecture that generates mel spectrograms from MIDI files without control values, aimed at relasitic playback, and is trained on the MAESTRO dataset.
Project Information
- Category: ML
- Stack: python, pytorch
