---
pipeline_tag: text-to-audio
tags:
- audio
- mozart
- piano
- concert
- music
- generation
- open
- open-source
- cpu
- local
- tiny
- small
---

# Welcome to TinyMozart v2 85M!
This model is an unconditional MIDI classic piano music generator trained on the Google MAESTRO V3 MIDI dataset.
<br>
Note, that this is the **second** version of our TinyMozart model series.

## Sample
<audio controls src="https://cdn-uploads.huggingface.co/production/uploads/697f2832c2c5e4daa93cece7/oIWWZdm1AKhht_uMSnOvq.mpga"></audio>

## How to use it
To use this model and generate your own music, just download `model.pt` and `use.py` and run it with Python3. Have fun :D

## Training code
Use `train.py`.

## Training data
We trained on [https://storage.googleapis.com/magentadata/datasets/maestro/v3.0.0/maestro-v3.0.0-midi.zip](https://storage.googleapis.com/magentadata/datasets/maestro/v3.0.0/maestro-v3.0.0-midi.zip).

## Limitations
We know, this model isn't perfect, but it's a good beginning to understand how to generate music with AI models.