Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Using deep learning to generate music in MIDI format.
| Date | Stars |
|---|---|
| 2026-07-31 | 553 |
| 2026-08-05 | 553 |
| 2026-08-06 | 553 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# MusicAutobot Using Deep Learning to generate pop music! You can also experiment through the web app - [musicautobot.com](http://musicautobot.com)  ## Overview Recent advances in NLP have produced amazing [results](https://transformer.huggingface.co/) in generating text. [Transformer](http://jalammar.github.io/illustrated-transformer/) architecture is a big reason behind this. This project aims to leverage these powerful language models and apply them to music. It's built on top of the fast.ai [library](https://github.com/fastai/fastai) ## Implementation **MusicTransformer** - This basic model uses [Transformer-XL](https://github.com/kimiyoung/transformer-xl) to take a sequence of music notes and predict the next note. **MultitaskTransformer** - Built on top of MusicTransformer, this model is trained on multiple tasks. * Next Note Prediction (same as MusicTransformer) * [BERT](https://github.com/google-research/bert) Token Masking * Sequence To Sequence Translation - Using chords to predict melody and vice versa. Training on multiple tasks means we can generate some really cool predictions (Check out this [Notebook](notebooks/multitask_transformer/Generate.ipynb)): 1. [Harmonization](http://musicautobot.com/#/predict/2b4f5e6613f366bad7b4f39c61be32b9) - generate accompanying chords 2. [Melody](http://musicautobot.com/#/predict/3087b73963aaa2bae62424808a251628) - new melody from existing chord progression 3. Remix [tune](http://musicautobot.com/#/predict/1bbfcb942133414a5664a35a7e7b5612) - new song in the rhythm of a reference song 4. Remix [beat](http://musicautobot.com/#/predict/71d7ff59f67fffa98614c841101e1b6b) - same tune, different rhythm ## How it works Details are explained in this 4 part series: * [Part I](https://towardsdatascience.com/creating-a-pop-music-generator-with-the-transformer-5867511b382a) - Creating a Pop Music Generator * [Part II](https://towardsdatascience.com/practical-tips-for-training-a-music-model-755c62560ec2) - Implementation details * [Part III](https://towardsdatascience.com/a-multitask-music-model-with-bert-transformer-xl-and-seq2seq-3d80bd2ea08e) - Multitask Transformer * [Part IV](https://towardsdatascience.com/how-to-remix-the-chainsmokers-with-a-music-bot-6b920359248c) - Composing a song with Multitask ## Example Notebooks 1. Play with predictions on Google Colab * [MusicTransformer Generate](https://colab.research.google.com/github/bearpelican/musicautobot/blob/master/notebooks/music_transformer/Generate_colab.ipynb) - Loads a pretrained model and shows how to generate/predict new notes * [MultitaskTransformer Generate](https://colab.research.google.com/github/bearpelican/musicautobot/blob/master/notebooks/multitask_transformer/Generate_colab.ipynb) - Loads a pretrained model and shows how to harmonize, generate new melodies, and remix existing songs. 2. MusicTransformer * [Train](notebooks/music_transformer/Train.ipynb) - End to end example on how to create a dataset from midi files and train a model from scratch * [Generate](notebooks/music_tranformer/Generate.ipynb) - Loads a pretrained model and shows how to generate/predict new notes 3. MultitaskTransformer * [Train](notebooks/multitask_transformer/Train.ipynb) - End to end example on creating a seq2seq and masked dataset for multitask training. * [Generate](notebooks/multitask_tranformer/Generate.ipynb) - Loads a pretrained model and shows how to harmonize, generate new melodies, and remix existing songs. 4. Data Encoding * [Midi2Tensor](notebooks/data_encoding/Midi2Tensor.ipynb) - Shows how the libary internally encodes midi files to tensors for training. * [MusicItem](notebooks/data_encoding/MusicItem-Transforms.ipynb) - MusicItem is a wrapper that makes it easy to manipulate midi data. Convert midi to tensor, apply data transformations, even play music or display the notes within browser. ## Pretrained Models Pretrained models are available as Mu
Excerpt of 8,260 characters
Read on GitHub
Andrew Shaw
476
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:8371bf608cfe2f06, llm:Repository description: 'Using deep learning to generate music in MIDI format.' Language: Jupyter Notebook. No topics provided.
matched fp:8371bf608cfe2f06, llm:Repository description: 'Using deep learning to generate music in MIDI format.' Language: Jupyter Notebook. No topics provided.
matched fp:8371bf608cfe2f06, llm:Repository description: 'Using deep learning to generate music in MIDI format.' Language: Jupyter Notebook. No topics provided.