OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Deepjazz

Deepjazz is an experimental Python project that trains a two-layer LSTM on MIDI input and generates jazz music with Keras and Theano. It includes MIDI preprocessing and a command-line generator.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Deepjazz — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Deepjazz to explore or adapt an experimental MIDI-based jazz-generation implementation, especially when you can work with its Python preprocessing code.

Who it is for

  • Developers studying LSTM-based music generation in Python.
  • Musicians or researchers experimenting with jazz generation from MIDI input.

Documented uses

  • Generate jazz music from a supplied MIDI file.
  • Inspect a Python implementation combining MIDI preprocessing with a two-layer LSTM.
  • Experiment with CPU or NVIDIA CUDA GPU execution of the generator.
Before you choose

Requirements & limits

  • Keras, Theano and music21 are listed dependencies.
  • The README provides separate CPU and GPU commands; GPU support is described for NVIDIA cards with CUDA.
  • Using MIDI files beyond the existing preprocessing requires modifying preprocess.py to select the melody part.
  • The README states that the project is no longer actively developed.
  • The code license is Apache-2.0.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Run the documented generator on CPU

Start the Python jazz-generation script using the README’s CPU command.

Before you start

  • Python project files from the Deepjazz repository
  • Keras, Theano and music21 dependencies
  • A supplied MIDI input for the project’s learning workflow

Steps

  1. Install the listed dependencies.

  2. From the project environment, run the command below, replacing the placeholder with an epoch count.

    Documented command
    python generator.py [# of epochs]

Expected outcome

The documented generator command starts the CPU workflow for jazz generation.

Guide limitations

  • The README says preprocessing must be modified to use other MIDI files; the relevant melody MIDI part must be selected.
  • The project is no longer actively developed.
Documentation 1 ↗
Source-based practical guide

Run the documented generator with NVIDIA CUDA

Start the Python jazz-generation script through Theano’s documented GPU configuration.

Before you start

  • Python project files and the Keras, Theano and music21 dependencies
  • An NVIDIA card with Theano’s CUDA backend

Steps

  1. Install the listed dependencies and prepare the NVIDIA CUDA setup required by Theano.

  2. Run THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python generator.py [# of epochs], replacing the placeholder with an epoch count.

Expected outcome

The documented command starts the generator with Theano configured for GPU execution.

Guide limitations

  • The README identifies formal GPU support only for NVIDIA cards using the CUDA backend.
  • Other MIDI files require changes to preprocess.py to select the relevant melody part.
  • The project is no longer actively developed.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
10 October 2026
Original website title
jisungk/deepjazz · Official homepage title · retained in its source language
How we review sourcesSuggest a correction