asigalov61 / tegridy-tools

Symbolic Music NLP Artificial Intelligence Toolkit

tegridy-tools

---

Symbolic music artificial Intelligence toolkit for rapid prototyping, design and evaluation of the symbolic music AI architectures, systems and models

---

Install

Recommended and required for the latest version of the modules/files:

!git clone --depth 1 https://github.com/asigalov61/tegridy-tools

---

Hassle-free automatic install , setup, and all core modules imports

!curl -O https://raw.githubusercontent.com/asigalov61/tegridy-tools/main/install/setup.sh
# !wget https://raw.githubusercontent.com/asigalov61/tegridy-tools/main/install/setup.sh
!chmod +x setup.sh
!bash setup.sh
!rm setup.sh

---

After install you can import/re-import modules like so...

Standard Python auto-imports/re-imports

# CPU setup and imports
!python ./tegridy-tools/install/setup.py

# GPU setup and imports
!python ./tegridy-tools/install/setup_cuda.py

Google Colab auto-imports/re-imports

# CPU setup and imports
%run ./tegridy-tools/install/setup.py

# GPU setup and imports
%run ./tegridy-tools/install/setup_cuda.py

Manual imports/re-imports

import TMIDIX
import TMELODIES
import HaystackSearch
import midi_to_colab_audio
from x_transformer_1_23_2 import *
import random

---

Quick Start Guide

Start by checking out example code snippets and scripts

Or you can check out many practical, specific and detailed Jupyter/Google Colab Notebooks

---

Detailed core modules documentation

tegridy-tools docs

---

@inproceedings{lev2024tegridytools,
    title       = {tegridy-tools: Symbolic Music NLP Artificial Intelligence Toolkit},
    author      = {Aleksandr Lev},
    booktitle   = {GitHub},
    year        = {2024},
}

---

Project Los Angeles

Tegridy Code 2024

tegridy-tools's Issues

How can I train a model ?

Hi, I come here because the author of this repo : https://github.com/asigalov61/Optimus-VIRTUOSO/blob/main/optimus_virtuoso_composer.py

is not online anymore, and I'm trying to fork his code to experiment further, but I don't know how to train a model or what kind of format do I need to use ( as in MIDI files ) to train new models. for now I have adapted these functions to work locally but I need to know how to create a dataset that is usable ( I'm new to ML ) Would you please guide me a little / give some direction so I can proceed ?

Thank you very much