Dart NeoML

Dart NeoML is a port of the cross-platform library NeoML made by the company ABBYY.

Index

• Dart NeoML
  • Index
  • About
  • Motivation
    • Why was NeoML chosen?
  • Features
  • Getting started
  • Usage
  • Contribution

About

This package will help you easily create and train and use machine learning and neural network models in the Dart language based on the NeoML framework

Motivation

Among the neural network and machine learning libraries on dart that worked with C++ under the hood. Since C++ achieves high performance, with dart-ffi this can be achieved on the basis of the NeoML end-to-end framework.

Why was NeoML chosen?

• Cross-platform framework that works on all known platforms Windows, Linux, macOS, Android, iOS (which have been fully supported since Flutter 3.0.0).
• High performance. This was discussed by NeoML in an article on Habr.
• Usabillity for mobile. It is possible to serialize the trained model and use it on your mobile devices.
• Backend. Machine learning models are also used on the backend, where Dart also takes its place (for example, AppWrite). In the same Django framework, you can also use machine learning models, since Python is a favorable environment for this.

Features

TODO: List what your package can do. Maybe include images, gifs, or videos.

Getting started

TODO: List prerequisites and provide or point to information on how to start using the package.

Usage

TODO: Include short and useful examples for package users. Add longer examples to /example folder.

const like = 'sample';

Contribution

Create PR or offers in Issues. Become a part of the OpenSource community 😊

But before you start, you need to deploy the project locally.

To do this, you need to follow a number of steps:

1. Clone project with submodules:

git clone --recurse-submodules -j8 https://github.com/shigomany/dart-neoml.git

2. If you have already cloned the repository, you must run this command to clone the submodules.

git submodule update --init --recursive