mrx-link 2.0.0.dev2

A simple AI/ML modeling tool, Link™ ensures a smoother flow and a better experience throughout the model development cycle.

Link

A simple AI/ML modeling tool developed as an extension of JupyterLab, Link ensures a smoother flow and a better experience throughout the model development cycle.

Discussions | Docs | Link SDK | Release Note | Website | Youtube | Technical Support

Usage Environment

Supported Platform	Linux (AMD64, ARM64) / macOS (Intel, Apple Silicon) / Windows
Python Version	3.8, 3.9, 3.10
Requirements	Git

Getting Started

We recommend installing our program in a new virtual environment.

Step 1. Install Link with the following command.

python3 -m pip install --upgrade mrx-link

Step 2. After installation, run JupyterLab.

python3 -m jupyterlab

If you wish to install it as a Desktop Application, please download from here.
And if you want to try Link on your browser, please click here.

Main Features

Pipeline Creation

Create a pipeline with notebook cells to improve code readability and reproducibility.

Caching Management

Store execution results for respective cells to minimize redundant cell executions.

Remote Resources

Optimize the use of server resources by using separate resources for different pipeline components.

Hyper-Parameter Optimizer

Link provides a hyper-parameter feature, which works to find optimal hyper-parameters in an automated manner.

Version Control

Link provides a version control feature, where you can easily view changes to your pipeline and merge conflicts.

Link SDK

Link provides an SDK, where you can easily run the Link notebook file with the command line interface, and create, run, export pipelines with Python code. (Learn more)

Other Features

• Convert your pipelines to Kubeflow (.yaml) pipelines.
• Convert your components to executable Python (.py) code.

Examples

• Link Introduction
• Deep Q-Network Reinforcement Learning for the CartPole Environment
• Denoising Autoencoder with PyTorch
• Iris Classification
• MNIST Image Classification with PyTorch
• MNIST Image Generation with VAE
• Sequential Data Classification
• Spiral-Distributed Data Classification
• Titanic Data EDA and Modeling
• XGBoost Example

Help and Support

• Product Page: https://link.makinarocks.ai
• Official Documentation: https://makinarocks-link.readme.io/docs
• Release Notes: https://makinarocks-link.readme.io/changelog
• Technical Support: Contact Technical Support for any questions or issues

Team

• Chan Baek (@chan-python)
• Daesung Kim (@JakeisDaeSung)
• Dohu Jeon (@adc0612)
• Eunyoo Byeon (@EunyooByeon)
• Hongji Kim (@HongJi-Kim)
• Hooncheol Shin (@hunhoon21)
• Hwiyeon Cho (@hwyncho)
• Hyeonseok Shin (@hyeonseok)
• Jaeseong Choe (@sorrychoe)
• Jonghyeok Lee (@daisy0y)
• Jungmin Park (@jungminpark1234)
• Ryeong Yi (@ryeongYi)
• Sangwoo Shim (@borishim)
• Woojin Cha (@Xena123kiki)
• Yongjin Shin (@YongjinShin306)

Contributors

• Byoungwan Kim (@bwkim71)
• Hwan Chang (@aiworld-git)