🚀 skainet-jupyter: Kotlin-Powered Deep Learning in Jupyter

Experience the future of machine learning development with Kotlin’s type safety, performance, and expressiveness in interactive Jupyter notebooks. Build neural networks, train models, and experiment with AI using a modern, statically-typed language.

🎯 What is skainet-notebook?

skainet-jupyter brings the power of the skainet deep learning framework to Jupyter notebooks, combining Kotlin’s developer-friendly features with interactive data science workflows. Whether you’re prototyping neural networks, analyzing datasets, or building production-ready ML models, this project provides a robust foundation for AI development.

✨ Current Status & Capabilities

🟢 Active Development - This project is actively maintained and growing

🧠 Neural Networks & Deep Learning

• ✅ Multi-Layer Perceptrons (MLP) - Fully functional feed-forward networks

💻 Developer Experience

• ✅ Type-Safe ML Code - Kotlin’s compile-time safety prevents runtime errors

• ✅ Interactive Development - Real-time experimentation in Jupyter notebooks

• ✅ IntelliJ IDEA Integration - First-class IDE support with debugging

• ✅ Rich Documentation - Comprehensive examples and tutorials

Using with Jupyter Notebook

Option 1: IntelliJ IDEA

You can create Kotlin notebooks directly in IntelliJ IDEA using one of these methods:

1. Within a Project:

   • Right-click on source root/folder in Project view

   • Select New → Kotlin Notebook

2. Scratch Notebook:

   • Press kbd:[Cmd+Shift+N] (macOS) or kbd:[Ctrl+Alt+Shift+Insert] (Windows/Linux)

   • Select Kotlin Notebook

3. From Welcome Screen:

   • Navigate to Kotlin Notebooks tab

   • Click New Notebook

Option 2: Jupyter Installation

1. Install Jupyter and Kotlin Kernel:

   $ pip install jupyter
   $ jupyter kernelspec install kotlin

2. Create a New Notebook:

   • Launch Jupyter: jupyter notebook

   • Select "New" → "Kotlin"

3. Basic Usage:

   // Example code
   val data = DataFrame.readCsv("data.csv")
   data.head(5).show()

Local Development Setup

To build and run the jupyter notebooks locally, you need to first build and publish the skainet library dependency to your local Maven repository.

Prerequisites

• JDK 17 or higher

• Git

Building skainet Library Locally

1. Clone the skainet repository:

   git clone https://github.com/sk-ai-net/skainet.git
   cd skainet
   # Checkout specific working commit (WIP)
   git checkout 6f61c63

2. Build and publish to Maven local:

   ./gradlew publishToMavenLocal

3. Navigate back to this project and build:

   cd ../skainet-jupyter-gh
   ./gradlew publishToMavenLocal

Consuming from Kotlin Notebook

Since we have publishing skainet wrapper for Kotlin Notebook into local maven repository, we can now consume by adding dependency on the library from localMaven like follows:

USE {
    repositories {
        mavenLocal()
    }
    dependencies {
        implementation("sk.ainet:kotlin-notebook:0.0.1")
    }
}