djl
An Engine-Agnostic Deep Learning Framework in Java
What is djl?
An Engine-Agnostic Deep Learning Framework in Java
Overview of djl
What it is
Deep Java Library (DJL) is an open-source, high-level, engine-agnostic Java framework for deep learning. It is designed to be easy to get started with and simple to use for Java developers. DJL provides a native Java development experience and functions like any other regular Java library. It is engine-agnostic, meaning users do not have to choose between deep learning engines when creating projects and can switch engines at any point. DJL also provides automatic CPU/GPU choice based on hardware configuration.
Key features
- Engine-agnostic: supports multiple deep learning engines and allows switching between them.
- Native Java development experience: integrates with standard Java tools and IDEs.
- Automatic CPU/GPU selection based on hardware configuration.
- Ergonomic API interface designed to guide best practices for deep learning tasks.
- Supports both inference and training workflows with simple pseudocode examples.
- Includes a model zoo for pre-trained models.
Use cases
DJL is suitable for Java developers who want to build, train, and deploy deep learning models without needing to be machine learning experts. It can be used for tasks such as object detection (e.g., using a pre-trained ResNet50 model) and training neural networks (e.g., an MLP on the MNIST dataset). It is designed to integrate models into Java applications easily.
Getting started
To build from source, check out the code and use Gradle:sh# for Linux/macOS:./gradlew build# for Windows:gradlew buildTo skip unit tests during build:sh# for Linux/macOS:./gradlew build -x test# for Windows:gradlew build -x testFor importing into Eclipse, run ./gradlew eclipse (Linux/macOS) or gradlew eclipse (Windows) and then import the Gradle project in Eclipse.
License and pricing
This project is licensed under the Apache-2.0 License. It is open-source and free to use.
Based on the project's public README and repository metadata.