Dl4j-spring-boot-starter?

DL4J
1

Hi DL4J team

First of all congrats for this project/initiative. As currently the entire java AI community seems to be hypnoitized by LLM’s gen AI stuff, it is comforting to see that there are still ppl interested in developing the frameworks for core AI algorithms.

I’ve been working on a small Spring Boot starter to simplify using Deeplearning4j in backend applications, and I wanted to share it with you and get some feedback ( more over the idea of having the possibility of configuring neural nets in yaml, using a spring boot starter - because the code I have currently is a minimum beta-like version )

The idea is to make DL4J more “Spring-native” by allowing neural networks to be defined declaratively via YAML instead of wiring everything in Java code.

Example:

dl4j:
  input-size: 3
  model:
    layers:
      - type: dense
        size: 16
        activation: relu
      - type: output
        size: 1
        activation: sigmoid
        loss: mse

The starter provides:

  • auto-configuration for DL4J

  • YAML-driven model builder

  • simple training + prediction support

  • model save/load functionality

It’s currently an early v0.1.0 beta, focused on basic feedforward networks and developer experience.

Repo: https://github.com/darius1973/spring-boot-starter-dl4j

I’d really appreciate any feedback, especially:

  • whether this aligns well with DL4J design

  • if something like this could be useful for the community

  • any suggestions for improvements or integration

If there already is a similar topic - let me know and I’ll move this msg there

Thank you