ML in Java, YES it's possible!
summary
Claims Java can train and run models end to end and shows concrete tools: JSR381 (Visre) builds a Kagel hot dog classifier in ~10 lines plus config, trains a one‑layer CNN on an M1 32GB machine and exports a usable model; Deeplearning4J uses ND4J, offers a Deeplearning4J model zoo, UI server, transfer learning and low‑level control but verbose code; djl delegates to PyTorch/TensorFlow engines and provides a Docker model server; Tribuo targets enterprise pipelines and classical ML (classification, regression, anomaly detection).
AI-written summary · about
chapters (9)
- 0:00Introduction and speaker background
- 7:30Machine learning fundamentals
- 12:46Challenges of ML in Java
- 14:35JSR 381 VisRec framework
- 24:54Eclipse Deeplearning4j framework
- 33:47AWS DGL framework overview
- 40:49Oracle Tribuo framework
- 43:41Choosing the right Java ML framework
- 45:08Conclusion and Q&A
related talks
47mML in Java, YES it's possible!
50mML in Java, YES it's possible!
58mPatterns, Predictions and Programming
43mWhy Java and the JVM Will Dominate the Future of Machine Learning, AI, and Big Data
48mML on the JVM
25mDeeplearning4j: How to run AI workloads on Jakarta EE compliant servers
34mBuilding Machine Learning-powered Applications with JS
01h 01mDeep Learning and Java