Java-native tools for running and fine-tuning language models - on-prem, air-gapped, or in the cloud. No hype, no SaaS, no Python.
Java-native distributed LLM inference and fine-tuning. Runs open-source GGUF models locally, in a cluster, or embedded as a JVM library. OpenAI-compatible REST API included. No Python. No NCCL. No InfiniBand required.
We are introducing Java in the ML field providing pure distributed JVM inference and training
On-prem first and network-ready. No mandatory cloud dependency, no SaaS lock-in. No Python, GIL, pure JVM orchestration using performance-proven engine
JFR metrics. Utilising custom flight-recorder events across hot paths - instrumentation-driven development that tells you exactly what is happening inside your runtime
Honest documentation. Solid testing. Clear issue tracking. Real benchmarks.
Fork the project, pick up an open board issue, move to QA and add a comment ref to your fork.
Please send us your performance report - just provide a Metrics summary, especially if you are using GPU. Mail to dev@ml.cab and list your: GPU card details, juno startup command, conversation log, JFR Metrics Summary section.
Tried Juno on your setup? Found a rough edge? Open an issue on the board. Specific, reproducible reports move things forward fastest.