nexis-ml (Python)

nexis-ml is the Python engine for the ML Suite — a hobby-grade ML tool to create, train, and inspect small models on small data, with metrics streamed live to the Nexis terminal (or plain text in any other shell).

It’s the full-featured engine: the Nexis ML Lab ships the UI (panels, live charts, run browser) and spawns nexis-ml to do the actual work. It’s also useful standalone — every command works in a plain shell.

Install

pip install nexis-ml[torch]     # into the active virtualenv
pipx install nexis-ml[torch]    # or: on PATH in every shell

The Nexis ML Lab panel can install or download it for you automatically.

What it does

• Train tiny MLPs on tabular data (a spreadsheet / CSV).
• Train a tiny GPT you can watch learn to write.
• Train a small CNN over folders of images.
• Stream metrics live, save every run, and inspect checkpoints.

What it is not

Not an MLOps platform, not distributed training, not an LLM serving stack. Train a small model on your spreadsheet, watch the loss curve drop, and poke at the checkpoint.

Interop

nexis-ml speaks the shared NDJSON protocol and writes the shared run store, so its runs are interchangeable with the Rust engine nexis-ml-rs. Nexis renders runs from either engine identically.

No Python?

If you’d rather not install a Python/PyTorch toolchain, use the single-binary Rust engine nexis-ml-rs instead — the ML Lab panel auto-detects whichever is present.