Skip to content

localml

A local ML experimentation platform that runs entirely on an Apple Silicon workstation. It implements the core architecture of a production ML platform at local scale: a Python SDK, framework adapters, experiment tracking, a model registry, artifact storage, evaluation jobs, and local model serving.

Install

pip install localml

Quickstart

import localml as ml

ml.configure(api_url="http://localhost:8000", token="local-dev-token")

with ml.start_run(project="local", config={"model": "tiny-llm"}):
    ml.log_params({"batch_size": 4})
    ml.log_metrics({"baseline_accuracy": 0.82})

Architecture

flowchart LR
    User[SDK / CLI / Notebook] --> API[FastAPI control plane]
    API --> MLflow[MLflow<br/>tracking + registry]
    API --> DB[(Postgres<br/>metadata)]
    API --> Store[(MinIO<br/>artifacts)]
    API --> Queue[Redis<br/>job queue]
    API --> Serving[Serving proxy<br/>OpenAI-compatible: Ollama / MLX]
    Queue --> Worker[Worker]
    Worker --> Store
    Worker --> DB

The control plane is the source of truth for platform metadata. MLflow holds experiment tracking state and model registry records, MinIO stores artifacts, and Redis carries transient evaluation job state for the worker.