Quickstart

30-Second Path

uv tool install --python 3.12 yolomatic
yolomatic

Choose Configure Model, select a dataset, generate a config, then choose Train Model.

Repository Path

uv sync
uv run yolomatic

Useful follow-up commands:

uv run yolomatic-train
uv run yolomatic-predict
uv run yolomatic-benchmark
uv run yolomatic-tensorboard

Expected Dataset Layout

datasets/my_dataset/
  data.yaml
  train/images/
  train/labels/
  valid/images/
  valid/labels/
  test/images/
  test/labels/

COCO JSON datasets are used for Detectron2, SAM 3.1 fine-tuning, and benchmark validation sets.

Related pages: First training run, Datasets, Configuration.