Start training

Pick a dataset, a model, and an input mode. The examples are organized into short pages so you can copy one config and run its command.

Step	Page	What you do
1	Install	Install Fusion with [visin]; optionally connect to Visin
2	ZOD, Waymo, or Iseauto	Download and prepare one dataset
3	Choose a model	Select RGB, LiDAR, or fusion; copy or download its JSON
4	Run and inspect the outputs	Train, test, visualize, and benchmark

First example

After preparing ZOD, run:

ZOD + CLFTv2 + fusion

visin-fusion run -c configs/examples/zod/clftv2/fusion.json --upload --benchmark-device cuda

Use --benchmark-device cpu on a CPU host. With .env configured, Fusion reports to Visin; otherwise outputs remain local. All examples run one epoch and use separate output directories. No extra experiments repository is needed.

Want a longer experiment? Customize a run.