Tutorials

Note

Tutorials will be added as the API matures.

Planned:

• Writing your first RunConfig

• Resuming a training run after a crash

• Wrapping a new torch_brain model in orion.models.*

• Authoring a custom Benchmark

• Setting up periodic in-training benchmarks

• Building a multi-stage TrainingPipeline (pretrain → finetune)

• Using the run registry to diff and compare runs

• Bursting from Polaris to cloud via SkyPilot