Inference & Deployment

  • We provide an inference deployment system that supports both synchronous inference and asynchronous inference via SimpleRTC.

  • By simply wrapping your model as a Server, you can enable communication with deployment nodes, facilitating local or even remote (cross-machine) synchronous/asynchronous inference deployment.

  • The system integrates with the app to allow inference start/stop and takeover, supporting workflows such as DAgger-style data collection.

sync deployment

Synchronous Inference Deployment Flowchart

async deployment

Asynchronous Inference Deployment Flowchart