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.
Synchronous Inference Deployment Flowchart¶
Asynchronous Inference Deployment Flowchart¶