Exploy Documentation¶
Welcome to the Exploy documentation!
Exploy is a toolkit for exporting and deploying reinforcement learning policies. It provides tools for exporting RL environments and policies from simulation frameworks and deploying them using ONNX Runtime for efficient inference.
Contents:
- Exploy
- Exporter Tutorial
- Prerequisites
- Step 1 — Define the Data Source and the Environment
- Step 2 — Implement the Exportable Environment Adapter
- Step 3 — Register Inputs, Outputs, and Memory
- Step 4 — Create an Actor
- Step 5 — Export to ONNX
- Step 6 — Load and Run the Exported Model
- Step 7 — Evaluate the Export
- Advanced: Using Torch Modules in Observations
- Advanced: RNN-Based Actors
- Putting It All Together
- Controller Tutorial
- API Reference