The core of Agent's interation. Implements Reinforcement Learning APIs to enable interaction with AI Traineree.
The Agent is provided as a Docker image. Please see GitHub's Agents Bar - Agent container repository to access the image.
In order to build the development image, including the full Pytoch dependency, run
docker build --target dev . -t agent_service:dev_version
In order to build the production image, including the Pytoch+cpu dependency (no support for CUDA), run
docker build --target prod . -t agent_service:latest