Agent Zoo

Numerous pre-built and trained agents are hosted at zoo/policies . These agents should be installed prior to being used in SMARTS environments or benchmarks.

$ cd <path>/SMARTS
# Install a zoo agent.
$ scl zoo install <agent path>
# e.g., scl zoo install zoo/policies/interaction_aware_motion_prediction

Note

To build a wheel, execute:

$ scl zoo build <agent path>
# e.g., scl zoo build zoo/policies/interaction_aware_motion_prediction

Contribute agents

Note

This section is only for contributing to the agent zoo.

First, add the new agent to zoo/policies. It should contain

  • package setup file,

  • inference code with pre-built model, and

  • algorithm explanation.

Then, register the newly added zoo agent in zoo/policies/__init__.py.

Available zoo agents

Agent locator and path

Benchmark or Environment

Action space

Source

Remarks

zoo.policies:interaction-aware-motion-prediction-agent-v0
zoo/policies/interaction_aware_motion_prediction

driving_smarts_2022==0.0

TargetPose

code

Contributed as part of NeurIPS 2022 Driving SMARTS competition.

zoo.policies:control-and-supervised-learning-agent-v0
zoo/policies/control_and_supervised_learning

driving_smarts_2022==0.0

TargetPose

code

Contributed as part of NeurIPS 2022 Driving SMARTS competition.

zoo.policies:discrete-soft-actor-critic-agent-v0
zoo/policies/discrete_soft_actor_critic

driving_smarts_2022==0.0

TargetPose

code

Contributed as part of NeurIPS 2022 Driving SMARTS competition.