tutorial_vpg

This is an example to add a simple VPG algorithm.

class SimpleVPG(env_spec, policy, sampler)

Simple Vanilla Policy Gradient.

Parameters
  • env_spec (EnvSpec) – Environment specification.

  • policy (garage.tf.policies.StochasticPolicy) – Policy.

  • sampler (garage.sampler.Sampler) – Sampler.

init_opt(self)

Initialize optimizer and build computation graph.

train(self, trainer)

Obtain samplers and start actual training for each epoch.

Parameters

trainer (Trainer) – Experiment trainer.

tutorial_vpg(ctxt=None)

Train VPG with PointEnv environment.

Parameters

ctxt (ExperimentContext) – The experiment configuration used by Trainer to create the Snapshotter.