trpo_swimmer_ray_sampler

This is an example to train a task with TRPO algorithm.

Uses Ray sampler instead of on_policy vectorized sampler. Here it runs Swimmer-v2 environment with 40 iterations.

trpo_swimmer_ray_sampler(ctxt=None, seed=1)

tf_trpo_swimmer.

Parameters

  • ctxt (garage.experiment.ExperimentContext) – The experiment configuration used by Trainer to create the snapshotter.

  • seed (int) – Used to seed the random number generator to produce determinism.