Help protect the Great Barrier Reef with TensorFlow on Kaggle Join Challenge

Module: tf_agents.train.triggers

Common triggers useful during training.

See interval_trigger.IntervalTrigger.

Classes

class PolicySavedModelTrigger: Triggers saves policy checkpoints an agent's policy.

class ReverbCheckpointTrigger: Checkpoints data from Reverb replay buffer.

class StepPerSecondLogTrigger: Logs train_steps_per_second.

ENV_STEP_METADATA_KEY 'env_step'