tf_agents.bandits.policies.policy_utilities.InfoFields

View source on GitHub

Strings which can be used in the policy info fields.

Class Variables

  • BANDIT_POLICY_TYPE = 'bandit_policy_type'
  • PREDICTED_REWARDS_MEAN = 'predicted_rewards_mean'
  • PREDICTED_REWARDS_SAMPLED = 'predicted_rewards_sampled'