ray.rllib.policy.eager_tf_policy_v2.EagerTFPolicyV2.action_distribution_fn#

EagerTFPolicyV2.action_distribution_fn(model: ModelV2, *, obs_batch: numpy.array | jnp.ndarray | tf.Tensor | torch.Tensor, state_batches: numpy.array | jnp.ndarray | tf.Tensor | torch.Tensor, **kwargs) → Tuple[numpy.array | jnp.ndarray | tf.Tensor | torch.Tensor, type, List[numpy.array | jnp.ndarray | tf.Tensor | torch.Tensor]][source]#

Action distribution function for this Policy.

Parameters:

  • model – Underlying model.

  • obs_batch – Observation tensor batch.

  • state_batches – Action sampling state batch.

Returns:

Distribution input. ActionDistribution class. State outs.