ray.rllib.policy.sample_batch.SampleBatch.agent_steps#

SampleBatch.agent_steps() → int[source]#

Returns the same as len(self) (number of steps in this batch).

To make this compatible with MultiAgentBatch.agent_steps().