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().