- class ray.rllib.policy.sample_batch.MultiAgentBatch(policy_batches: Dict[str, ray.rllib.policy.sample_batch.SampleBatch], env_steps: int)#
A batch of experiences from multiple agents in the environment.
Mapping from policy ids to SampleBatches of experiences.
The number of env steps in this batch.
Initialize a MultiAgentBatch instance.
The number of agent steps (there are >= 1 agent steps per env step).
self(already a MultiAgentBatch).
Compresses each policy batch (per column) in place.
Deep-copies self into a new MultiAgentBatch.
Decompresses each policy batch (per column), if already compressed.
The number of env steps (there are >= 1 agent steps per env step).
The overall size in bytes of all policy batches (all columns).
Returns k-step batches holding data for each agent at those steps.
Returns SampleBatch or MultiAgentBatch, depending on given policies.