2e3655e8a9b1e37fa6e29f11db02a6b53cfb5928,rllib/agents/a3c/a3c_tf_policy.py,,postprocess_advantages,#Any#Any#Any#Any#,48

Before Change

                           other_agent_batches=None,
                           episode=None):
    completed = sample_batch[SampleBatch.DONES][-1]
    if completed:
        last_r = 0.0
    else:
        next_state = []
        for i in range(policy.num_state_tensors()):
            next_state.append(sample_batch["state_out_{}".format(i)][-1])
        last_r = policy._value(sample_batch[SampleBatch.NEXT_OBS][-1],
                               sample_batch[SampleBatch.ACTIONS][-1],
                               sample_batch[SampleBatch.REWARDS][-1],
                               *next_state)
    return compute_advantages(
        sample_batch, last_r, policy.config["gamma"], policy.config["lambda"],
        policy.config["use_gae"], policy.config["use_critic"])

After Change

                           episode=None):

    // Stub serving backward compatibility.
    deprecation_warning(
        old="rllib.agents.a3c.a3c_tf_policy.postprocess_advantages",
        new="rllib.evaluation.postprocessing.compute_gae_for_sample_batch",
        error=False)

    return compute_gae_for_sample_batch(policy, sample_batch,
                                        other_agent_batches, episode)

Instances