57544b1ff9f97d4da9f64d25c8ea5a3d8d247ffc,rllib/examples/two_trainer_workflow.py,,,#,91

Before Change

        default_policy=None,
        execution_plan=custom_training_workflow)

    tune.run(
        MyTrainer,
        stop={"training_iteration": args.num_iters},
        config={
            "rollout_fragment_length": 50,
            "num_workers": 0,
            "env": "multi_agent_cartpole",
            "multiagent": {
                "policies": policies,
                "policy_mapping_fn": policy_mapping_fn,
                "policies_to_train": ["dqn_policy", "ppo_policy"],
            },
        })

After Change

    if args.as_test:
        check_learning_achieved(results, args.stop_reward)

    ray.shutdown()

Instances