208e5d91d7f88cc1ab10ab1a4bfdd856b2691671,tensorforce/tests/test_trpo_agent.py,TestTRPOAgent,test_discrete,#TestTRPOAgent#,30

Before Change

    def test_discrete(self):
        environment = MinimalTest(continuous=False)

        config = {
            "batch_size": 16,
            "override_line_search": False,
            "cg_iterations": 20,
            "use_gae": False,
            "normalize_advantage": False,
            "gae_lambda": 0.97,
            "cg_damping": 0.001,
            "line_search_steps": 20,
            "max_kl_divergence": 0.05,
            "max_episode_length": 4,
            "continuous": False,
            "state_shape": (2,),
            "actions": 2,
            "gamma": 0.99
        }

        config = create_config(config)
        tf.reset_default_graph()

After Change

    def test_discrete(self):
        environment = MinimalTest(continuous=False)
        config = Configuration(
            batch_size=8,
            learning_rate=0.0001,
            cg_iterations=20,
            cg_damping=0.001,
            line_search_steps=20,
            max_kl_divergence=0.05,
            states=environment.states,
            actions=environment.actions
        )
        network_builder = layered_network_builder(layers_config=[{"type": "dense", "size": 32}])
        agent = TRPOAgent(config=config, network_builder=network_builder)
        runner = Runner(agent=agent, environment=environment)

Instances