244aafdcf89ae814975c8c4e3faf0bd4995c7878,rllib/utils/exploration/tests/test_curiosity.py,TestCuriosity,test_curiosity_on_large_frozen_lake,#TestCuriosity#,88

Before Change

            }
            trainer = ppo.PPOTrainer(config=config)
            rewards_w = 0.0
            for _ in range(num_iterations):
                result = trainer.train()
                rewards_w += result["episode_reward_mean"]
                print(result)
            rewards_w /= num_iterations
            trainer.stop()

            // W/o Curiosity.

After Change

            for _ in range(num_iterations):
                result = trainer.train()
                print(result)
                if result["episode_reward_mean"] >= 0.001:
                    print("Learnt something!")
                    learnt = True
                    break
            trainer.stop()
            self.assertTrue(learnt)

            // // W/o Curiosity. Expect to learn nothing.

Instances