38c115e1bf3862010daf57107b90deee7626c8de,tests/agents_tests/test_ppo.py,TestPPO,_test_abc,#TestPPO#Any#Any#Any#Any#,60

Before Change

        // Test
        n_test_runs = 100
        n_succeeded = 0
        for _ in range(n_test_runs):
            total_r = 0.0
            obs = test_env.reset()
            done = False
            reward = 0.0
            while not done:
                action = agent.act(obs)
                obs, reward, done, _ = test_env.step(action)
                total_r += reward
            agent.stop_episode()
            if np.isclose(total_r, successful_return):
                n_succeeded += 1

        if require_success:
            self.assertGreater(n_succeeded, 0.8 * n_test_runs)

        // Save

After Change

        // Test
        n_test_runs = 100
        eval_returns = run_evaluation_episodes(
            test_env,
            agent,
            n_runs=n_test_runs,
            max_episode_len=max_episode_len,
        )
        n_succeeded = np.sum(np.asarray(eval_returns) >= successful_return)

        if require_success:
            self.assertGreater(n_succeeded, 0.8 * n_test_runs)

Instances