Function code

rllib/utils/debug/memory.py:60–70

()

        action_sample = action_space.sample()

        def code():
            ts = 0
            env.reset()
            while True:
                # If masking is used, try something like this:
                # np.random.choice(
                #    action_space.n, p=(obs["action_mask"] / sum(obs["action_mask"])))
                _, _, done, _, _ = env.step(action_sample)
                ts += 1
                if done:
                    break

        test = _test_some_code_for_memory_leaks(
            desc="Looking for leaks in env, running through episodes.",