PPO playing impala-BreakoutNoFrameskip-v4 from https://github.com/sgoodfriend/rl-algo-impls/tree/e47a44c4d891f48885af0b1605b30d19fc67b5af
b05c680
| import gym | |
| import numpy as np | |
| from gym.wrappers.monitoring.video_recorder import VideoRecorder | |
| from stable_baselines3.common.vec_env.base_vec_env import ( | |
| VecEnvStepReturn, | |
| VecEnvObs, | |
| ) | |
| class VecEpisodeRecorder(gym.Wrapper): | |
| def __init__(self, env, base_path: str, max_video_length: int = 3600): | |
| super().__init__(env) | |
| self.base_path = base_path | |
| self.max_video_length = max_video_length | |
| self.video_recorder = None | |
| self.recorded_frames = 0 | |
| def step(self, actions: np.ndarray) -> VecEnvStepReturn: | |
| obs, rew, dones, infos = self.env.step(actions) | |
| # Using first env to record episodes | |
| if self.video_recorder: | |
| self.video_recorder.capture_frame() | |
| self.recorded_frames += 1 | |
| if dones[0] and infos[0].get("episode"): | |
| episode_info = { | |
| k: v.item() if hasattr(v, "item") else v | |
| for k, v in infos[0]["episode"].items() | |
| } | |
| self.video_recorder.metadata["episode"] = episode_info | |
| if dones[0] or self.recorded_frames > self.max_video_length: | |
| self._close_video_recorder() | |
| return obs, rew, dones, infos | |
| def reset(self) -> VecEnvObs: | |
| obs = self.env.reset() | |
| self._start_video_recorder() | |
| return obs | |
| def _start_video_recorder(self) -> None: | |
| self._close_video_recorder() | |
| self.video_recorder = VideoRecorder( | |
| self.env, | |
| base_path=self.base_path, | |
| ) | |
| self.video_recorder.capture_frame() | |
| self.recorded_frames = 1 | |
| def _close_video_recorder(self) -> None: | |
| if self.video_recorder: | |
| self.video_recorder.close() | |
| self.video_recorder = None | |