https://colab.research.google.com/drive/1ZmFhb0j0yqPuMtgrkeqYkQ3v8qm0Xr1l?usp=sharing
Up until several weeks ago, this Google Colab notebook worked fine - all it is is setting up a Gym environment, having the agent randomly choose actions, and then record and visualize the result - all within a Google Colab notebook. There's not even any actual learning going on here.
Google Colab suddenly decided that it's only going to support Tensorflow 2, which has introduced a host of compatibility issues, and almost every Colab notebook that I've found that demonstrates Deep reinforcement learning with Open AI Gym is now simply broken.
In the notebook linked above, I've gotten at least everything to run except for the last cell. Traceback included below.
gym/envs/classic_control/rendering.py - seems to throw an error with .get_data() : TypeError: get_data() missing 2 required positional arguments: 'format' and 'pitch'
I've found a couple links regarding correctly calling this new get_data() with format and pitch, but I'm not sure what's the most elegant way to solve this issue. Should I try to try to locally rewrite this function call so it takes in the correct argument sequence?
TypeError Traceback (most recent call last)
<ipython-input-19-0c4a02550ae0> in <module>
1 # Record & show video
----> 2 record_video(env_id, video_length=2000, prefix='random_agent')
3 show_videos('videos', prefix='random_agent')
7 frames
<ipython-input-15-a4b0623bbfac> in record_video(env_id, video_length, prefix, video_folder)
23 name_prefix=prefix)
24
---> 25 obs = eval_env.reset()
26 print("env.reset() shape: %s" %str(obs.shape))
27
/usr/local/lib/python3.7/dist-packages/stable_baselines3/common/vec_env/vec_video_recorder.py in reset(self)
66 def reset(self) -> VecEnvObs:
67 obs = self.venv.reset()
---> 68 self.start_video_recorder()
69 return obs
70
/usr/local/lib/python3.7/dist-packages/stable_baselines3/common/vec_env/vec_video_recorder.py in start_video_recorder(self)
78 )
79
---> 80 self.video_recorder.capture_frame()
81 self.recorded_frames = 1
82 self.recording = True
/usr/local/lib/python3.7/dist-packages/gym/wrappers/monitoring/video_recorder.py in capture_frame(self)
130
131 render_mode = "ansi" if self.ansi_mode else "rgb_array"
--> 132 frame = self.env.render(mode=render_mode)
133
134 if frame is None:
/usr/local/lib/python3.7/dist-packages/stable_baselines3/common/vec_env/dummy_vec_env.py in render(self, mode)
85 """
86 if self.num_envs == 1:
---> 87 return self.envs[0].render(mode=mode)
88 else:
89 return super().render(mode=mode)
/usr/local/lib/python3.7/dist-packages/gym/core.py in render(self, mode, **kwargs)
293
294 def render(self, mode="human", **kwargs):
--> 295 return self.env.render(mode, **kwargs)
296
297 def close(self):
/usr/local/lib/python3.7/dist-packages/gym/envs/box2d/bipedal_walker.py in render(self, mode)
562 self.viewer.draw_polyline(f + [f[0]], color=(0, 0, 0), linewidth=2)
563
--> 564 return self.viewer.render(return_rgb_array=mode == "rgb_array")
565
566 def close(self):
/usr/local/lib/python3.7/dist-packages/gym/envs/classic_control/rendering.py in render(self, return_rgb_array)
134 buffer = pyglet.image.get_buffer_manager().get_color_buffer()
135 image_data = buffer.get_image_data()
--> 136 arr = np.frombuffer(image_data.get_data(), dtype=np.uint8)
137 # In https://github.com/openai/gym-http-api/issues/2, we
138 # discovered that someone using Xmonad on Arch was having
TypeError: get_data() missing 2 required positional arguments: 'format' and 'pitch'
