Capture videos during training doesn't work

[frt03]

Hi,

I tried Capture videos during training in README.md; i.e. running

python train.py capture_video=True capture_video_freq=1500 capture_video_len=100 force_render=False

but got the following error:

Traceback (most recent call last):
  File "train.py", line 156, in launch_rlg_hydra
    runner.run({
  File "/opt/conda/lib/python3.8/site-packages/rl_games/torch_runner.py", line 120, in run
    self.run_train(args)
  File "/opt/conda/lib/python3.8/site-packages/rl_games/torch_runner.py", line 101, in run_train
    agent.train()
  File "/opt/conda/lib/python3.8/site-packages/rl_games/common/a2c_common.py", line 1162, in train
    self.obs = self.env_reset()
  File "/opt/conda/lib/python3.8/site-packages/rl_games/common/a2c_common.py", line 470, in env_reset
    obs = self.vec_env.reset()
  File "/home/root/furuta/IsaacGymEnvs/isaacgymenvs/utils/rlgames_utils.py", line 159, in reset
    return self.env.reset()
  File "/opt/conda/lib/python3.8/site-packages/gym/wrappers/record_video.py", line 109, in reset
    self.start_video_recorder()
  File "/opt/conda/lib/python3.8/site-packages/gym/wrappers/record_video.py", line 121, in start_video_recorder
    self.video_recorder = video_recorder.VideoRecorder(
  File "/opt/conda/lib/python3.8/site-packages/gym/wrappers/monitoring/video_recorder.py", line 61, in __init__
    self.render_mode = env.render_mode
AttributeError: 'Ant' object has no attribute 'render_mode'

Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.
Exception ignored in: <function VideoRecorder.__del__ at 0x7f1372356ca0>
Traceback (most recent call last):
  File "/opt/conda/lib/python3.8/site-packages/gym/wrappers/monitoring/video_recorder.py", line 269, in __del__
    self.close()
  File "/opt/conda/lib/python3.8/site-packages/gym/wrappers/monitoring/video_recorder.py", line 225, in close
    self.env.close()
AttributeError: 'Ant' object has no attribute 'close'

I guess it was caused by the API change of Open AI Gym. Can we fix this somehow? or could you add the valid version of gym in setup.py? FYI, I used gym==0.25.2.

Thank you for your great effort!

[Touutae-lab]

Maybe this is the answer you were looking for #75

[Touutae-lab]

For now i suggest you to use gym==0.24 instead of gym==0.25.

Maybe this is the answer you were looking for #75

[frt03]

@Touutae-lab Thanks!

For your reference, I tried gym==0.25.1 and gym==0.24.1. gym==0.25.1 raised another API issue and gym==0.24.1 showed some warnings (but it seemed to work). So, I'm using gym==0.23.1now.