eval error
Closed this issue · 5 comments
after the first turn of training loop, I got the error, and I notice the program stoped when body=1 and link 13, the program always stop here File "/media/u/all_code/SurRol2/SurRoL/rl/train.py", line 9, in main exp.train() File "/media/u/all_code/SurRol2/SurRoL/rl/trainers/rl_trainer.py", line 109, in train score = self.eval() File "/media/u/all_code/SurRol2/SurRoL/rl/trainers/rl_trainer.py", line 182, in eval episode, _, env_steps = self.eval_sampler.sample_episode(is_train=False, render=True) File "/media/u/all_code/SurRol2/SurRoL/rl/modules/samplers.py", line 25, in sample_episode self.init() File "/media/u/all_code/SurRol2/SurRoL/rl/modules/samplers.py", line 18, in init self._episode_reset() File "/media/u/all_code/SurRol2/SurRoL/rl/modules/samplers.py", line 55, in _episode_reset self._obs = self._reset_env() File "/media/u/all_code/SurRol2/SurRoL/rl/modules/samplers.py", line 59, in _reset_env return self._env.reset() File "/home/u/anaconda3/envs/surrol3/lib/python3.7/site-packages/gym/wrappers/time_limit.py", line 26, in reset return self.env.reset(**kwargs) File "/home/u/anaconda3/envs/surrol3/lib/python3.7/site-packages/gym/wrappers/order_enforcing.py", line 18, in reset return self.env.reset(**kwargs) File "/media/u/all_code/SurRol2/SurRoL/surrol/gym/surrol_goalenv.py", line 19, in reset return super().reset() File "/media/u/all_code/SurRol2/SurRoL/surrol/gym/surrol_env.py", line 147, in reset self._sample_goal_callback() File "/media/u/all_code/SurRol2/SurRoL/surrol/tasks/needle_pick.py", line 104, in _sample_goal_callback pos_obj, orn_obj = get_link_pose(self.obj_id, self.obj_link1) File "/media/u/all_code/SurRol2/SurRoL/surrol/utils/pybullet_utils.py", line 725, in get_link_pose link_state = get_link_state(body, link) File "/media/u/all_code/SurRol2/SurRoL/surrol/utils/pybullet_utils.py", line 707, in get_link_state return LinkState(*p.getLinkState(body, link))
the error is: pybullet.error: getLinkState failed.
Hi,
This error may be caused by the environment creation, guessed from my experience. May I ask which SurRoL are you using? The one that follows the instructions of this repo, or the official SurRoL repo?
Hi Tao Huang， thanks for your reply, currently I clone the repo of SurRoL-v2, and all the env settings follow this repo.
Hi Junhui,
I see. I would suggest you use our provided SurRoL (submodule) or official SurRoLv1 to make sure our RL implementation is correct. Once done, you may raise an issue in the SurRoLv2 repo to ask for suggestions, as I have little idea about the development of SurRoLv2 compared with v1 and the issue related to simulation itself is not the focus of this repo.
If you still encounter any problems when using SurRoLv1 and DEX, please let me know.
Closing because of inactivity.