TypeError: __init__() got an unexpected keyword argument 'actor_class'

[ammar-n-abbas]

Hi,

while running the training files through rosrun as defined in the readme, I am getting this error:

  **File "~/catkin_ws/src/ur_openai_gym/ur_rl/scripts/tf2rl_sac.py", line 83, in <module>**
    policy = SAC(
  File "~.local/lib/python3.8/site-packages/tf2rl/algos/sac.py", line 50, in __init__
    super().__init__(
  File "~/.local/lib/python3.8/site-packages/tf2rl/algos/policy_base.py", line 86, in __init__
    super().__init__(memory_capacity=memory_capacity, **kwargs)
**TypeError: __init__() got an unexpected keyword argument 'actor_class'**

[cambel]

Hi,

Are you using the docker version or did you install it from source?

In any case, the problem seems to be the version of the tf2rl library. Make sure you are using this version https://github.com/cambel/tf2rl/tree/dev (the dev branch)

[ammar-n-abbas]

@cambel thank you for your reply. The issue is resolved with your branch

However, now it shows this error:

''' GYM Environment:simulation/task_space.yaml Grasping pluging can't be loaded ft_sensor: False, ee_link: tool0, robot_urdf: ur3e [ WARN] [1692696708.704822850]: link 'robotiq_coupler' material 'flat_black' undefined. [ WARN] [1692696708.704878883]: link 'robotiq_coupler' material 'flat_black' undefined. [ WARN] [1692696708.713430526]: link 'robotiq_coupler' material 'flat_black' undefined. [ WARN] [1692696708.713589309]: link 'robotiq_coupler' material 'flat_black' undefined. connecting to: /scaled_pos_joint_traj_controller/command ACTION SPACES TYPE Box(6,) OBSERVATION SPACES TYPE Box(12,) Segmentation fault (core dumped) '''

[cambel]

That may be a problem with the python cpprb library. Please re-install the library to this specific version pip install cpprb==10.0.0

[ammar-n-abbas]

@cambel thanks a lot this error is resolved now.

Getting this attribute error

AttributeError: module 'ur_control.transformations' has no attribute 'pose_from_angular_veloticy'

[cambel]

@ammar-n-abbas Sorry about that, I don't have time right now to fix the issues with this repo. But I have a hopefully working example here. It is self-contained so you shouldn't have the error from above that just means incompatibility with a dependency.

[ammar-n-abbas]

@cambel you are right, it was due to tensorflow and tensorflow_probability compatibility issues it's working fine now.

Thanks

Thank you for sharing another work example, really interesting...

[cambel]

glad it works for you now

[pvela2017]

what version of tensorflow and tensorflow_probabilty are you using? I have the same problem