Robosuite skrl documentation not working

[skyang1]

I tried running the example to train the environment TwoArmLift with skrl. But something went wrong
I use skrl=1.3 and robosuite==1.5
I tried changing RoboSuite to version 1.4 but still getting the same error

[skrl:WARNING] Using `from skrl.envs.torch import ...` is deprecated and will be removed in future versions.
[skrl:WARNING]  - Import loaders using `from skrl.envs.loaders.torch import ...`
[skrl:WARNING]  - Import wrappers using `from skrl.envs.wrappers.torch import ...`
[skrl:INFO] Environment wrapper: 'auto' (class: robosuite.environments.manipulation.two_arm_env.TwoArmEnv)
[skrl:INFO] Environment wrapper: Robosuite
G:\downLoad\Miniconda\miniconda\envs\skrl-robosuite\lib\site-packages\gym\spaces\box.py:127: UserWarning: WARN: Box bound precision lowered by casting to float32
  logger.warn(f"Box bound precision lowered by casting to {self.dtype}")
Traceback (most recent call last):
  File "G:\new_study\graduation\proj1\two_arm_lift.py", line 114, in <module>
    trainer = SequentialTrainer(cfg=cfg_trainer, env=env, agents=agent)
  File "G:\downLoad\Miniconda\miniconda\envs\skrl-robosuite\lib\site-packages\skrl\trainers\torch\sequential.py", line 55, in __init__
    self.agents.init(trainer_cfg=self.cfg)
  File "G:\downLoad\Miniconda\miniconda\envs\skrl-robosuite\lib\site-packages\skrl\agents\torch\td3\td3.py", line 203, in init
    self.memory.create_tensor(name="states", size=self.observation_space, dtype=torch.float32)
  File "G:\downLoad\Miniconda\miniconda\envs\skrl-robosuite\lib\site-packages\skrl\memories\torch\base.py", line 197, in create_tensor
    size = self._get_space_size(size, keep_dimensions)
  File "G:\downLoad\Miniconda\miniconda\envs\skrl-robosuite\lib\site-packages\skrl\memories\torch\base.py", line 111, in _get_space_size
    raise ValueError("keep_dimensions=True cannot be used with Dict spaces")
ValueError: keep_dimensions=True cannot be used with Dict spaces

[Toni-SM]

Hi @skyang1

This is a known issue that is fixed in version 1.4.0 (unreleased).

skrl/CHANGELOG.md

Line 18 in e5c6b81

- Make flattened tensor storage in memory the default option (revert changed introduced in version 1.3.0)

If you want to try it, please, install skrl from the develop branch:

pip install git+https://github.com/Toni-SM/skrl.git@develop

[skyang1]

Thank you very much for your assistance
I made an attempt, but encountered a new issue.

[skrl:INFO] Environment wrapper: 'auto' (class: robosuite.environments.manipulation.two_arm_env.TwoArmEnv)
[skrl:INFO] Environment wrapper: Robosuite
G:\downLoad\Miniconda\miniconda\envs\skrl4\lib\site-packages\gymnasium\spaces\box.py:235: UserWarning: WARN: Box low's precision lowered by casting to float32, current low.dtype=float64
  gym.logger.warn(
G:\downLoad\Miniconda\miniconda\envs\skrl4\lib\site-packages\gymnasium\spaces\box.py:305: UserWarning: WARN: Box high's precision lowered by casting to float32, current high.dtype=float64
  gym.logger.warn(
Traceback (most recent call last):
  File "G:\new_study\graduation\proj1\two_arm_lift.py", line 78, in <module>
    models["policy"] = DeterministicActor(env.observation_space, env.action_space, device)
  File "G:\new_study\graduation\skrl\skrl\envs\wrappers\torch\robosuite_envs.py", line 37, in observation_space
    return convert_gym_space(self._observation_space)
  File "G:\new_study\graduation\skrl\skrl\utils\spaces\torch\spaces.py", line 35, in convert_gym_space
    raise ValueError(f"Unsupported space ({space})")
ValueError: Unsupported space (Dict('gripper0_to_handle0': Box(-inf, inf, (3,), float32), 'gripper1_to_handle1': Box(-inf, inf, (3,), float32), 'handle0_xpos': Box(-inf, inf, (3,), float32), 'handle1_xpos': Box(-inf, inf, (3,), float32), 'object-state': Box(-inf, inf, (19,), float32), 'pot_pos': Box(-inf, inf, (3,), float32), 'pot_quat': Box(-inf, inf, (4,), float32), 'robot0_eef_pos': Box(-inf, inf, (3,), float32), 'robot0_eef_quat': Box(-inf, inf, (4,), float32), 'robot0_gripper_qpos': Box(-inf, inf, (2,), float32), 'robot0_gripper_qvel': Box(-inf, inf, (2,), float32), 'robot0_joint_pos_cos': Box(-inf, inf, (7,), float32), 'robot0_joint_pos_sin': Box(-inf, inf, (7,), float32), 'robot0_joint_vel': Box(-inf, inf, (7,), float32), 'robot0_proprio-state': Box(-inf, inf, (32,), float32), 'robot1_eef_pos': Box(-inf, inf, (3,), float32), 'robot1_eef_quat': Box(-inf, inf, (4,), float32), 'robot1_gripper_qpos': Box(-inf, inf, (2,), float32), 'robot1_gripper_qvel': Box(-inf, inf, (2,), float32), 'robot1_joint_pos_cos': Box(-inf, inf, (7,), float32), 'robot1_joint_pos_sin': Box(-inf, inf, (7,), float32), 'robot1_joint_vel': Box(-inf, inf, (7,), float32), 'robot1_proprio-state': Box(-inf, inf, (32,), float32)))