search

qgallouedec / panda-gym

Set of robotic environments based on PyBullet physics engine and gymnasium.
MIT License
506 stars 109 forks source link

I trained panda push using sb3, however i got this error when loading the model. Observation spaces do not match #18

Closed isaacncz closed 9 months ago

isaacncz commented 2 years ago 
raise ValueError(f"Observation spaces do not match: {observation_space} != {env.observation_space}")

ValueError: Observation spaces do not match: Dict(achieved_goal:Box([-10. -10. -10.], [10. 10. 10.], (3,), float32), desired_goal:Box([-10. -10. -10.], [10. 10. 10.], (3,), float32), observation:Box([-10. -10. -10. -10. -10. -10.], [10. 10. 10. 10. 10. 10.], (6,), float32)) != Dict(achieved_goal:Box([-10. -10. -10.], [10. 10. 10.], (3,), float32), desired_goal:Box([-10. -10. -10.], [10. 10. 10.], (3,), float32), observation:Box([-10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10.], [10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10.], (18,), float32))

MH-Hou commented 11 months ago

Hi,

I kind of encountered a very similar problem when I tried to load a trained model from HuggingFace for PickandPlace-v1. Could you please tell me how you solved this problem of observation space mismatch?

qgallouedec commented 11 months ago

Can you share the code used ? What version of sb3 and panda-gym do you use?

MH-Hou commented 11 months ago

Thanks for your quick reply!

Below is the code I ran: test_env = gym.make("PandaPickAndPlace-v1", render=False) checkpoint = load_from_hub(repo_id="sb3/tqc-PandaPickAndPlace-v1", filename="tqc-PandaPickAndPlace-v1.zip") model = TQC.load(checkpoint, env=test_env)

After I ran the code in the terminal, it complained that `raise ValueError(f"Observation spaces do not match: {observation_space} != {env.observation_space}") ValueError: Observation spaces do not match: Dict('achieved_goal': Box(-10.0, 10.0, (3,), float32), 'desired_goal': Box(-10.0, 10.0, (3,), float32), 'observation': Box([-10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. -10. 0.], [10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10. 10.

  1. 1.], (20,), float32)) != Dict('achieved_goal': Box(-10.0, 10.0, (3,), float32), 'desired_goal': Box(-10.0, 10.0, (3,), float32), 'observation': Box(-10.0, 10.0, (19,), float32)) `

For your information, I used the stablebaselines3 of version 2.1.0 and panda_gym of version 1.1.1

I also tried to load the model you uploaded to the huggingface by using the code: test_env = gym.make("PandaPickAndPlace-v1", render=False) checkpoint = load_from_hub(repo_id="qgallouedec/tqc-PandaPickAndPlace-v1-4094880237", filename=" tqc-PandaPickAndPlace-v1.zip") model = TQC.load(checkpoint, env=test_env)

Then this time it complained that raise KeyError("The observation_space and action_space were not given, can't verify new environments") KeyError: "The observation_space and action_space were not given, can't verify new environments"

Could you please tell me what I should do to solve these issues?