When can I get the dataset?

[Ethan-Chen-plus]

I've submitted the Request access, please.

[AnthonyHuo]

you can go to https://github.com/NVlabs/mimicgen repo, they have updated the dataset.

[Ethan-Chen-plus]

Thanks a lot. I have got the dataset. But an error occurs:

---------------------------------------------------------------------------
Exception                                 Traceback (most recent call last)
File /data/ckw/micromamba/envs/sdp/lib/python3.9/site-packages/hydra/_internal/instantiate/_instantiate2.py:92, in _call_target(_target_, _partial_, args, kwargs, full_key)
     91 try:
---> 92     return _target_(*args, **kwargs)
     93 except Exception as e:

File /data/ckw/07robo/04sdp/SDP/diffusion_policy/env_runner/robomimic_image_runner.py:218, in RobomimicImageRunner.__init__(self, output_dir, dataset_path, shape_meta, n_train, n_train_vis, train_start_idx, n_test, n_test_vis, test_start_seed, max_steps, n_obs_steps, n_action_steps, render_obs_key, fps, crf, past_action, abs_action, tqdm_interval_sec, n_envs)
    216     env_init_fn_dills.append(dill.dumps(init_fn))
--> 218 env = AsyncVectorEnv(env_fns, dummy_env_fn=dummy_env_fn)
    219 # env = SyncVectorEnv(env_fns)

File /data/ckw/07robo/04sdp/SDP/diffusion_policy/gym_util/async_vector_env.py:102, in AsyncVectorEnv.__init__(self, env_fns, dummy_env_fn, observation_space, action_space, shared_memory, copy, context, daemon, worker)
    101     dummy_env_fn = env_fns[0]
--> 102 dummy_env = dummy_env_fn()
    103 self.metadata = dummy_env.metadata

File /data/ckw/07robo/04sdp/SDP/diffusion_policy/env_runner/robomimic_image_runner.py:127, in RobomimicImageRunner.__init__.<locals>.dummy_env_fn()
    126 def dummy_env_fn():
--> 127     robomimic_env = create_env(
    128             env_meta=env_meta, 
    129             shape_meta=shape_meta,
    130             enable_render=False
    131         )
    132     return MultiStepWrapper(
    133         VideoRecordingWrapper(
    134             RobomimicImageWrapper(
   (...)
    153         max_episode_steps=max_steps
    154     )

File /data/ckw/07robo/04sdp/SDP/diffusion_policy/env_runner/robomimic_image_runner.py:33, in create_env(env_meta, shape_meta, enable_render)
     31 ObsUtils.initialize_obs_modality_mapping_from_dict(modality_mapping)
---> 33 env = EnvUtils.create_env_from_metadata(
     34     env_meta=env_meta,
     35     render=False, 
     36     render_offscreen=enable_render,
     37     use_image_obs=enable_render, 
     38 )
     39 return env

File /data/ckw/07robo/04sdp/robomimic_pkg/robomimic/utils/env_utils.py:218, in create_env_from_metadata(env_meta, env_name, render, render_offscreen, use_image_obs)
    216 env_kwargs = env_meta["env_kwargs"]
--> 218 env = create_env(
    219     env_type=env_type,
    220     env_name=env_name,  
    221     render=render, 
    222     render_offscreen=render_offscreen, 
    223     use_image_obs=use_image_obs, 
    224     **env_kwargs,
    225 )
    226 check_env_version(env, env_meta)

File /data/ckw/07robo/04sdp/robomimic_pkg/robomimic/utils/env_utils.py:169, in create_env(env_type, env_name, render, render_offscreen, use_image_obs, **kwargs)
    168 env_class = get_env_class(env_type=env_type)
--> 169 env = env_class(
    170     env_name=env_name, 
    171     render=render, 
    172     render_offscreen=render_offscreen, 
    173     use_image_obs=use_image_obs,
    174     postprocess_visual_obs=True,
    175     **kwargs,
    176 )
    177 print("Created environment with name {}".format(env_name))

File /data/ckw/07robo/04sdp/robomimic_pkg/robomimic/envs/env_robosuite.py:93, in EnvRobosuite.__init__(self, env_name, render, render_offscreen, use_image_obs, postprocess_visual_obs, **kwargs)
     92 self._init_kwargs = deepcopy(kwargs)
---> 93 self.env = robosuite.make(self._env_name, **kwargs)
     95 if self._is_v1:
     96     # Make sure joint position observations and eef vel observations are active

File /data/ckw/07robo/04sdp/robosuite_pkg/robosuite/environments/base.py:35, in make(env_name, *args, **kwargs)
     34 if env_name not in REGISTERED_ENVS:
---> 35     raise Exception(
     36         "Environment {} not found. Make sure it is a registered environment among: {}".format(
     37             env_name, ", ".join(REGISTERED_ENVS)
     38         )
     39     )
     40 return REGISTERED_ENVS[env_name](*args, **kwargs)

Exception: Environment Square_D0 not found. Make sure it is a registered environment among: Lift, Stack, NutAssembly, NutAssemblySingle, NutAssemblySquare, NutAssemblyRound, PickPlace, PickPlaceSingle, PickPlaceMilk, PickPlaceBread, PickPlaceCereal, PickPlaceCan, Door, Wipe, ToolHang, TwoArmLift, TwoArmPegInHole, TwoArmHandover, TwoArmTransport

The above exception was the direct cause of the following exception:

InstantiationException                    Traceback (most recent call last)
Cell In[15], line 1
----> 1 workspace.run()

File /data/ckw/07robo/04sdp/SDP/diffusion_policy/workspace/train_diffusion_transformer_hybrid_workspace.py:126, in TrainDiffusionTransformerHybridWorkspace.run(self)
    124 # env_runner3: BaseImageRunner
    125 for i in range(cfg.task_num):
--> 126     env_runners.append(hydra.utils.instantiate(cfg[f'task{i}'].env_runner, output_dir=self.output_dir))
    127     assert isinstance(env_runners[i], BaseImageRunner)
    130 # configure logging

File /data/ckw/micromamba/envs/sdp/lib/python3.9/site-packages/hydra/_internal/instantiate/_instantiate2.py:222, in instantiate(config, *args, **kwargs)
    219     _convert_ = config.pop(_Keys.CONVERT, ConvertMode.NONE)
    220     _partial_ = config.pop(_Keys.PARTIAL, False)
--> 222     return instantiate_node(
    223         config, *args, recursive=_recursive_, convert=_convert_, partial=_partial_
    224     )
    225 elif OmegaConf.is_list(config):
    226     # Finalize config (convert targets to strings, merge with kwargs)
    227     config_copy = copy.deepcopy(config)

File /data/ckw/micromamba/envs/sdp/lib/python3.9/site-packages/hydra/_internal/instantiate/_instantiate2.py:339, in instantiate_node(node, convert, recursive, partial, *args)
    334                 value = instantiate_node(
    335                     value, convert=convert, recursive=recursive
    336                 )
    337             kwargs[key] = _convert_node(value, convert)
--> 339     return _call_target(_target_, partial, args, kwargs, full_key)
    340 else:
    341     # If ALL or PARTIAL non structured, instantiate in dict and resolve interpolations eagerly.
    342     if convert == ConvertMode.ALL or (
    343         convert == ConvertMode.PARTIAL
    344         and node._metadata.object_type in (None, dict)
    345     ):

File /data/ckw/micromamba/envs/sdp/lib/python3.9/site-packages/hydra/_internal/instantiate/_instantiate2.py:97, in _call_target(_target_, _partial_, args, kwargs, full_key)
     95 if full_key:
     96     msg += f"\nfull_key: {full_key}"
---> 97 raise InstantiationException(msg) from e

InstantiationException: Error in call to target 'diffusion_policy.env_runner.robomimic_image_runner.RobomimicImageRunner':
Exception('Environment Square_D0 not found. Make sure it is a registered environment among: Lift, Stack, NutAssembly, NutAssemblySingle, NutAssemblySquare, NutAssemblyRound, PickPlace, PickPlaceSingle, PickPlaceMilk, PickPlaceBread, PickPlaceCereal, PickPlaceCan, Door, Wipe, ToolHang, TwoArmLift, TwoArmPegInHole, TwoArmHandover, TwoArmTransport')
full_key: task0.env_runner

How to solve it? Thanks for your helping!!

[AnthonyHuo]

oops! It seems new version do not have squre_d0 dataset, you can modify the dataset used in config.yaml, I will also check if I can find the orginal data link.

[Ethan-Chen-plus]

Thanks a lot. I use this script: https://github.com/NVlabs/mimicgen/blob/main/mimicgen/scripts/download_datasets.py

python download_datasets.py --dataset_type core --tasks coffee_d0 coffee_preparation_d0 hammer_cleanup_d0 kitchen_d0 mug_cleanup_d0 nut_assembly_d0 pick_place_d0 square_d0 stack_d0 stack_three_d0 threading_d0 three_piece_assembly_d0

It seemed to be ok and I can see the dataset. Also I changed the dataset path: It can smoothly generate Zarr.

However this still occurs:

InstantiationException: Error in call to target 'diffusion_policy.env_runner.robomimic_image_runner.RobomimicImageRunner':
Exception('Environment Square_D0 not found. Make sure it is a registered environment among: Lift, Stack, NutAssembly, NutAssemblySingle, NutAssemblySquare, NutAssemblyRound, PickPlace, PickPlaceSingle, PickPlaceMilk, PickPlaceBread, PickPlaceCereal, PickPlaceCan, Door, Wipe, ToolHang, TwoArmLift, TwoArmPegInHole, TwoArmHandover, TwoArmTransport')
full_key: task0.env_runner

Thanks. I need your help please.

[Ethan-Chen-plus]

Hello! I'm here to ask whether you have solved it and could you please give me more infomation for how to use the download_dataset.py or generate more config for the data? Or maybe I've used the wrong version. I sincerely need your help.@AnthonyHuo

[Ethan-Chen-plus]

@AnthonyHuo Thanks a lot. This issue has been solved. Here: https://github.com/NVlabs/mimicgen/issues/41 https://github.com/ARISE-Initiative/robomimic/issues/205

[Ethan-Chen-plus]

@AnthonyHuo Thanks a lot. This issue has been solved. Here: https://github.com/NVlabs/mimicgen/issues/41 https://github.com/ARISE-Initiative/robomimic/issues/205