Datasets:

wintermelontree
/

robot_policy_overfitting_study

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
avg_steps: double
checkpoint_path: string
episodes: list<item: struct<episode_index: int64, max_reward: double, seed: int64, steps: int64, success: doub (... 26 chars omitted)
  child 0, item: struct<episode_index: int64, max_reward: double, seed: int64, steps: int64, success: double, total_r (... 14 chars omitted)
      child 0, episode_index: int64
      child 1, max_reward: double
      child 2, seed: int64
      child 3, steps: int64
      child 4, success: double
      child 5, total_reward: double
num_episodes: int64
policy_state_key: string
success_rate: double
epoch_0150: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0050: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0100: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0300: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0650: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
...
: double>
      child 0, item: double
epoch_0550: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0700: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0800: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0500: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0200: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0600: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
epoch_0750: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
  child 0, success_rate_mean: double
  child 1, per_pass_success_rates: list<item: double>
      child 0, item: double
to
{'epoch_0050': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0100': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0150': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0200': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0250': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0300': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0350': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0400': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0450': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0500': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0550': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0600': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0650': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0700': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0750': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0800': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 149, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 129, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 544, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 3052, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2559, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2584, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 558, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 440, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 361, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2381, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2309, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              avg_steps: double
              checkpoint_path: string
              episodes: list<item: struct<episode_index: int64, max_reward: double, seed: int64, steps: int64, success: doub (... 26 chars omitted)
                child 0, item: struct<episode_index: int64, max_reward: double, seed: int64, steps: int64, success: double, total_r (... 14 chars omitted)
                    child 0, episode_index: int64
                    child 1, max_reward: double
                    child 2, seed: int64
                    child 3, steps: int64
                    child 4, success: double
                    child 5, total_reward: double
              num_episodes: int64
              policy_state_key: string
              success_rate: double
              epoch_0150: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0050: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0100: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0300: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0650: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
              ...
              : double>
                    child 0, item: double
              epoch_0550: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0700: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0800: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0500: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0200: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0600: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              epoch_0750: struct<success_rate_mean: double, per_pass_success_rates: list<item: double>>
                child 0, success_rate_mean: double
                child 1, per_pass_success_rates: list<item: double>
                    child 0, item: double
              to
              {'epoch_0050': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0100': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0150': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0200': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0250': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0300': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0350': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0400': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0450': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0500': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0550': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0600': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0650': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0700': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0750': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}, 'epoch_0800': {'success_rate_mean': Value('float64'), 'per_pass_success_rates': List(Value('float64'))}}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

Robot policy overfitting study: data and checkpoints

Companion release for the policy-learning-dynamics code. Five MimicGen manipulation tasks (mug_cleanup_d1, square_d2, threading_d2, hammer_cleanup_d1, three_piece_assembly_d1), image-based diffusion policies, absolute end-effector actions with a rotation-6d block.

folder	content
`bc_pretraining_data/`	per task: `<task>.hdf5` (the 1000 MimicGen demos: environment metadata for rollouts and rewards), `<task>_image_abs.zarr.tar` (the image dataset every policy and generator was trained on), `<task>_lowdim_abs.zarr.tar` (simulator-state version of the same demos)
`bc_pretraining_ckpts/<task>_image/`	behaviour-cloning run: `.hydra/config.yaml` and `checkpoints/epoch_0050.ckpt` ... `epoch_0800.ckpt` (16 per task)
`joint_generator_ckpts/<task>_image/`	joint (image, state, action) generator run: config and the checkpoints the distillation runs sampled from
`joint_aug_ckpts/<task>_image/`	behaviour cloning with generated-sample distillation: config, 16 checkpoints, and the closed-loop eval results of the run

Each checkpoint holds cfg (the resolved training config, with data paths written as ${oc.env:MIMICGEN_DATA_ROOT}/core/...), model_state, ema_state (the weights used for every evaluation), normalizer_state, epoch and global_step. Optimizer state was dropped. release_manifest.json lists every file with the run it came from.

The code repo's README explains how to download a subset (scripts/release/download_release.py) and rerun the five experiments.

License: the checkpoints and configs are released under the MIT License, like the code. The demonstrations in bc_pretraining_data/ are MimicGen data and remain under the MimicGen license.

Downloads last month: 362

Total file size:

394 GB