Hugging Face's logo

Datasets:
Hoshipu
/
reflex-b1k-task1-2f

Tasks:
Robotics
Modalities:
Video
Size:
n<1K
Tags:
BEHAVIOR-1K
OmniGibson
R1Pro
failure-recovery
License:
Dataset card Data Studio Files
xet
 Community
Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
recording: struct<root: string, episode: struct<task_id: int64, instance_id: int64, index: int64>>
  child 0, root: string
  child 1, episode: struct<task_id: int64, instance_id: int64, index: int64>
      child 0, task_id: int64
      child 1, instance_id: int64
      child 2, index: int64
success: bool
steps: int64
q_score: null
restored_from: struct<recording: struct<root: string, episode: struct<task_id: int64, instance_id: int64, index: in (... 36 chars omitted)
  child 0, recording: struct<root: string, episode: struct<task_id: int64, instance_id: int64, index: int64>>
      child 0, root: string
      child 1, episode: struct<task_id: int64, instance_id: int64, index: int64>
          child 0, task_id: int64
          child 1, instance_id: int64
          child 2, index: int64
  child 1, step: int64
  child 2, variant: string
goals: list<item: int64>
  child 0, item: int64
num_samples: int64
task_obs_keys: list<item: null>
  child 0, item: null
robot_type: string
ins_id_mapping: string
scene_file: string
n_steps: int64
config: string
n_episodes: int64
to
{'n_episodes': Value('int64'), 'n_steps': Value('int64'), 'num_samples': Value('int64'), 'robot_type': Value('string'), 'success': Value('bool'), 'config': Value('string'), 'scene_file': Value('string'), 'task_obs_keys': List(Value('null')), 'ins_id_mapping': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1835, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 621, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                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
              recording: struct<root: string, episode: struct<task_id: int64, instance_id: int64, index: int64>>
                child 0, root: string
                child 1, episode: struct<task_id: int64, instance_id: int64, index: int64>
                    child 0, task_id: int64
                    child 1, instance_id: int64
                    child 2, index: int64
              success: bool
              steps: int64
              q_score: null
              restored_from: struct<recording: struct<root: string, episode: struct<task_id: int64, instance_id: int64, index: in (... 36 chars omitted)
                child 0, recording: struct<root: string, episode: struct<task_id: int64, instance_id: int64, index: int64>>
                    child 0, root: string
                    child 1, episode: struct<task_id: int64, instance_id: int64, index: int64>
                        child 0, task_id: int64
                        child 1, instance_id: int64
                        child 2, index: int64
                child 1, step: int64
                child 2, variant: string
              goals: list<item: int64>
                child 0, item: int64
              num_samples: int64
              task_obs_keys: list<item: null>
                child 0, item: null
              robot_type: string
              ins_id_mapping: string
              scene_file: string
              n_steps: int64
              config: string
              n_episodes: int64
              to
              {'n_episodes': Value('int64'), 'n_steps': Value('int64'), 'num_samples': Value('int64'), 'robot_type': Value('string'), 'success': Value('bool'), 'config': Value('string'), 'scene_file': Value('string'), 'task_obs_keys': List(Value('null')), 'ins_id_mapping': Value('string')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1379, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 956, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1702, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1887, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

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.

n_episodes
int64
n_steps
int64
num_samples
int64
robot_type
string
success
bool
config
string
scene_file
string
task_obs_keys
list
ins_id_mapping
string
1
10,536
10,536
R1Pro
false
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
1
1,768
1,768
R1Pro
true
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
1
10,536
10,536
R1Pro
false
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
1
2,257
2,257
R1Pro
true
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
1
10,536
10,536
R1Pro
false
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
1
3,811
3,811
R1Pro
true
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
1
10,536
10,536
R1Pro
false
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
1
1,665
1,665
R1Pro
true
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
1
10,536
10,536
R1Pro
false
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
1
1,087
1,087
R1Pro
true
"{\"env\": {\"action_frequency\": 30, \"rendering_frequency\": 30, \"physics_frequency\": 120, \"dev(...TRUNCATED)
"\"{\\n \\\"versions\\\": {\\n \\\"omnigibson\\\": {\\n \\\"version\\\": \\\"3.(...TRUNCATED)
[]
{"0": "background"}
End of preview.

ReFLEX · BEHAVIOR-1K task 1 (picking_up_trash) · end-of-episode recovery (ablation arm 2f)

This is ablation arm 2f of the ReFLEX project: recovery from the end of each failed episode, with no failure localization, no failure modes and no augmentation. It is a companion to arms 2b (Hoshipu/reflex-b1k-task1-recoveries), 2c (Hoshipu/reflex-b1k-task1-2c) and 2d (Hoshipu/reflex-b1k-task1-2d). It uses the same task, policy (pi0.5 2025ckpt2), stack and file layout. See the 2b card for the layout and the conversion notes.

How it was made

  1. Restart point: the last frame of a failed rollout (step n_steps), always. The agent does not watch the rollout's videos or look for the failure.
  2. Recovery: one Claude Code agent per failed rollout (PROMPT.md) restores that frame in OmniGibson 3.9.2 and writes a classical planner (IK, interpolation, simple navigation; privileged state reads, actions only through the 23-D R1Pro command) that finishes the whole task: every can inside the trash can, none taken out. A run counts when the BDDL goal holds and nothing that was inside is lost; the agent then repeats the unchanged planner once to confirm.
  3. Order and volume: the 363 task-1 failures were shuffled (seed 2026) and started 8 at a time until about 2.5 h of recordings, to match arms 2b, 2c and 2d. 158 agents were started and 143 succeeded (91%). Agents were stopped after 2.5 h; one whose last run had already finished the task was resumed once to confirm and report (NOTES.md). Each success's final run was replayed with recording on; 9 replays did not reproduce the goal and are left out.

Contents

Clips 134, one per rollout
Frames 285,404 at 30 fps, 2.64 h (27 % of the 200 task-1 human demos)
Whole task finished within the clip 134 (all)
Agent cost $391 of Claude tokens over 158 agents; 121 simulator GPU-hours

In scored.json, planner is 2f-agent/ep<id>, mode is end_of_episode and goal_fact is task_done. Recovery episodes use index 9. Paths are relative (root: "."). The parquets use the 2025 layout: actions are one row late and the base velocity is in the world frame. Convert them with mix_dataset.py before training. Videos are RGB only.
Downloads last month
105
Total file size:
2.15 GB