Hugging Face's logo

Datasets:
erl-hub
/
behaviour1k-Qwen3-DINOv3

Tasks:
Robotics
Tags:
pls-vla
behavior-1k
b1k
vision-language-action
skill-latents
License:
Dataset card Data Studio Files
xet
 Community
1
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
episode: struct<episode_id: string, episode_index: int64, fps: double, instruction: string, n_frames: int64,  (... 86 chars omitted)
  child 0, episode_id: string
  child 1, episode_index: int64
  child 2, fps: double
  child 3, instruction: string
  child 4, n_frames: int64
  child 5, n_samples: int64
  child 6, n_segments: int64
  child 7, source: string
  child 8, split: string
  child 9, task_index: int64
fingerprint: string
manifest: list<item: struct<chunk_end_frame: int64, chunk_id: int64, chunk_start_frame: int64, end_frame: int6 (... 339 chars omitted)
  child 0, item: struct<chunk_end_frame: int64, chunk_id: int64, chunk_start_frame: int64, end_frame: int64, frame_id (... 327 chars omitted)
      child 0, chunk_end_frame: int64
      child 1, chunk_id: int64
      child 2, chunk_start_frame: int64
      child 3, end_frame: int64
      child 4, frame_idx: int64
      child 5, instruction: string
      child 6, is_mixed_subtask: bool
      child 7, pre_index: int64
      child 8, progress: double
      child 9, row_index: int64
      child 10, sample_id: int64
      child 11, segment_id: int64
      child 12, split: string
      child 13, start_frame: int64
      child 14, subtask: string
      child 15, subtask_end_frame: int64
      child 16, subtask_id: int64
      child 17, subtask_start_frame: int64
      child 18, target_index: int64
      child 19, video_id: string
      child 20, vis_index: int64
segments: list<item: struct<chunk_id: int64, end_frame: int64, e
...
nt64, end_frame: int64, episode_id: string, identity_mse: double, is_mixed_subtask (... 184 chars omitted)
      child 0, chunk_id: int64
      child 1, end_frame: int64
      child 2, episode_id: string
      child 3, identity_mse: double
      child 4, is_mixed_subtask: bool
      child 5, max_abs: double
      child 6, mean_frame_norm: double
      child 7, n_sampled: int64
      child 8, segment_id: int64
      child 9, split: string
      child 10, start_frame: int64
      child 11, subtask: string
      child 12, target_norm: double
      child 13, within_variance: double
video: struct<episode_index: int64, n_frames: int64, n_segments: int64, source: string, split: string, task (... 32 chars omitted)
  child 0, episode_index: int64
  child 1, n_frames: int64
  child 2, n_segments: int64
  child 3, source: string
  child 4, split: string
  child 5, task_index: int64
  child 6, video_id: string
n_segments: int64
arrays: struct<actions: list<item: int64>, frames: list<item: int64>, pre: list<item: int64>, proprio: list< (... 73 chars omitted)
  child 0, actions: list<item: int64>
      child 0, item: int64
  child 1, frames: list<item: int64>
      child 0, item: int64
  child 2, pre: list<item: int64>
      child 0, item: int64
  child 3, proprio: list<item: int64>
      child 0, item: int64
  child 4, segment_targets: list<item: int64>
      child 0, item: int64
  child 5, vis: list<item: int64>
      child 0, item: int64
n_samples: int64
split: string
episode_id: string
to
{'arrays': {'actions': List(Value('int64')), 'frames': List(Value('int64')), 'pre': List(Value('int64')), 'proprio': List(Value('int64')), 'segment_targets': List(Value('int64')), 'vis': List(Value('int64'))}, 'episode_id': Value('string'), 'n_samples': Value('int64'), 'n_segments': Value('int64'), 'split': 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 1827, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, 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 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              episode: struct<episode_id: string, episode_index: int64, fps: double, instruction: string, n_frames: int64,  (... 86 chars omitted)
                child 0, episode_id: string
                child 1, episode_index: int64
                child 2, fps: double
                child 3, instruction: string
                child 4, n_frames: int64
                child 5, n_samples: int64
                child 6, n_segments: int64
                child 7, source: string
                child 8, split: string
                child 9, task_index: int64
              fingerprint: string
              manifest: list<item: struct<chunk_end_frame: int64, chunk_id: int64, chunk_start_frame: int64, end_frame: int6 (... 339 chars omitted)
                child 0, item: struct<chunk_end_frame: int64, chunk_id: int64, chunk_start_frame: int64, end_frame: int64, frame_id (... 327 chars omitted)
                    child 0, chunk_end_frame: int64
                    child 1, chunk_id: int64
                    child 2, chunk_start_frame: int64
                    child 3, end_frame: int64
                    child 4, frame_idx: int64
                    child 5, instruction: string
                    child 6, is_mixed_subtask: bool
                    child 7, pre_index: int64
                    child 8, progress: double
                    child 9, row_index: int64
                    child 10, sample_id: int64
                    child 11, segment_id: int64
                    child 12, split: string
                    child 13, start_frame: int64
                    child 14, subtask: string
                    child 15, subtask_end_frame: int64
                    child 16, subtask_id: int64
                    child 17, subtask_start_frame: int64
                    child 18, target_index: int64
                    child 19, video_id: string
                    child 20, vis_index: int64
              segments: list<item: struct<chunk_id: int64, end_frame: int64, e
              ...
              nt64, end_frame: int64, episode_id: string, identity_mse: double, is_mixed_subtask (... 184 chars omitted)
                    child 0, chunk_id: int64
                    child 1, end_frame: int64
                    child 2, episode_id: string
                    child 3, identity_mse: double
                    child 4, is_mixed_subtask: bool
                    child 5, max_abs: double
                    child 6, mean_frame_norm: double
                    child 7, n_sampled: int64
                    child 8, segment_id: int64
                    child 9, split: string
                    child 10, start_frame: int64
                    child 11, subtask: string
                    child 12, target_norm: double
                    child 13, within_variance: double
              video: struct<episode_index: int64, n_frames: int64, n_segments: int64, source: string, split: string, task (... 32 chars omitted)
                child 0, episode_index: int64
                child 1, n_frames: int64
                child 2, n_segments: int64
                child 3, source: string
                child 4, split: string
                child 5, task_index: int64
                child 6, video_id: string
              n_segments: int64
              arrays: struct<actions: list<item: int64>, frames: list<item: int64>, pre: list<item: int64>, proprio: list< (... 73 chars omitted)
                child 0, actions: list<item: int64>
                    child 0, item: int64
                child 1, frames: list<item: int64>
                    child 0, item: int64
                child 2, pre: list<item: int64>
                    child 0, item: int64
                child 3, proprio: list<item: int64>
                    child 0, item: int64
                child 4, segment_targets: list<item: int64>
                    child 0, item: int64
                child 5, vis: list<item: int64>
                    child 0, item: int64
              n_samples: int64
              split: string
              episode_id: string
              to
              {'arrays': {'actions': List(Value('int64')), 'frames': List(Value('int64')), 'pre': List(Value('int64')), 'proprio': List(Value('int64')), 'segment_targets': List(Value('int64')), 'vis': List(Value('int64'))}, 'episode_id': Value('string'), 'n_samples': Value('int64'), 'n_segments': Value('int64'), 'split': 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 1369, 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 948, 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 1694, 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 1880, 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.

arrays
dict
episode_id
string
n_samples
int64
n_segments
int64
split
string
{ "actions": [ 64, 23 ], "frames": [ 64 ], "pre": [ 64, 4096 ], "proprio": [ 64, 61 ], "segment_targets": [ 8, 4096 ], "vis": [ 64, 768 ] }
task0000_ep00000000
64
8
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000001
24
3
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000002
32
4
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000003
16
2
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000004
24
3
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000005
24
3
train
{ "actions": [ 40, 23 ], "frames": [ 40 ], "pre": [ 40, 4096 ], "proprio": [ 40, 61 ], "segment_targets": [ 5, 4096 ], "vis": [ 40, 768 ] }
task0000_ep00000006
40
5
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000007
32
4
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000008
32
4
train
{ "actions": [ 48, 23 ], "frames": [ 48 ], "pre": [ 48, 4096 ], "proprio": [ 48, 61 ], "segment_targets": [ 6, 4096 ], "vis": [ 48, 768 ] }
task0000_ep00000009
48
6
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000010
32
4
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000011
16
2
train
{ "actions": [ 64, 23 ], "frames": [ 64 ], "pre": [ 64, 4096 ], "proprio": [ 64, 61 ], "segment_targets": [ 8, 4096 ], "vis": [ 64, 768 ] }
task0000_ep00000012
64
8
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000013
16
2
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000014
24
3
train
{ "actions": [ 48, 23 ], "frames": [ 48 ], "pre": [ 48, 4096 ], "proprio": [ 48, 61 ], "segment_targets": [ 6, 4096 ], "vis": [ 48, 768 ] }
task0000_ep00000015
48
6
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000016
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000017
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000018
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000019
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000020
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000021
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000022
16
2
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000023
32
4
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000024
24
3
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000025
24
3
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000026
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000027
16
2
train
{ "actions": [ 40, 23 ], "frames": [ 40 ], "pre": [ 40, 4096 ], "proprio": [ 40, 61 ], "segment_targets": [ 5, 4096 ], "vis": [ 40, 768 ] }
task0000_ep00000028
40
5
test
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000029
24
3
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000030
24
3
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000031
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000032
16
2
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000033
24
3
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000034
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000035
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000036
16
2
test
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000037
32
4
train
{ "actions": [ 40, 23 ], "frames": [ 40 ], "pre": [ 40, 4096 ], "proprio": [ 40, 61 ], "segment_targets": [ 5, 4096 ], "vis": [ 40, 768 ] }
task0000_ep00000038
40
5
train
{ "actions": [ 40, 23 ], "frames": [ 40 ], "pre": [ 40, 4096 ], "proprio": [ 40, 61 ], "segment_targets": [ 5, 4096 ], "vis": [ 40, 768 ] }
task0000_ep00000039
40
5
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000040
16
2
val
{ "actions": [ 80, 23 ], "frames": [ 80 ], "pre": [ 80, 4096 ], "proprio": [ 80, 61 ], "segment_targets": [ 10, 4096 ], "vis": [ 80, 768 ] }
task0000_ep00000041
80
10
val
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000042
24
3
test
{ "actions": [ 48, 23 ], "frames": [ 48 ], "pre": [ 48, 4096 ], "proprio": [ 48, 61 ], "segment_targets": [ 6, 4096 ], "vis": [ 48, 768 ] }
task0000_ep00000043
48
6
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000044
32
4
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000045
16
2
train
{ "actions": [ 48, 23 ], "frames": [ 48 ], "pre": [ 48, 4096 ], "proprio": [ 48, 61 ], "segment_targets": [ 6, 4096 ], "vis": [ 48, 768 ] }
task0000_ep00000046
48
6
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000047
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000048
16
2
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000049
24
3
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000050
16
2
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000051
24
3
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000052
24
3
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000053
16
2
val
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000054
24
3
train
{ "actions": [ 48, 23 ], "frames": [ 48 ], "pre": [ 48, 4096 ], "proprio": [ 48, 61 ], "segment_targets": [ 6, 4096 ], "vis": [ 48, 768 ] }
task0000_ep00000055
48
6
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000056
16
2
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000057
32
4
val
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000058
32
4
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000059
16
2
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000060
24
3
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000061
24
3
val
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000062
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000063
16
2
val
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000064
24
3
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000065
16
2
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000066
24
3
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000067
24
3
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000068
32
4
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000069
32
4
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000070
24
3
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000071
24
3
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000072
16
2
train
{ "actions": [ 72, 23 ], "frames": [ 72 ], "pre": [ 72, 4096 ], "proprio": [ 72, 61 ], "segment_targets": [ 9, 4096 ], "vis": [ 72, 768 ] }
task0000_ep00000073
72
9
test
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000074
16
2
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000075
24
3
train
{ "actions": [ 48, 23 ], "frames": [ 48 ], "pre": [ 48, 4096 ], "proprio": [ 48, 61 ], "segment_targets": [ 6, 4096 ], "vis": [ 48, 768 ] }
task0000_ep00000076
48
6
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000077
32
4
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000078
32
4
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000079
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000080
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000081
16
2
train
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000082
32
4
test
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000083
32
4
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000084
16
2
train
{ "actions": [ 56, 23 ], "frames": [ 56 ], "pre": [ 56, 4096 ], "proprio": [ 56, 61 ], "segment_targets": [ 7, 4096 ], "vis": [ 56, 768 ] }
task0000_ep00000085
56
7
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000086
24
3
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000087
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000088
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000089
16
2
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000090
24
3
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000091
24
3
train
{ "actions": [ 56, 23 ], "frames": [ 56 ], "pre": [ 56, 4096 ], "proprio": [ 56, 61 ], "segment_targets": [ 7, 4096 ], "vis": [ 56, 768 ] }
task0000_ep00000092
56
7
train
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000093
24
3
test
{ "actions": [ 24, 23 ], "frames": [ 24 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ], "vis": [ 24, 768 ] }
task0000_ep00000094
24
3
test
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000095
16
2
val
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000096
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000097
16
2
train
{ "actions": [ 16, 23 ], "frames": [ 16 ], "pre": [ 16, 4096 ], "proprio": [ 16, 61 ], "segment_targets": [ 2, 4096 ], "vis": [ 16, 768 ] }
task0000_ep00000098
16
2
val
{ "actions": [ 32, 23 ], "frames": [ 32 ], "pre": [ 32, 4096 ], "proprio": [ 32, 61 ], "segment_targets": [ 4, 4096 ], "vis": [ 32, 768 ] }
task0000_ep00000099
32
4
val
End of preview.

PLS-VLA skill features

Per-frame conditioned features e_t = Phi(f_t, L_sub^(j), L), mean-pooled skill latents S_j, aligned proprioception q_t, actions a_t, and subtask progress p_t. DINOv3 visual features V(f_t) are also included. Training inputs/targets for the Primitive Skill Composer VLA Skill Predictor. Versions from different trajectory corpora share this Hub repo; each folder is named {source}-task{NNNN} (e.g. b1k-task0000).

Source

  • source_dataset: b1k (BEHAVIOR-1K)
  • version: b1k-task0000-q0p15

Timelines

Aggregation chunk boundaries come from the keyframe detector. Per-frame subtask text and validity come from exact source-dataset GT intervals; annotation gaps are masked, and one chunk may contain multiple subtasks (subtask_source=behavior1k_skill).

This version

  • feature_source: vlm_prelogit
  • segment_source: keyframe
  • subtask_source: behavior1k_skill
  • schema_version: 3
  • sampling: fixed K=8 per chunk, both endpoints included
  • d_pre: 4096 d_vis: 768
  • samples: 8640, segments: 1080, episodes: 200
  • code revision: unknown created: 2026-08-17T23:55:37.090216+00:00

Reading these features honestly

K is part of every result. S_j is the mean of K sampled frames, so the zero-parameter predictor S_hat_t := e_t has error sigma^2 (K-1)/K against a floor of sigma^2/K. An identity gain quoted without its K is meaningless.

Feature sources are not comparable. hosted_embedding is a retrieval-trained embedding model reached over an API; vlm_prelogit is a VLM's last hidden state. They differ in width and in what they encode. Never pool them.

Chunk ≠ subtask. Per-frame subtask is the conditioning VLM interval; skill / segment_targets are pooled over keyframe chunk bounds. Mixed chunks are expected under dual timelines.

Do not mix source corpora blindly. Versions from different source_dataset ids may differ in embodiment, action space, and camera rig — treat them as separate distributions unless you deliberately align them.

Layout 

<version>/            # e.g. b1k-task0000
  config.json          provenance: source_dataset, models, K, git rev
  episodes.parquet     episode_id, task, instruction, split
  segments.parquet     aggregation chunks (keyframe bounds, is_mixed_subtask)
  subtasks.parquet     VLM-plan conditioning intervals (schema >= 3)
  manifest.parquet     one row per sample; chunk + subtask bounds, progress
  shards/<episode>/    pre.npy, vis.npy, segment_targets.npy, proprio.npy, actions.npy
                       frames.npz (sample-aligned JPEG f_t), meta.json

Use 

from pls_vla.hub import load_published_tensors

tensors = load_published_tensors("b1k-task0000-q0p15", split="train")
# -> SkillTensors(vis, pre, skill, progress[, subtask])
# subtask = per-frame VLM conditioning; skill = chunk-pooled S_j

Versions in this repo

Folder names encode the trajectory source ({source}-task{NNNN}).

  • b1k-task0000
  • b1k-task0000-q0.3
  • b1k-task0000-q0.65
  • b1k-task0000-q0p15
  • b1k-task0000-q0p15
Downloads last month
137
Total file size:
27.7 GB