Hugging Face's logo

Datasets:
erl-hub
/
behaviour1k-Qwen3-features

Tasks:
Robotics
Modalities:
Tabular
Text
Size:
1K - 10K
Tags:
behavior-1k
vision-language-action
skill-latents
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
arrays
dict
episode_id
stringlengths
19
19
n_samples
int64
8
80
n_segments
int64
1
10
split
stringclasses
3 values
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000000
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000001
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000002
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000003
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000004
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000005
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000006
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000007
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000008
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000009
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000010
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000011
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000012
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000013
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000014
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000015
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000016
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000017
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000018
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000019
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000020
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000021
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000022
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000023
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000024
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000025
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000026
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000027
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000028
24
3
test
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000029
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000030
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000031
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000032
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000033
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000034
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000035
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000036
24
3
test
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000037
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000038
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000039
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000040
24
3
val
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000041
24
3
val
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000042
24
3
test
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000043
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000044
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000045
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000046
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000047
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000048
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000049
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000050
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000051
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000052
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000053
24
3
val
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000054
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000055
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000056
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000057
24
3
val
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000058
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000059
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000060
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000061
24
3
val
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000062
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000063
24
3
val
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000064
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000065
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000066
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000067
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000068
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000069
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000070
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000071
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000072
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000073
24
3
test
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000074
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000075
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000076
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000077
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000078
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000079
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000080
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000081
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000082
24
3
test
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000083
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000084
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000085
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000086
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000087
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000088
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000089
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000090
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000091
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000092
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000093
24
3
test
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000094
24
3
test
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000095
24
3
val
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000096
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000097
24
3
train
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000098
24
3
val
{ "actions": [ 24, 23 ], "pre": [ 24, 4096 ], "proprio": [ 24, 61 ], "segment_targets": [ 3, 4096 ] }
task0000_ep00000099
24
3
val
End of preview. Expand in Data Studio

BEHAVIOR-1K Qwen3 skill features

Per-frame conditioned features e_t = Phi(f_t, L_sub^(j), L), mean-pooled primitive skill latents S_j, aligned proprioception q_t, actions a_t, and subtask progress p_t. These are the inputs and targets for a Primitive Skill Composer VLA Skill Predictor.

Ground-truth primitives come from BEHAVIOR-1K's hand-authored primitive_annotation, so the segmentation is human-labelled rather than predicted, and nothing here depends on a keyframe detector.

This version

  • version: v1-hosted-task0022
  • feature_source: hosted_embedding
  • subtask_source: behavior1k_primitive
  • sampling: fixed K=8 per primitive, both endpoints included
  • d_pre: 4096 d_vis: None
  • samples: 3200, segments: 400, episodes: 200
  • code revision: unknown created: 2026-08-03T17:15:24.562363+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.

Layout 

<version>/
  config.json          provenance: models, K, feature/subtask source, git rev
  episodes.parquet     episode_id, task, instruction, split
  segments.parquet     episode_id, segment_id, start/end_frame, L_sub
  manifest.parquet     one row per sample; frame_idx, progress, indices
  shards/<episode>/    pre.npy, vis.npy, segment_targets.npy, meta.json

Use 

from pls_vla.hub import load_published_tensors

tensors = load_published_tensors("v1-hosted-task0022", split="train")
# -> SkillTensors(vis, pre, skill, progress)

Versions in this repo

  • v1-hosted-task0000
  • v1-hosted-task0001
  • v1-hosted-task0002
  • v1-hosted-task0003
  • v1-hosted-task0004
  • v1-hosted-task0005
  • v1-hosted-task0006
  • v1-hosted-task0007
  • v1-hosted-task0008
  • v1-hosted-task0009
  • v1-hosted-task0010
  • v1-hosted-task0011
  • v1-hosted-task0012
  • v1-hosted-task0015
  • v1-hosted-task0016
  • v1-hosted-task0017
  • v1-hosted-task0018
  • v1-hosted-task0022
  • v1-hosted-task0022
Downloads last month
185
Number of rows:
3,600
Total file size:
1.26 GB