Hugging Face's logo

Datasets:
VGalaxies666
/
vlabench_primitive_ft_lerobot_video

Tasks:
Robotics
Modalities:
Tabular
Formats:
parquet
Size:
1K - 10K
Tags:
lerobot
robotics
vlabench
manipulation
benchmark:eval-yaml
tsfile
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
episode_index
int64
0
5k
⌀
length
int64
52
232
⌀
dataset_from_index
int64
0
947k
⌀
dataset_to_index
int64
93
948k
⌀
videos/image/chunk_index
int64
0
4
⌀
videos/image/file_index
int64
0
999
⌀
videos/image/from_timestamp
float64
0
0
⌀
videos/image/to_timestamp
float64
5.2
23.2
⌀
videos/second_image/chunk_index
int64
0
4
⌀
videos/second_image/file_index
int64
0
999
⌀
videos/second_image/from_timestamp
float64
0
0
⌀
videos/second_image/to_timestamp
float64
5.2
23.2
⌀
videos/wrist_image/chunk_index
int64
0
4
⌀
videos/wrist_image/file_index
int64
0
999
⌀
videos/wrist_image/from_timestamp
float64
0
0
⌀
videos/wrist_image/to_timestamp
float64
5.2
23.2
⌀
data/chunk_index
int64
0
3
⌀
data/file_index
int64
0
999
⌀
0
93
0
93
0
0
0
9.3
0
0
0
9.3
0
0
0
9.3
0
0
1
74
74
148
0
1
0
7.4
0
1
0
7.4
0
1
0
7.4
0
0
2
78
156
234
0
2
0
7.8
0
2
0
7.8
0
2
0
7.8
0
0
3
105
315
420
0
3
0
10.5
0
3
0
10.5
0
3
0
10.5
0
1
4
77
308
385
0
4
0
7.7
0
4
0
7.7
0
4
0
7.7
0
1
5
73
365
438
0
5
0
7.3
0
5
0
7.3
0
5
0
7.3
0
2
6
78
468
546
0
6
0
7.8
0
6
0
7.8
0
6
0
7.8
0
2
7
58
406
464
0
7
0
5.8
0
7
0
5.8
0
7
0
5.8
0
3
8
180
1,440
1,620
0
8
0
18
0
8
0
18
0
8
0
18
0
3
9
76
684
760
0
9
0
7.6
0
9
0
7.6
0
9
0
7.6
0
4
10
104
1,040
1,144
0
10
0
10.4
0
10
0
10.4
0
10
0
10.4
0
5
11
122
1,342
1,464
0
11
0
12.2
0
11
0
12.2
0
11
0
12.2
0
6
12
95
1,140
1,235
0
12
0
9.5
0
12
0
9.5
0
12
0
9.5
0
7
13
66
858
924
0
13
0
6.6
0
13
0
6.6
0
13
0
6.6
0
7
14
97
1,358
1,455
0
14
0
9.7
0
14
0
9.7
0
14
0
9.7
0
8
15
154
2,310
2,464
0
15
0
15.4
0
15
0
15.4
0
15
0
15.4
0
8
16
107
1,712
1,819
0
16
0
10.7
0
16
0
10.7
0
16
0
10.7
0
9
17
66
1,122
1,188
0
17
0
6.6
0
17
0
6.6
0
17
0
6.6
0
10
18
137
2,466
2,603
0
18
0
13.7
0
18
0
13.7
0
18
0
13.7
0
10
19
85
1,615
1,700
0
19
0
8.5
0
19
0
8.5
0
19
0
8.5
0
11
20
76
1,520
1,596
0
20
0
7.6
0
20
0
7.6
0
20
0
7.6
0
12
21
152
3,192
3,344
0
21
0
15.2
0
21
0
15.2
0
21
0
15.2
0
12
22
179
3,938
4,117
0
22
0
17.9
0
22
0
17.9
0
22
0
17.9
0
13
23
68
1,564
1,632
0
23
0
6.8
0
23
0
6.8
0
23
0
6.8
0
14
24
62
1,488
1,550
0
24
0
6.2
0
24
0
6.2
0
24
0
6.2
0
15
25
156
3,900
4,056
0
25
0
15.6
0
25
0
15.6
0
25
0
15.6
0
15
26
130
3,380
3,510
0
26
0
13
0
26
0
13
0
26
0
13
0
16
27
185
4,995
5,180
0
27
0
18.5
0
27
0
18.5
0
27
0
18.5
0
17
28
69
1,932
2,001
0
28
0
6.9
0
28
0
6.9
0
28
0
6.9
0
18
29
72
2,088
2,160
0
29
0
7.2
0
29
0
7.2
0
29
0
7.2
0
19
30
98
2,940
3,038
0
30
0
9.8
0
30
0
9.8
0
30
0
9.8
0
19
31
84
2,604
2,688
0
31
0
8.4
0
31
0
8.4
0
31
0
8.4
0
19
32
120
3,840
3,960
0
32
0
12
0
32
0
12
0
32
0
12
0
20
33
179
5,907
6,086
0
33
0
17.9
0
33
0
17.9
0
33
0
17.9
0
21
34
83
2,822
2,905
0
34
0
8.3
0
34
0
8.3
0
34
0
8.3
0
22
35
76
2,660
2,736
0
35
0
7.6
0
35
0
7.6
0
35
0
7.6
0
23
36
105
3,780
3,885
0
36
0
10.5
0
36
0
10.5
0
36
0
10.5
0
23
37
127
4,699
4,826
0
37
0
12.7
0
37
0
12.7
0
37
0
12.7
0
24
38
143
5,434
5,577
0
38
0
14.3
0
38
0
14.3
0
38
0
14.3
0
25
39
159
6,201
6,360
0
39
0
15.9
0
39
0
15.9
0
39
0
15.9
0
26
40
145
5,800
5,945
0
40
0
14.5
0
40
0
14.5
0
40
0
14.5
0
27
41
91
3,731
3,822
0
41
0
9.1
0
41
0
9.1
0
41
0
9.1
0
28
42
67
2,814
2,881
0
42
0
6.7
0
42
0
6.7
0
42
0
6.7
0
29
43
136
5,848
5,984
0
43
0
13.6
0
43
0
13.6
0
43
0
13.6
0
29
44
164
7,216
7,380
0
44
0
16.4
0
44
0
16.4
0
44
0
16.4
0
30
45
147
6,615
6,762
0
45
0
14.7
0
45
0
14.7
0
45
0
14.7
0
31
46
71
3,266
3,337
0
46
0
7.1
0
46
0
7.1
0
46
0
7.1
0
32
47
60
2,820
2,880
0
47
0
6
0
47
0
6
0
47
0
6
0
33
48
158
7,584
7,742
0
48
0
15.8
0
48
0
15.8
0
48
0
15.8
0
33
49
74
3,626
3,700
0
49
0
7.4
0
49
0
7.4
0
49
0
7.4
0
34
50
98
4,900
4,998
0
50
0
9.8
0
50
0
9.8
0
50
0
9.8
0
35
51
152
7,752
7,904
0
51
0
15.2
0
51
0
15.2
0
51
0
15.2
0
35
52
92
4,784
4,876
0
52
0
9.2
0
52
0
9.2
0
52
0
9.2
0
36
53
158
8,374
8,532
0
53
0
15.8
0
53
0
15.8
0
53
0
15.8
0
37
54
87
4,698
4,785
0
54
0
8.7
0
54
0
8.7
0
54
0
8.7
0
38
55
99
5,445
5,544
0
55
0
9.9
0
55
0
9.9
0
55
0
9.9
0
39
56
76
4,256
4,332
0
56
0
7.6
0
56
0
7.6
0
56
0
7.6
0
39
57
159
9,063
9,222
0
57
0
15.9
0
57
0
15.9
0
57
0
15.9
0
40
58
154
8,932
9,086
0
58
0
15.4
0
58
0
15.4
0
58
0
15.4
0
41
59
66
3,894
3,960
0
59
0
6.6
0
59
0
6.6
0
59
0
6.6
0
42
60
94
5,640
5,734
0
60
0
9.4
0
60
0
9.4
0
60
0
9.4
0
42
61
115
7,015
7,130
0
61
0
11.5
0
61
0
11.5
0
61
0
11.5
0
43
62
64
3,968
4,032
0
62
0
6.4
0
62
0
6.4
0
62
0
6.4
0
44
63
158
9,954
10,112
0
63
0
15.8
0
63
0
15.8
0
63
0
15.8
0
44
64
74
4,736
4,810
0
64
0
7.4
0
64
0
7.4
0
64
0
7.4
0
45
65
79
5,135
5,214
0
65
0
7.9
0
65
0
7.9
0
65
0
7.9
0
46
66
78
5,148
5,226
0
66
0
7.8
0
66
0
7.8
0
66
0
7.8
0
46
67
163
10,921
11,084
0
67
0
16.3
0
67
0
16.3
0
67
0
16.3
0
47
68
83
5,644
5,727
0
68
0
8.3
0
68
0
8.3
0
68
0
8.3
0
48
69
168
11,592
11,760
0
69
0
16.8
0
69
0
16.8
0
69
0
16.8
0
49
70
117
8,190
8,307
0
70
0
11.7
0
70
0
11.7
0
70
0
11.7
0
50
71
164
11,644
11,808
0
71
0
16.4
0
71
0
16.4
0
71
0
16.4
0
51
72
107
7,704
7,811
0
72
0
10.7
0
72
0
10.7
0
72
0
10.7
0
52
73
142
10,366
10,508
0
73
0
14.2
0
73
0
14.2
0
73
0
14.2
0
53
74
171
12,654
12,825
0
74
0
17.1
0
74
0
17.1
0
74
0
17.1
0
54
75
81
6,075
6,156
0
75
0
8.1
0
75
0
8.1
0
75
0
8.1
0
55
76
66
5,016
5,082
0
76
0
6.6
0
76
0
6.6
0
76
0
6.6
0
56
77
122
9,394
9,516
0
77
0
12.2
0
77
0
12.2
0
77
0
12.2
0
56
78
181
14,118
14,299
0
78
0
18.1
0
78
0
18.1
0
78
0
18.1
0
57
79
124
9,796
9,920
0
79
0
12.4
0
79
0
12.4
0
79
0
12.4
0
58
80
202
16,160
16,362
0
80
0
20.2
0
80
0
20.2
0
80
0
20.2
0
59
81
195
15,795
15,990
0
81
0
19.5
0
81
0
19.5
0
81
0
19.5
0
60
82
109
8,938
9,047
0
82
0
10.9
0
82
0
10.9
0
82
0
10.9
0
61
83
163
13,529
13,692
0
83
0
16.3
0
83
0
16.3
0
83
0
16.3
0
62
84
136
11,424
11,560
0
84
0
13.6
0
84
0
13.6
0
84
0
13.6
0
63
85
172
14,620
14,792
0
85
0
17.2
0
85
0
17.2
0
85
0
17.2
0
64
86
89
7,654
7,743
0
86
0
8.9
0
86
0
8.9
0
86
0
8.9
0
65
87
172
14,964
15,136
0
87
0
17.2
0
87
0
17.2
0
87
0
17.2
0
66
88
151
13,288
13,439
0
88
0
15.1
0
88
0
15.1
0
88
0
15.1
0
67
89
78
6,942
7,020
0
89
0
7.8
0
89
0
7.8
0
89
0
7.8
0
68
90
116
10,440
10,556
0
90
0
11.6
0
90
0
11.6
0
90
0
11.6
0
69
91
69
6,279
6,348
0
91
0
6.9
0
91
0
6.9
0
91
0
6.9
0
69
92
186
17,112
17,298
0
92
0
18.6
0
92
0
18.6
0
92
0
18.6
0
70
93
73
6,789
6,862
0
93
0
7.3
0
93
0
7.3
0
93
0
7.3
0
71
94
148
13,912
14,060
0
94
0
14.8
0
94
0
14.8
0
94
0
14.8
0
72
95
92
8,740
8,832
0
95
0
9.2
0
95
0
9.2
0
95
0
9.2
0
73
96
68
6,528
6,596
0
96
0
6.8
0
96
0
6.8
0
96
0
6.8
0
74
97
139
13,483
13,622
0
97
0
13.9
0
97
0
13.9
0
97
0
13.9
0
74
98
107
10,486
10,593
0
98
0
10.7
0
98
0
10.7
0
98
0
10.7
0
75
99
145
14,355
14,500
0
99
0
14.5
0
99
0
14.5
0
99
0
14.5
0
76
End of preview. Expand in Data Studio

VLABench Primitive Tasks — LeRobot v3.0 (TsFile)

Apache TsFile version of VLABench/vlabench_primitive_ft_lerobot_video.

Overview

This dataset is organized in the LeRobot v3.0 format and is used for integrating VLABench into the LeRobot framework officially. Compared with the v2.0 and the RLDS versions, this release stores the visual observations in a video-compressed format rather than as individual image files, giving better storage efficiency and data-loading performance.

It contains demonstrations for 10 primitive manipulation tasks on a Franka Panda arm, collected across 128 concrete language-instructed task variants (e.g. "Pick up the mahjong of 1 pin", "Insert the rose into the vase"). Models trained on this data can be evaluated on multiple generalization tracks in VLABench.

The 10 primitive task categories are: add_condiment, insert_flower, select_book, select_chemistry_tube, select_drink, select_fruit, select_mahjong, select_painting, select_poker, select_toy.

  • Robot: Franka Panda, 7-DoF.
  • Episodes: 5,000 (episode_index 0–4999).
  • Frames: 575,101 time-series rows.
  • Tasks: 128 language-instructed variants (task_index 0–127); the full task_index → instruction mapping is in meta/tasks.parquet.
  • Sampling rate: 10 fps (from meta/info.json).

Baseline models: pi05-primitive-10task, pi0-fast-primitive-10task, pi0-primitive-10task.

Schema (TsFile structure)

All 5,000 episodes are stored in a single TsFile table; episode_index and task_index are TAG columns (the TsFile device dimension), so one episode is selected with WHERE episode_index=0.

  • Time (INT64, milliseconds) — round(timestamp * 1000); restarts at 0 for each episode, stepping by 100 ms (10 fps).
  • episode_index (TAG) — source episode id, 0–4999.
  • task_index (TAG) — source task id, 0–127; resolve to the instruction text via meta/tasks.parquet.
  • frame_index (FIELD, INT64) — frame number within the episode.
  • sample_index (FIELD, INT64) — the source global index column, renamed.
  • state_0 … state_6 (FIELD, FLOAT) — robot proprioceptive state, flattened from the 7-element state vector.
  • actions_0 … actions_6 (FIELD, FLOAT) — action command, flattened from the 7-element actions vector.

The source timestamp column is dropped because it equals Time / 1000 seconds. No other columns or rows are dropped.

Videos

The three camera video streams of the original dataset (image, second_image, wrist_image) are NOT included in this repository. They are large and non–time-series; obtain them from the original dataset: https://huggingface.co/datasets/VLABench/vlabench_primitive_ft_lerobot_video (the videos/ directory).

Usage

Read the .tsfile file with the Apache TsFile Java or Python SDK.

Source & license

Downloads last month
41
Number of rows:
5,128
Total file size:
29.7 MB