Hugging Face's logo

Datasets:
DEEL-AI
/
Hackathon_Team01_v3

Tasks:
Robotics
Modalities:
Tabular
Video
Formats:
parquet
Size:
10K - 100K
Tags:
LeRobot
Libraries:
Datasets
pandas
License:
Dataset card Data Studio Files
xet
 Community
1
ampicard commited on
Commit
5debe34
·
verified ·
1 Parent(s): d60e9b2

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +211 -162
README.md CHANGED
@@ -1,162 +1,211 @@
1
- ---
2
- license: apache-2.0
3
- task_categories:
4
- - robotics
5
- tags:
6
- - LeRobot
7
- configs:
8
- - config_name: default
9
- data_files: data/*/*.parquet
10
- ---
11
-
12
- This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).
13
-
14
- ## Dataset Description
15
-
16
-
17
-
18
- - **Homepage:** [More Information Needed]
19
- - **Paper:** [More Information Needed]
20
- - **License:** apache-2.0
21
-
22
- ## Dataset Structure
23
-
24
- [meta/info.json](meta/info.json):
25
- ```json
26
- {
27
- "codebase_version": "v3.0",
28
- "robot_type": "so101_follower",
29
- "total_episodes": 50,
30
- "total_frames": 31189,
31
- "total_tasks": 1,
32
- "chunks_size": 1000,
33
- "data_files_size_in_mb": 100,
34
- "video_files_size_in_mb": 500,
35
- "fps": 30,
36
- "splits": {
37
- "train": "0:50"
38
- },
39
- "data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
40
- "video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4",
41
- "features": {
42
- "action": {
43
- "dtype": "float32",
44
- "names": [
45
- "shoulder_pan.pos",
46
- "shoulder_lift.pos",
47
- "elbow_flex.pos",
48
- "wrist_flex.pos",
49
- "wrist_roll.pos",
50
- "gripper.pos"
51
- ],
52
- "shape": [
53
- 6
54
- ]
55
- },
56
- "observation.state": {
57
- "dtype": "float32",
58
- "names": [
59
- "shoulder_pan.pos",
60
- "shoulder_lift.pos",
61
- "elbow_flex.pos",
62
- "wrist_flex.pos",
63
- "wrist_roll.pos",
64
- "gripper.pos"
65
- ],
66
- "shape": [
67
- 6
68
- ]
69
- },
70
- "observation.images.left": {
71
- "dtype": "video",
72
- "shape": [
73
- 480,
74
- 640,
75
- 3
76
- ],
77
- "names": [
78
- "height",
79
- "width",
80
- "channels"
81
- ],
82
- "info": {
83
- "video.height": 480,
84
- "video.width": 640,
85
- "video.codec": "av1",
86
- "video.pix_fmt": "yuv420p",
87
- "video.is_depth_map": false,
88
- "video.fps": 30,
89
- "video.channels": 3,
90
- "has_audio": false
91
- }
92
- },
93
- "observation.images.front": {
94
- "dtype": "video",
95
- "shape": [
96
- 480,
97
- 640,
98
- 3
99
- ],
100
- "names": [
101
- "height",
102
- "width",
103
- "channels"
104
- ],
105
- "info": {
106
- "video.height": 480,
107
- "video.width": 640,
108
- "video.codec": "av1",
109
- "video.pix_fmt": "yuv420p",
110
- "video.is_depth_map": false,
111
- "video.fps": 30,
112
- "video.channels": 3,
113
- "has_audio": false
114
- }
115
- },
116
- "timestamp": {
117
- "dtype": "float32",
118
- "shape": [
119
- 1
120
- ],
121
- "names": null
122
- },
123
- "frame_index": {
124
- "dtype": "int64",
125
- "shape": [
126
- 1
127
- ],
128
- "names": null
129
- },
130
- "episode_index": {
131
- "dtype": "int64",
132
- "shape": [
133
- 1
134
- ],
135
- "names": null
136
- },
137
- "index": {
138
- "dtype": "int64",
139
- "shape": [
140
- 1
141
- ],
142
- "names": null
143
- },
144
- "task_index": {
145
- "dtype": "int64",
146
- "shape": [
147
- 1
148
- ],
149
- "names": null
150
- }
151
- }
152
- }
153
- ```
154
-
155
-
156
- ## Citation
157
-
158
- **BibTeX:**
159
-
160
- ```bibtex
161
- [More Information Needed]
162
- ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ task_categories:
4
+ - robotics
5
+ tags:
6
+ - LeRobot
7
+ configs:
8
+ - config_name: default
9
+ data_files: data/*/*.parquet
10
+ ---
11
+
12
+ This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).
13
+
14
+ ## Dataset Description
15
+
16
+ This dataset contains the final batch of teleoperated demonstrations collected during a two-day hackathon using the LeRobot library and SO-101 robot arms in a leader–follower configuration.
17
+ Each episode shows the follower arm picking two colored cubes (one after the other) and placing each into the matching colored cross within a 2×2 grid. Two RGB cameras were used:
18
+
19
+ Top camera: mounted above the workspace for a clear 2D view of the arm, cubes, and grid.
20
+
21
+ Front/low camera: slightly above the ground, facing the arm and grid to provide better z-axis cues and arm self-perception.
22
+
23
+ Despite cardboard backgrounding, the room’s illumination varied over time and is deliberately preserved in the data, as it proved to be a limiting factor and may be valuable for robustness research.
24
+
25
+ This dataset is intended for vision-based imitation learning (e.g., behavior cloning, goal-conditioned policies), multi-view fusion, and perception-control studies on tabletop manipulation.
26
+
27
+ ### Use Cases
28
+
29
+ - **Imitation Learning**: Behavior cloning from teleop demonstrations.
30
+
31
+ - **Multiview Perception**: Fusing top + front perspectives for depth inference without explicit depth sensors.
32
+
33
+ - **Robustness to Lighting**: Evaluating policy sensitivity to illumination drift.
34
+
35
+ - **State–Action Alignment**: Leveraging synchronized proprioception and images.
36
+
37
+
38
+ ## Data Collection
39
+
40
+ ### Teleoperation Setup
41
+
42
+ - **Leader–Follower**: Human teleoperates a leader arm; follower SO-101 replicates motion to generate demonstrations.
43
+
44
+ - **Workspace**: Tabletop with a 2×2 grid. Each cell contains a colored cross; two colored cubes must be placed on matching crosses.
45
+
46
+ - **Cameras**:
47
+
48
+ - **Top**: overhead, full scene.
49
+
50
+ - **Front**: low angle, emphasizes depth and arm self-pose.
51
+
52
+ - **Background control**: Cardboard panels; lighting varies during the day and is preserved in data.
53
+
54
+ ### Episode Protocol
55
+
56
+ 1- Move to pre-grasp; localize target cube(s) visually.
57
+
58
+ 2- Grasp first cube; transport; place on correct colored cross.
59
+
60
+ 3- Repeat for second cube.
61
+
62
+ 4- Return to neutral.
63
+
64
+
65
+ ## Known limitations
66
+
67
+ - **Lighting drift**: Significant variation during the day; expect distribution shift. Consider color constancy or data augmentation.
68
+
69
+ - **Camera motion**: Cameras are fixed for the batch, but small nudges may occur; rely on metadata intrinsics/extrinsics if provided.
70
+
71
+ - **Occlusions**: Self-occlusion of the gripper and cubes in certain positions, especially from left camera during close approach.
72
+
73
+ - **No depth**: RGB only
74
+
75
+
76
+ - **Homepage:** [More Information Needed]
77
+ - **Paper:** [More Information Needed]
78
+ - **License:** apache-2.0
79
+
80
+ ## Dataset Structure
81
+
82
+ [meta/info.json](meta/info.json):
83
+ ```json
84
+ {
85
+ "codebase_version": "v3.0",
86
+ "robot_type": "so101_follower",
87
+ "total_episodes": 50,
88
+ "total_frames": 31189,
89
+ "total_tasks": 1,
90
+ "chunks_size": 1000,
91
+ "data_files_size_in_mb": 100,
92
+ "video_files_size_in_mb": 500,
93
+ "fps": 30,
94
+ "splits": {
95
+ "train": "0:50"
96
+ },
97
+ "data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
98
+ "video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4",
99
+ "features": {
100
+ "action": {
101
+ "dtype": "float32",
102
+ "names": [
103
+ "shoulder_pan.pos",
104
+ "shoulder_lift.pos",
105
+ "elbow_flex.pos",
106
+ "wrist_flex.pos",
107
+ "wrist_roll.pos",
108
+ "gripper.pos"
109
+ ],
110
+ "shape": [
111
+ 6
112
+ ]
113
+ },
114
+ "observation.state": {
115
+ "dtype": "float32",
116
+ "names": [
117
+ "shoulder_pan.pos",
118
+ "shoulder_lift.pos",
119
+ "elbow_flex.pos",
120
+ "wrist_flex.pos",
121
+ "wrist_roll.pos",
122
+ "gripper.pos"
123
+ ],
124
+ "shape": [
125
+ 6
126
+ ]
127
+ },
128
+ "observation.images.left": {
129
+ "dtype": "video",
130
+ "shape": [
131
+ 480,
132
+ 640,
133
+ 3
134
+ ],
135
+ "names": [
136
+ "height",
137
+ "width",
138
+ "channels"
139
+ ],
140
+ "info": {
141
+ "video.height": 480,
142
+ "video.width": 640,
143
+ "video.codec": "av1",
144
+ "video.pix_fmt": "yuv420p",
145
+ "video.is_depth_map": false,
146
+ "video.fps": 30,
147
+ "video.channels": 3,
148
+ "has_audio": false
149
+ }
150
+ },
151
+ "observation.images.front": {
152
+ "dtype": "video",
153
+ "shape": [
154
+ 480,
155
+ 640,
156
+ 3
157
+ ],
158
+ "names": [
159
+ "height",
160
+ "width",
161
+ "channels"
162
+ ],
163
+ "info": {
164
+ "video.height": 480,
165
+ "video.width": 640,
166
+ "video.codec": "av1",
167
+ "video.pix_fmt": "yuv420p",
168
+ "video.is_depth_map": false,
169
+ "video.fps": 30,
170
+ "video.channels": 3,
171
+ "has_audio": false
172
+ }
173
+ },
174
+ "timestamp": {
175
+ "dtype": "float32",
176
+ "shape": [
177
+ 1
178
+ ],
179
+ "names": null
180
+ },
181
+ "frame_index": {
182
+ "dtype": "int64",
183
+ "shape": [
184
+ 1
185
+ ],
186
+ "names": null
187
+ },
188
+ "episode_index": {
189
+ "dtype": "int64",
190
+ "shape": [
191
+ 1
192
+ ],
193
+ "names": null
194
+ },
195
+ "index": {
196
+ "dtype": "int64",
197
+ "shape": [
198
+ 1
199
+ ],
200
+ "names": null
201
+ },
202
+ "task_index": {
203
+ "dtype": "int64",
204
+ "shape": [
205
+ 1
206
+ ],
207
+ "names": null
208
+ }
209
+ }
210
+ }
211
+ ```