yxma commited on
Commit
53f3821
·
verified ·
1 Parent(s): cfda4b8

Drop 2026-03-23 right-only pilot (outdated); recompute totals across 27 bimanual recordings

Browse files
README.md CHANGED
@@ -33,7 +33,7 @@ Dense, contact-rich, synchronized multimodal interaction data collected from **h
33
 
34
  ![Tactile intensity timeline](figures/contact_intensity_full.png)
35
 
36
- > **138 min of robot-free human-hand multimodal interaction · 88 min (66 %) of confirmed bimanual tactile contact · 4,136 distinct contact events · 240 k frames @ 30 Hz across 3 × RGB-D + 2 × GelSight + 3-body OptiTrack**
37
 
38
  ## What's different about this dataset
39
 
@@ -52,8 +52,8 @@ Dense, contact-rich, synchronized multimodal interaction data collected from **h
52
  |---|---|
53
  | Embodiment | **Human hands (no robot)** — handheld GelSight sensors with motion-capture rigid bodies |
54
  | Intended use | Dynamics / world-model learning over short multimodal windows. Sample short trajectories (1 s – 10 s); recording-file boundaries are not action boundaries. |
55
- | Total synchronized duration | **138.4 min** at 30 Hz (239,759 multimodal frames) |
56
- | Bimanual tactile-contact time | **87.9 min — 66 % of frames** (4,136 contact events, median 0.73 s) |
57
  | Cameras | 3× Intel RealSense D415 (color + depth), 480×640, 30 FPS |
58
  | Tactile | 2× GelSight Mini (left, right), handheld |
59
  | Motion capture | OptiTrack VRPN, 3 rigid bodies, ~120 Hz |
@@ -64,7 +64,6 @@ Dense, contact-rich, synchronized multimodal interaction data collected from **h
64
 
65
  | Date | Kind | Active sensors | Notes |
66
  |---|---|---|---|
67
- | 2026-03-23 | pilot | **right only** | Early trials. The left GelSight and left rigid body were not in use; ignore `tactile_left*` and `sensor_left_pose` for these recordings. |
68
  | 2026-05-10 | session | left + right | First full bimanual session. |
69
  | 2026-05-11 | session | left + right | Largest session. A handful of GelSight LED-flicker frames + one mocap teleport; see [`bad_frames.json`](bad_frames.json). |
70
 
@@ -104,7 +103,6 @@ import json
104
  with open("bad_frames.json") as f:
105
  bad = json.load(f)["episodes"]
106
  # Drop ~0.085 % of frames flagged in bad_frames.json — see docs/quality.md
107
- # For 2026-03-23 recordings, also ignore the left-sensor fields (right-only pilot).
108
  ```
109
 
110
  ## Example dataloader — short contact-rich windows
@@ -127,13 +125,13 @@ ds = ReactWindowDataset(
127
  min_contact_fraction=0.6, # ≥ 60 % of window frames must have contact
128
  which_sensors="any", # "any" | "both" | "left" | "right"
129
  skip_bad_frames=True,
130
- respect_active_sensors=True, # mask out left modalities for 2026-03-23 pilot
131
  )
132
  print(len(ds), "windows")
133
  loader = DataLoader(ds, batch_size=8, shuffle=True, num_workers=2)
134
  ```
135
 
136
- With the defaults shown above, the dataset assembles **9,230 contact-rich 16-frame windows** across the 27 bimanual recordings. Each sample is a dict of `(T, …)` tensors plus metadata (`episode`, `frame_start`, `active_sensors`, …).
137
 
138
  ### Example output
139
 
 
33
 
34
  ![Tactile intensity timeline](figures/contact_intensity_full.png)
35
 
36
+ > **126 min of robot-free human-hand multimodal interaction · 81 min (66 %) of confirmed bimanual tactile contact · 221,621 frames @ 30 Hz across 3 × RGB-D + 2 × GelSight + 3-body OptiTrack**
37
 
38
  ## What's different about this dataset
39
 
 
52
  |---|---|
53
  | Embodiment | **Human hands (no robot)** — handheld GelSight sensors with motion-capture rigid bodies |
54
  | Intended use | Dynamics / world-model learning over short multimodal windows. Sample short trajectories (1 s – 10 s); recording-file boundaries are not action boundaries. |
55
+ | Total synchronized duration | **126.0 min** at 30 Hz (221,621 multimodal frames) |
56
+ | Bimanual tactile-contact time | **81.4 min — 66 % of frames** (median event duration 0.73 s) |
57
  | Cameras | 3× Intel RealSense D415 (color + depth), 480×640, 30 FPS |
58
  | Tactile | 2× GelSight Mini (left, right), handheld |
59
  | Motion capture | OptiTrack VRPN, 3 rigid bodies, ~120 Hz |
 
64
 
65
  | Date | Kind | Active sensors | Notes |
66
  |---|---|---|---|
 
67
  | 2026-05-10 | session | left + right | First full bimanual session. |
68
  | 2026-05-11 | session | left + right | Largest session. A handful of GelSight LED-flicker frames + one mocap teleport; see [`bad_frames.json`](bad_frames.json). |
69
 
 
103
  with open("bad_frames.json") as f:
104
  bad = json.load(f)["episodes"]
105
  # Drop ~0.085 % of frames flagged in bad_frames.json — see docs/quality.md
 
106
  ```
107
 
108
  ## Example dataloader — short contact-rich windows
 
125
  min_contact_fraction=0.6, # ≥ 60 % of window frames must have contact
126
  which_sensors="any", # "any" | "both" | "left" | "right"
127
  skip_bad_frames=True,
128
+ respect_active_sensors=True,
129
  )
130
  print(len(ds), "windows")
131
  loader = DataLoader(ds, batch_size=8, shuffle=True, num_workers=2)
132
  ```
133
 
134
+ With the defaults shown above, the dataset assembles **~9.2 k contact-rich 16-frame windows** across the 27 recordings. Each sample is a dict of `(T, …)` tensors plus metadata (`episode`, `frame_start`, `active_sensors`, …).
135
 
136
  ### Example output
137
 
bad_frames.json CHANGED
@@ -3,84 +3,13 @@
3
  "tau_velocity_mps": 5.0,
4
  "buffer_frames": 3,
5
  "summary": {
6
- "n_episodes": 30,
7
- "total_frames": 239759,
8
- "total_bad_frames": 202,
9
- "bad_fraction_overall": 0.0008425126898260337,
10
- "n_episodes_with_bad_frames": 10
11
  },
12
  "episodes": {
13
- "2026-03-23/episode_000": {
14
- "n_frames": 2141,
15
- "duration_s": 72.09878873825073,
16
- "intensity_spikes": [],
17
- "pose_teleports_L": [],
18
- "pose_teleports_R": [],
19
- "total_bad_frames": 0,
20
- "bad_fraction": 0.0
21
- },
22
- "2026-03-23/episode_001": {
23
- "n_frames": 7142,
24
- "duration_s": 295.3777325153351,
25
- "intensity_spikes": [],
26
- "pose_teleports_L": [
27
- [
28
- 6476,
29
- 6483
30
- ],
31
- [
32
- 6974,
33
- 6981
34
- ]
35
- ],
36
- "pose_teleports_R": [
37
- [
38
- 6734,
39
- 6741
40
- ],
41
- [
42
- 7122,
43
- 7129
44
- ]
45
- ],
46
- "total_bad_frames": 32,
47
- "bad_fraction": 0.004480537664519743
48
- },
49
- "2026-03-23/episode_002": {
50
- "n_frames": 8855,
51
- "duration_s": 372.1804549694061,
52
- "intensity_spikes": [],
53
- "pose_teleports_L": [
54
- [
55
- 4378,
56
- 4385
57
- ],
58
- [
59
- 4478,
60
- 4485
61
- ]
62
- ],
63
- "pose_teleports_R": [
64
- [
65
- 4538,
66
- 4545
67
- ],
68
- [
69
- 4558,
70
- 4565
71
- ],
72
- [
73
- 4598,
74
- 4605
75
- ],
76
- [
77
- 8621,
78
- 8628
79
- ]
80
- ],
81
- "total_bad_frames": 48,
82
- "bad_fraction": 0.005420666290231507
83
- },
84
  "2026-05-10/episode_000": {
85
  "n_frames": 6891,
86
  "duration_s": 235.86456942558289,
 
3
  "tau_velocity_mps": 5.0,
4
  "buffer_frames": 3,
5
  "summary": {
6
+ "n_episodes": 27,
7
+ "total_frames": 221621,
8
+ "total_bad_frames": 122,
9
+ "bad_fraction_overall": 0.0005504893489335396,
10
+ "n_episodes_with_bad_frames": 8
11
  },
12
  "episodes": {
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13
  "2026-05-10/episode_000": {
14
  "n_frames": 6891,
15
  "duration_s": 235.86456942558289,
docs/caveats.md CHANGED
@@ -9,18 +9,6 @@ React is a **dense multimodal interaction stream** intended for learning **dynam
9
  - A demonstration / policy-learning dataset. "Episodes" here are just file boundaries — they don't carry semantic / action structure.
10
  - Comparable apples-to-apples with BridgeData V2, DROID, RT-1, ALOHA, etc. on "number of demos." The relevant comparison is *hours of synchronized multimodal contact-rich interaction*.
11
 
12
- ## Per-session caveats
13
-
14
- ### 2026-03-23 — right-sensor-only pilot
15
-
16
- The first 3 recording files (2026-03-23/episode_{000,001,002}) are **early pilot trials with only the right GelSight sensor in use**. For these files:
17
-
18
- - `tactile_left` is essentially flat (peak intensity ≤ 1.4 vs ≥ 12 when actively used). Treat as empty.
19
- - `sensor_left_pose` is not informative — the left rigid body was either not mounted or stationary off-workspace; the few "pose teleport" events flagged in `bad_frames.json` for these files are mocap noise on an unused tracker.
20
- - `tactile_right` and `sensor_right_pose` are valid and contribute ~7.5 min of useful single-sensor contact data.
21
-
22
- Dataloaders consuming these files should either mask the left modalities or skip the date entirely if they require bimanual data. The `active_sensors` field in [`../tasks.json`](../tasks.json) records this explicitly.
23
-
24
  ## Other caveats
25
 
26
  - **Missing / dropped files** on `motherboard/2026-05-11`:
 
9
  - A demonstration / policy-learning dataset. "Episodes" here are just file boundaries — they don't carry semantic / action structure.
10
  - Comparable apples-to-apples with BridgeData V2, DROID, RT-1, ALOHA, etc. on "number of demos." The relevant comparison is *hours of synchronized multimodal contact-rich interaction*.
11
 
 
 
 
 
 
 
 
 
 
 
 
 
12
  ## Other caveats
13
 
14
  - **Missing / dropped files** on `motherboard/2026-05-11`:
docs/quality.md CHANGED
@@ -6,24 +6,20 @@ A small fraction of frames contain known sensor artifacts. The repo ships a [`..
6
 
7
  | | |
8
  |---|---:|
9
- | Total synchronized frames | 239,759 (138.4 min @ 30 Hz) |
10
- | Recording files | 30 |
11
- | Frames flagged in `bad_frames.json` | **202 (0.084 %)** |
12
- | Recording files with ≥1 flagged frame | 10 |
13
 
14
  ## Failure modes — quantitative breakdown
15
 
16
  | # | Mode | Frames | % of dataset | Files | Symptom | Example |
17
  |---|---|---:|---:|---:|---|---|
18
- | 1 | **GelSight LED flicker** | 66 | **0.028 %** | 5 / 30 | 1–2 frames of uniform pink/magenta wash across the gel surface; adjacent frames normal. 2026-05-11 / {003, 007, 008, 011, 017}. | [`intensity_spike_overview.png`](../figures/dataset_figures/intensity_spike_samples/intensity_spike_overview.png) |
19
- | 2 | **OptiTrack pose teleport (>5 m/s)** | 136 | **0.057 %** | 5 / 30 | Position jumps by >5 cm in 33 ms — mocap lost lock and reacquired. 2026-03-23/{001, 002} (these are on the unused **left** tracker — see Notes), 2026-05-10/{001, 002}, 2026-05-11/015. | [`pose_teleport_samples/`](../figures/dataset_figures/pose_teleport_samples/) |
20
- | 3 | Sensor at rest (>30 s freeze) | 37,601 | **15.68 %** | 5 / 30 | Multi-minute periods with pose const and tactile const (sensor set down on the table). **Not corruption**, but degenerate for dynamics learning. 2026-05-11/{012: 5.4 min, 017: 10.7 min, 005: 1.3 min}, 2026-03-23/{001: 1.8 min, 002: 1.6 min} (latter two are on the unused left tracker). | [`data_quality_report.png`](../figures/dataset_figures/data_quality_report.png) |
21
 
22
- **Modes 1+2 together: 202 / 239,759 = 0.084 % of frames are corrupted in any modality.** Mode 3 is tracked separately because it is healthy data, just non-informative for dynamics. `bad_frames.json` covers modes 1 and 2 only; if you also want to skip rest periods, intersect with the velocity track from `sensor_*_pose`.
23
-
24
- ### Notes on the 2026-03-23 entries
25
-
26
- `2026-03-23/episode_{001,002}` were **right-sensor-only pilot trials** (see [`caveats.md`](caveats.md)). The "pose teleport" flags on those files are on the **left** tracker, which was not in use, so the position values were essentially random mocap noise. If you are masking out the left modalities for those files (which you should), these teleport intervals can also be ignored.
27
 
28
  ## Inspection figures
29
 
@@ -33,12 +29,7 @@ A small fraction of frames contain known sensor artifacts. The repo ships a [`..
33
 
34
  Per-file close-ups (reference frame, ±1 s neighbors, peak frame): [`figures/dataset_figures/intensity_spike_samples/`](../figures/dataset_figures/intensity_spike_samples/).
35
 
36
- **Mode 2 — OptiTrack pose teleport** (2 examples for 2026-03-23, where only the processed `.pt` is on the publishing machine; the other 4 affected files have GIFs in the same folder):
37
-
38
- ![Teleport 2026-03-23 ep001](../figures/dataset_figures/pose_teleport_samples/2026-03-23/episode_001_teleport.png)
39
- ![Teleport 2026-03-23 ep002](../figures/dataset_figures/pose_teleport_samples/2026-03-23/episode_002_teleport.png)
40
-
41
- GIFs (10 s playback at 2× speed) for the other 4: [`pose_teleport_samples/2026-05-{10,11}/`](../figures/dataset_figures/pose_teleport_samples/).
42
 
43
  **Modes 1+2+3 — top-6 worst-offenders chart** (tactile intensity + sensor velocity time-series):
44
 
@@ -54,11 +45,11 @@ Frame indices are inclusive on both ends and pre-padded by `buffer_frames` (3) o
54
  "tau_velocity_mps": 5.0,
55
  "buffer_frames": 3,
56
  "summary": {
57
- "n_episodes": 30,
58
- "total_frames": 239759,
59
- "total_bad_frames": 202,
60
- "bad_fraction_overall": 0.000843,
61
- "n_episodes_with_bad_frames": 10
62
  },
63
  "episodes": {
64
  "2026-05-11/episode_003": {
 
6
 
7
  | | |
8
  |---|---:|
9
+ | Total synchronized frames | 221,621 (126.0 min @ 30 Hz) |
10
+ | Recording files | 27 |
11
+ | Frames flagged in `bad_frames.json` | **122 (0.055 %)** |
12
+ | Recording files with ≥1 flagged frame | 8 |
13
 
14
  ## Failure modes — quantitative breakdown
15
 
16
  | # | Mode | Frames | % of dataset | Files | Symptom | Example |
17
  |---|---|---:|---:|---:|---|---|
18
+ | 1 | **GelSight LED flicker** | 66 | **0.030 %** | 5 / 27 | 1–2 frames of uniform pink/magenta wash across the gel surface; adjacent frames normal. 2026-05-11 / {003, 007, 008, 011, 017}. | [`intensity_spike_overview.png`](../figures/dataset_figures/intensity_spike_samples/intensity_spike_overview.png) |
19
+ | 2 | **OptiTrack pose teleport (>5 m/s)** | 64 | **0.029 %** | 3 / 27 | Position jumps by >5 cm in 33 ms — mocap lost lock and reacquired. 2026-05-10/{001, 002}, 2026-05-11/015. | [`pose_teleport_samples/`](../figures/dataset_figures/pose_teleport_samples/) |
20
+ | 3 | OptiTrack pose freeze ( 5 s held) | 31,376 | **14.16 %** | 3 / 27 | Held-pose intervals on an active sensor OptiTrack lost track and the recorder held the last sample. 2026-05-11/{017: 10.7 min, 012: 5.4 min, 005: 1.3 min}. **Not corruption**, but degenerate for dynamics; left-sensor only. | [`data_quality_report.png`](../figures/dataset_figures/data_quality_report.png) |
21
 
22
+ **Modes 1+2 together: 122 / 221621 = 0.055 % of frames are corrupted in any modality.** Mode 3 is tracked separately because it is healthy data, just non-informative for dynamics. `bad_frames.json` covers modes 1 and 2 only; if you also want to skip rest periods, intersect with the velocity track from `sensor_*_pose`.
 
 
 
 
23
 
24
  ## Inspection figures
25
 
 
29
 
30
  Per-file close-ups (reference frame, ±1 s neighbors, peak frame): [`figures/dataset_figures/intensity_spike_samples/`](../figures/dataset_figures/intensity_spike_samples/).
31
 
32
+ **Mode 2 — OptiTrack pose teleport.** GIFs (10 s playback at speed) for all 3 affected files: [`pose_teleport_samples/`](../figures/dataset_figures/pose_teleport_samples/).
 
 
 
 
 
33
 
34
  **Modes 1+2+3 — top-6 worst-offenders chart** (tactile intensity + sensor velocity time-series):
35
 
 
45
  "tau_velocity_mps": 5.0,
46
  "buffer_frames": 3,
47
  "summary": {
48
+ "n_episodes": 27,
49
+ "total_frames": 221621,
50
+ "total_bad_frames": 122,
51
+ "bad_fraction_overall": 0.00055,
52
+ "n_episodes_with_bad_frames": 8
53
  },
54
  "episodes": {
55
  "2026-05-11/episode_003": {
figures/dataset_figures/F8_per_episode_summary.csv CHANGED
@@ -1,31 +1,28 @@
1
- Date,Episode,Frames,Dur (s),Contact %,Max I_L,Max I_R,Drift L,Drift R
2
- 2026-03-23,episode_000,2141,72.1,51.3,1.30,12.40,1.30,1.35
3
- 2026-03-23,episode_001,7142,295.4,68.7,1.40,15.77,1.30,1.34
4
- 2026-03-23,episode_002,8855,372.2,64.8,1.54,18.89,1.53,1.43
5
- 2026-05-10,episode_000,6891,235.9,62.9,9.95,8.56,1.73,1.34
6
- 2026-05-10,episode_001,7621,258.2,59.6,9.15,10.84,1.44,1.33
7
- 2026-05-10,episode_002,10476,371.5,56.9,9.29,10.83,1.63,1.44
8
- 2026-05-10,episode_003,9262,311.7,71.1,7.36,5.11,1.37,2.30
9
- 2026-05-10,episode_004,2407,81.0,90.1,6.22,5.02,4.44,2.12
10
- 2026-05-10,episode_005,764,25.8,91.0,6.18,5.68,1.76,3.99
11
- 2026-05-10,episode_006,3926,136.1,89.7,6.13,5.76,1.32,1.31
12
- 2026-05-10,episode_007,1928,65.1,79.5,4.43,6.60,3.67,1.28
13
- 2026-05-10,episode_008,1305,44.0,71.6,4.36,5.17,1.31,3.35
14
- 2026-05-10,episode_009,1824,61.6,90.0,8.73,9.88,2.19,1.29
15
- 2026-05-10,episode_010,1035,35.7,74.0,7.34,5.85,1.37,5.44
16
- 2026-05-10,episode_011,2202,74.3,87.1,9.85,4.63,1.36,1.27
17
- 2026-05-11,episode_003,10032,338.2,55.4,82.94,23.81,1.84,1.73
18
- 2026-05-11,episode_004,2639,89.1,56.0,11.38,8.94,1.35,1.81
19
- 2026-05-11,episode_005,16927,571.1,73.3,17.03,11.07,1.64,1.52
20
- 2026-05-11,episode_006,11990,404.7,70.0,11.88,11.24,1.58,1.54
21
- 2026-05-11,episode_007,6909,238.5,68.8,9.08,62.96,1.63,1.56
22
- 2026-05-11,episode_008,13705,462.3,69.1,87.35,12.33,1.62,1.62
23
- 2026-05-11,episode_009,4366,147.5,73.0,12.53,10.27,1.44,1.43
24
- 2026-05-11,episode_010,6829,230.5,64.5,6.84,10.26,1.47,1.59
25
- 2026-05-11,episode_011,7300,261.2,57.7,41.30,9.54,1.36,1.43
26
- 2026-05-11,episode_012,24547,827.8,66.6,9.84,11.97,1.54,1.63
27
- 2026-05-11,episode_013,12724,434.5,33.5,3.96,4.79,1.37,1.47
28
- 2026-05-11,episode_014,2417,81.9,42.9,2.95,4.26,1.41,1.76
29
- 2026-05-11,episode_015,5960,202.3,38.7,4.01,6.45,1.37,1.36
30
- 2026-05-11,episode_016,11896,413.5,41.8,4.92,6.04,1.44,1.90
31
- 2026-05-11,episode_017,33739,1157.5,86.1,82.50,23.14,1.69,2.15
 
1
+ Date,Episode,Frames,Dur (s),Contact %,Max I_L,Max I_R,Drift L,Drift R
2
+ 2026-05-10,episode_000,6891,235.9,62.9,9.95,8.56,1.73,1.34
3
+ 2026-05-10,episode_001,7621,258.2,59.6,9.15,10.84,1.44,1.33
4
+ 2026-05-10,episode_002,10476,371.5,56.9,9.29,10.83,1.63,1.44
5
+ 2026-05-10,episode_003,9262,311.7,71.1,7.36,5.11,1.37,2.30
6
+ 2026-05-10,episode_004,2407,81.0,90.1,6.22,5.02,4.44,2.12
7
+ 2026-05-10,episode_005,764,25.8,91.0,6.18,5.68,1.76,3.99
8
+ 2026-05-10,episode_006,3926,136.1,89.7,6.13,5.76,1.32,1.31
9
+ 2026-05-10,episode_007,1928,65.1,79.5,4.43,6.60,3.67,1.28
10
+ 2026-05-10,episode_008,1305,44.0,71.6,4.36,5.17,1.31,3.35
11
+ 2026-05-10,episode_009,1824,61.6,90.0,8.73,9.88,2.19,1.29
12
+ 2026-05-10,episode_010,1035,35.7,74.0,7.34,5.85,1.37,5.44
13
+ 2026-05-10,episode_011,2202,74.3,87.1,9.85,4.63,1.36,1.27
14
+ 2026-05-11,episode_003,10032,338.2,55.4,82.94,23.81,1.84,1.73
15
+ 2026-05-11,episode_004,2639,89.1,56.0,11.38,8.94,1.35,1.81
16
+ 2026-05-11,episode_005,16927,571.1,73.3,17.03,11.07,1.64,1.52
17
+ 2026-05-11,episode_006,11990,404.7,70.0,11.88,11.24,1.58,1.54
18
+ 2026-05-11,episode_007,6909,238.5,68.8,9.08,62.96,1.63,1.56
19
+ 2026-05-11,episode_008,13705,462.3,69.1,87.35,12.33,1.62,1.62
20
+ 2026-05-11,episode_009,4366,147.5,73.0,12.53,10.27,1.44,1.43
21
+ 2026-05-11,episode_010,6829,230.5,64.5,6.84,10.26,1.47,1.59
22
+ 2026-05-11,episode_011,7300,261.2,57.7,41.30,9.54,1.36,1.43
23
+ 2026-05-11,episode_012,24547,827.8,66.6,9.84,11.97,1.54,1.63
24
+ 2026-05-11,episode_013,12724,434.5,33.5,3.96,4.79,1.37,1.47
25
+ 2026-05-11,episode_014,2417,81.9,42.9,2.95,4.26,1.41,1.76
26
+ 2026-05-11,episode_015,5960,202.3,38.7,4.01,6.45,1.37,1.36
27
+ 2026-05-11,episode_016,11896,413.5,41.8,4.92,6.04,1.44,1.90
28
+ 2026-05-11,episode_017,33739,1157.5,86.1,82.50,23.14,1.69,2.15
 
 
 
figures/dataset_figures/data_quality_breakdown.json CHANGED
@@ -1,10 +1,10 @@
1
  {
2
- "total_frames": 239759,
3
- "n_episodes": 30,
4
  "modes": {
5
  "gelsight_led_flicker": {
6
  "frames": 66,
7
- "pct": 0.027527642340850603,
8
  "episodes_affected": [
9
  "2026-05-11/episode_003",
10
  "2026-05-11/episode_007",
@@ -14,27 +14,24 @@
14
  ]
15
  },
16
  "pose_teleport": {
17
- "frames": 136,
18
- "pct": 0.05672362664175276,
19
  "episodes_affected": [
20
- "2026-03-23/episode_001",
21
- "2026-03-23/episode_002",
22
  "2026-05-10/episode_001",
23
  "2026-05-10/episode_002",
24
  "2026-05-11/episode_015"
25
  ]
26
  },
27
  "sensor_at_rest": {
28
- "frames": 37601,
29
- "pct": 15.6828315099746,
30
- "freeze_threshold_s": 30.0,
31
  "episodes_affected": [
32
- "2026-03-23/episode_001",
33
- "2026-03-23/episode_002",
34
  "2026-05-11/episode_005",
35
  "2026-05-11/episode_012",
36
  "2026-05-11/episode_017"
37
- ]
 
38
  }
39
  }
40
  }
 
1
  {
2
+ "total_frames": 221621,
3
+ "n_episodes": 27,
4
  "modes": {
5
  "gelsight_led_flicker": {
6
  "frames": 66,
7
+ "pct": 0.02978057133574887,
8
  "episodes_affected": [
9
  "2026-05-11/episode_003",
10
  "2026-05-11/episode_007",
 
14
  ]
15
  },
16
  "pose_teleport": {
17
+ "frames": 64,
18
+ "pct": 0.028878129780120113,
19
  "episodes_affected": [
 
 
20
  "2026-05-10/episode_001",
21
  "2026-05-10/episode_002",
22
  "2026-05-11/episode_015"
23
  ]
24
  },
25
  "sensor_at_rest": {
26
+ "frames": 31376,
27
+ "pct": 14.157503124703886,
28
+ "freeze_threshold_s": 5.0,
29
  "episodes_affected": [
 
 
30
  "2026-05-11/episode_005",
31
  "2026-05-11/episode_012",
32
  "2026-05-11/episode_017"
33
+ ],
34
+ "note": "Long held-pose intervals on active sensors. OptiTrack lost track and the recorder held the last pose. See docs/quality.md."
35
  }
36
  }
37
  }
figures/dataset_figures/data_quality_report.csv CHANGED
@@ -1,31 +1,28 @@
1
- name,frames,duration_s,iL_max,iL_p999,iR_max,iR_p999,freeze_L_s,freeze_R_s,drift_L,drift_R,L_max_vel,L_tele,L_nans,L_zeros,L_bad_quat,L_bbox_m,L_frozen_s,R_max_vel,R_tele,R_nans,R_zeros,R_bad_quat,R_bbox_m,R_frozen_s,severity
2
- 2026-03-23/episode_000,2141,72.09878873825073,1.2985996007919312,1.2982226610183716,12.399946212768555,11.215356826782227,0.06666666666666667,0.06666666666666667,1.2982226610183716,1.349454641342163,3.22448992729187,8,0,0,0,0.1160033643245697,26.033333333333335,1.2460277080535889,0,0,0,0,0.6654471158981323,0.4,WARN
3
- 2026-03-23/episode_001,7142,295.3777325153351,1.4029279947280884,1.3805484771728516,15.767017364501953,14.524497985839844,0.06666666666666667,0.5,1.3048450946807861,1.3419986963272095,10.00790786743164,39,0,0,0,0.36174994707107544,110.3,9.828115463256836,28,0,0,0,0.6532822847366333,0.0,SEVERE
4
- 2026-03-23/episode_002,8855,372.1804549694061,1.5370174646377563,1.530586838722229,18.894174575805664,17.257671356201172,0.06666666666666667,0.06666666666666667,1.5297207832336426,1.4312639236450195,8.620453834533691,25,0,0,0,0.3340376317501068,51.7,9.287128448486328,58,0,0,0,0.7628227472305298,0.7,SEVERE
5
- 2026-05-10/episode_000,6891,235.86456942558289,9.954959869384766,9.502667427062988,8.556878089904785,7.929441928863525,0.13333333333333333,0.13333333333333333,1.7265734672546387,1.3411930799484253,1.8689497709274292,0,0,0,0,0.7585326433181763,0.0,2.0365939140319824,1,0,0,0,0.6592375636100769,0.0,WARN
6
- 2026-05-10/episode_001,7621,258.1599111557007,9.14508056640625,8.607175827026367,10.844161033630371,8.153829574584961,0.13333333333333333,0.43333333333333335,1.4424065351486206,1.332179069519043,11.010350227355957,4,0,0,0,0.8438521027565002,0.8,2.739832878112793,1,0,0,0,0.8738973140716553,0.0,SEVERE
7
- 2026-05-10/episode_002,10476,371.4840340614319,9.285778045654297,8.520700454711914,10.832865715026855,10.622151374816895,0.13333333333333333,0.13333333333333333,1.6322435140609741,1.438342571258545,6.318398952484131,9,0,0,0,0.6564764976501465,0.0,6.197483062744141,13,0,0,0,0.7144262194633484,0.0,SEVERE
8
- 2026-05-10/episode_003,9262,311.72618341445923,7.358817100524902,7.206627368927002,5.112305641174316,4.872345447540283,0.13333333333333333,0.13333333333333333,1.3687100410461426,2.30434513092041,1.233119010925293,0,0,0,0,0.4686213731765747,0.0,1.6374657154083252,0,0,0,0,0.5558303594589233,0.13333333333333333,OK
9
- 2026-05-10/episode_004,2407,81.0497419834137,6.216858863830566,6.216858863830566,5.022139549255371,5.022139549255371,0.13333333333333333,0.13333333333333333,4.439228057861328,2.1246819496154785,0.513603150844574,0,0,0,0,0.18504935503005981,0.0,0.5007349252700806,0,0,0,0,0.24423444271087646,0.0,OK
10
- 2026-05-10/episode_005,764,25.767585039138794,6.17763614654541,6.17763614654541,5.684418678283691,5.684418678283691,0.13333333333333333,0.13333333333333333,1.7643053531646729,3.985346794128418,0.7517812252044678,0,0,0,0,0.28742340207099915,0.0,0.8874262571334839,0,0,0,0,0.1753755509853363,0.0,OK
11
- 2026-05-10/episode_006,3926,136.1196687221527,6.1306538581848145,6.070600509643555,5.757662773132324,5.637418746948242,0.13333333333333333,0.13333333333333333,1.3167400360107422,1.3133844137191772,0.9340700507164001,0,0,0,0,0.23434703052043915,0.0,1.2539730072021484,0,0,0,0,0.21440032124519348,0.0,OK
12
- 2026-05-10/episode_007,1928,65.14239430427551,4.431671142578125,4.431671142578125,6.599103927612305,6.599103927612305,0.1,0.13333333333333333,3.666048526763916,1.2817925214767456,0.7710904479026794,0,0,0,0,0.15755696594715118,0.0,0.528869092464447,0,0,0,0,0.19517114758491516,0.0,OK
13
- 2026-05-10/episode_008,1305,43.98420429229736,4.363437652587891,4.363437652587891,5.165870666503906,5.165870666503906,0.13333333333333333,0.1,1.3051360845565796,3.3461828231811523,0.6230773329734802,0,0,0,0,0.17741161584854126,0.0,0.5872477889060974,0,0,0,0,0.19946378469467163,0.0,OK
14
- 2026-05-10/episode_009,1824,61.569007873535156,8.726472854614258,8.726472854614258,9.881656646728516,9.881656646728516,0.13333333333333333,0.13333333333333333,2.193988800048828,1.2859500646591187,1.2424747943878174,0,0,0,0,0.4243673086166382,0.0,1.0909569263458252,0,0,0,0,0.527230978012085,0.0,OK
15
- 2026-05-10/episode_010,1035,35.73836064338684,7.338359832763672,7.338359832763672,5.845841407775879,5.845841407775879,0.13333333333333333,0.13333333333333333,1.367807149887085,5.444718360900879,0.7939414381980896,0,0,0,0,0.42937028408050537,0.0,1.0503939390182495,0,0,0,0,0.3026258051395416,0.0,OK
16
- 2026-05-10/episode_011,2202,74.28065299987793,9.853972434997559,9.772574424743652,4.625083923339844,4.589715003967285,0.1,0.13333333333333333,1.3586143255233765,1.274619460105896,0.5511309504508972,0,0,0,0,0.393202006816864,0.0,0.6120307445526123,0,0,0,0,0.42668619751930237,0.0,OK
17
- 2026-05-11/episode_003,10032,338.1943950653076,82.93767547607422,10.532184600830078,23.806249618530273,10.21254825592041,0.13333333333333333,0.13333333333333333,1.8386316299438477,1.727660059928894,1.8788384199142456,0,0,0,0,0.8178746700286865,0.06666666666666667,1.412979006767273,0,0,0,0,0.6480345129966736,0.0,SEVERE
18
- 2026-05-11/episode_004,2639,89.14375257492065,11.384664535522461,11.384664535522461,8.94444751739502,8.94444751739502,0.13333333333333333,0.13333333333333333,1.3487457036972046,1.8100842237472534,1.2397667169570923,0,0,0,0,0.6155400276184082,0.0,1.6692270040512085,0,0,0,0,0.6846281290054321,0.0,OK
19
- 2026-05-11/episode_005,16927,571.0894627571106,17.02862548828125,9.175427436828613,11.073382377624512,10.467161178588867,0.13333333333333333,0.13333333333333333,1.6357817649841309,1.521001935005188,1.6163232326507568,0,0,0,0,0.8027920722961426,80.93333333333334,2.561594247817993,2,0,0,0,0.7624298334121704,80.43333333333334,WARN
20
- 2026-05-11/episode_006,11990,404.6732089519501,11.875617980957031,9.629899978637695,11.244091033935547,10.21268367767334,0.13333333333333333,0.13333333333333333,1.581176996231079,1.5395671129226685,1.4471203088760376,0,0,0,0,0.6732650995254517,0.03333333333333333,1.1907745599746704,0,0,0,0,0.5994620323181152,0.0,OK
21
- 2026-05-11/episode_007,6909,238.52864241600037,9.078901290893555,8.939229965209961,62.9570198059082,9.775282859802246,0.13333333333333333,0.16666666666666666,1.6348042488098145,1.5581130981445312,2.244452714920044,1,0,0,0,0.6618192195892334,0.0,2.0695149898529053,1,0,0,0,0.6615707278251648,0.0,SEVERE
22
- 2026-05-11/episode_008,13705,462.2995111942291,87.34776306152344,10.651823043823242,12.331182479858398,10.890426635742188,0.16666666666666666,0.13333333333333333,1.6160932779312134,1.6228556632995605,1.2217721939086914,0,0,0,0,0.6075038313865662,0.0,1.1698782444000244,0,0,0,0,0.6313921809196472,0.0,SEVERE
23
- 2026-05-11/episode_009,4366,147.5375587940216,12.528318405151367,10.757293701171875,10.269855499267578,10.174468040466309,0.13333333333333333,0.13333333333333333,1.4370399713516235,1.433455467224121,1.58137845993042,0,0,0,0,0.4563417136669159,0.0,1.6967236995697021,0,0,0,0,0.5322908759117126,0.03333333333333333,OK
24
- 2026-05-11/episode_010,6829,230.4571714401245,6.838210105895996,6.482601165771484,10.263664245605469,10.12047290802002,0.13333333333333333,0.13333333333333333,1.4687319993972778,1.5896480083465576,0.8748554587364197,0,0,0,0,0.4127434492111206,0.0,1.208195686340332,0,0,0,0,0.4724993407726288,0.0,OK
25
- 2026-05-11/episode_011,7300,261.1779751777649,41.304351806640625,7.981741905212402,9.543685913085938,8.67159366607666,0.13333333333333333,0.13333333333333333,1.3645423650741577,1.4325497150421143,2.1780731678009033,1,0,0,0,0.3773902356624603,0.03333333333333333,4.200778961181641,2,0,0,0,0.35125088691711426,0.03333333333333333,WARN
26
- 2026-05-11/episode_012,24547,827.8351356983185,9.838443756103516,8.420958518981934,11.970463752746582,10.537294387817383,0.13333333333333333,0.13333333333333333,1.5424200296401978,1.629311203956604,1.3143892288208008,0,0,0,0,0.4577551484107971,323.93333333333334,1.4685096740722656,0,0,0,0,0.49013078212738037,323.9,WARN
27
- 2026-05-11/episode_013,12724,434.4760947227478,3.957453966140747,3.7450618743896484,4.789372444152832,4.407420635223389,0.16666666666666666,0.13333333333333333,1.3697080612182617,1.4694222211837769,4.977523326873779,4,0,0,0,0.7816130518913269,2.2666666666666666,3.7432634830474854,2,0,0,0,0.9271408319473267,0.43333333333333335,WARN
28
- 2026-05-11/episode_014,2417,81.90661072731018,2.9458799362182617,2.904794454574585,4.263512134552002,4.156301975250244,0.13333333333333333,0.13333333333333333,1.4061338901519775,1.7620596885681152,3.0673017501831055,1,0,0,0,0.7665240168571472,0.9666666666666667,2.029221534729004,1,0,0,0,0.796659529209137,0.43333333333333335,WARN
29
- 2026-05-11/episode_015,5960,202.26362705230713,4.010581016540527,3.8817710876464844,6.448170185089111,6.446977615356445,0.13333333333333333,0.13333333333333333,1.3719329833984375,1.35837984085083,3.437119245529175,6,0,0,0,0.7733251452445984,1.2,11.945237159729004,3,0,0,0,0.802914023399353,0.5666666666666667,SEVERE
30
- 2026-05-11/episode_016,11896,413.4506289958954,4.921731472015381,4.104878902435303,6.035311698913574,5.689062595367432,0.13333333333333333,0.13333333333333333,1.4356489181518555,1.904976725578308,3.8403403759002686,9,0,0,0,0.774321436882019,1.3666666666666667,1.4276446104049683,0,0,0,0,0.7977728843688965,0.16666666666666666,WARN
31
- 2026-05-11/episode_017,33739,1157.4502182006836,82.49847412109375,18.224760055541992,23.14059829711914,21.998035430908203,0.13333333333333333,0.13333333333333333,1.6930094957351685,2.150322437286377,1.0966863632202148,0,0,0,0,0.46664488315582275,640.8333333333334,1.7117093801498413,0,0,0,0,0.5299887657165527,640.9,SEVERE
 
1
+ name,frames,duration_s,iL_max,iL_p999,iR_max,iR_p999,freeze_L_s,freeze_R_s,drift_L,drift_R,L_max_vel,L_tele,L_nans,L_zeros,L_bad_quat,L_bbox_m,L_frozen_s,R_max_vel,R_tele,R_nans,R_zeros,R_bad_quat,R_bbox_m,R_frozen_s,severity
2
+ 2026-05-10/episode_000,6891,235.86456942558289,9.954959869384766,9.502667427062988,8.556878089904785,7.929441928863525,0.13333333333333333,0.13333333333333333,1.7265734672546387,1.3411930799484253,1.8689497709274292,0,0,0,0,0.7585326433181763,0.0,2.0365939140319824,1,0,0,0,0.6592375636100769,0.0,WARN
3
+ 2026-05-10/episode_001,7621,258.1599111557007,9.14508056640625,8.607175827026367,10.844161033630371,8.153829574584961,0.13333333333333333,0.43333333333333335,1.4424065351486206,1.332179069519043,11.010350227355957,4,0,0,0,0.8438521027565002,0.8,2.739832878112793,1,0,0,0,0.8738973140716553,0.0,SEVERE
4
+ 2026-05-10/episode_002,10476,371.4840340614319,9.285778045654297,8.520700454711914,10.832865715026855,10.622151374816895,0.13333333333333333,0.13333333333333333,1.6322435140609741,1.438342571258545,6.318398952484131,9,0,0,0,0.6564764976501465,0.0,6.197483062744141,13,0,0,0,0.7144262194633484,0.0,SEVERE
5
+ 2026-05-10/episode_003,9262,311.72618341445923,7.358817100524902,7.206627368927002,5.112305641174316,4.872345447540283,0.13333333333333333,0.13333333333333333,1.3687100410461426,2.30434513092041,1.233119010925293,0,0,0,0,0.4686213731765747,0.0,1.6374657154083252,0,0,0,0,0.5558303594589233,0.13333333333333333,OK
6
+ 2026-05-10/episode_004,2407,81.0497419834137,6.216858863830566,6.216858863830566,5.022139549255371,5.022139549255371,0.13333333333333333,0.13333333333333333,4.439228057861328,2.1246819496154785,0.513603150844574,0,0,0,0,0.18504935503005981,0.0,0.5007349252700806,0,0,0,0,0.24423444271087646,0.0,OK
7
+ 2026-05-10/episode_005,764,25.767585039138794,6.17763614654541,6.17763614654541,5.684418678283691,5.684418678283691,0.13333333333333333,0.13333333333333333,1.7643053531646729,3.985346794128418,0.7517812252044678,0,0,0,0,0.28742340207099915,0.0,0.8874262571334839,0,0,0,0,0.1753755509853363,0.0,OK
8
+ 2026-05-10/episode_006,3926,136.1196687221527,6.1306538581848145,6.070600509643555,5.757662773132324,5.637418746948242,0.13333333333333333,0.13333333333333333,1.3167400360107422,1.3133844137191772,0.9340700507164001,0,0,0,0,0.23434703052043915,0.0,1.2539730072021484,0,0,0,0,0.21440032124519348,0.0,OK
9
+ 2026-05-10/episode_007,1928,65.14239430427551,4.431671142578125,4.431671142578125,6.599103927612305,6.599103927612305,0.1,0.13333333333333333,3.666048526763916,1.2817925214767456,0.7710904479026794,0,0,0,0,0.15755696594715118,0.0,0.528869092464447,0,0,0,0,0.19517114758491516,0.0,OK
10
+ 2026-05-10/episode_008,1305,43.98420429229736,4.363437652587891,4.363437652587891,5.165870666503906,5.165870666503906,0.13333333333333333,0.1,1.3051360845565796,3.3461828231811523,0.6230773329734802,0,0,0,0,0.17741161584854126,0.0,0.5872477889060974,0,0,0,0,0.19946378469467163,0.0,OK
11
+ 2026-05-10/episode_009,1824,61.569007873535156,8.726472854614258,8.726472854614258,9.881656646728516,9.881656646728516,0.13333333333333333,0.13333333333333333,2.193988800048828,1.2859500646591187,1.2424747943878174,0,0,0,0,0.4243673086166382,0.0,1.0909569263458252,0,0,0,0,0.527230978012085,0.0,OK
12
+ 2026-05-10/episode_010,1035,35.73836064338684,7.338359832763672,7.338359832763672,5.845841407775879,5.845841407775879,0.13333333333333333,0.13333333333333333,1.367807149887085,5.444718360900879,0.7939414381980896,0,0,0,0,0.42937028408050537,0.0,1.0503939390182495,0,0,0,0,0.3026258051395416,0.0,OK
13
+ 2026-05-10/episode_011,2202,74.28065299987793,9.853972434997559,9.772574424743652,4.625083923339844,4.589715003967285,0.1,0.13333333333333333,1.3586143255233765,1.274619460105896,0.5511309504508972,0,0,0,0,0.393202006816864,0.0,0.6120307445526123,0,0,0,0,0.42668619751930237,0.0,OK
14
+ 2026-05-11/episode_003,10032,338.1943950653076,82.93767547607422,10.532184600830078,23.806249618530273,10.21254825592041,0.13333333333333333,0.13333333333333333,1.8386316299438477,1.727660059928894,1.8788384199142456,0,0,0,0,0.8178746700286865,0.06666666666666667,1.412979006767273,0,0,0,0,0.6480345129966736,0.0,SEVERE
15
+ 2026-05-11/episode_004,2639,89.14375257492065,11.384664535522461,11.384664535522461,8.94444751739502,8.94444751739502,0.13333333333333333,0.13333333333333333,1.3487457036972046,1.8100842237472534,1.2397667169570923,0,0,0,0,0.6155400276184082,0.0,1.6692270040512085,0,0,0,0,0.6846281290054321,0.0,OK
16
+ 2026-05-11/episode_005,16927,571.0894627571106,17.02862548828125,9.175427436828613,11.073382377624512,10.467161178588867,0.13333333333333333,0.13333333333333333,1.6357817649841309,1.521001935005188,1.6163232326507568,0,0,0,0,0.8027920722961426,80.93333333333334,2.561594247817993,2,0,0,0,0.7624298334121704,80.43333333333334,WARN
17
+ 2026-05-11/episode_006,11990,404.6732089519501,11.875617980957031,9.629899978637695,11.244091033935547,10.21268367767334,0.13333333333333333,0.13333333333333333,1.581176996231079,1.5395671129226685,1.4471203088760376,0,0,0,0,0.6732650995254517,0.03333333333333333,1.1907745599746704,0,0,0,0,0.5994620323181152,0.0,OK
18
+ 2026-05-11/episode_007,6909,238.52864241600037,9.078901290893555,8.939229965209961,62.9570198059082,9.775282859802246,0.13333333333333333,0.16666666666666666,1.6348042488098145,1.5581130981445312,2.244452714920044,1,0,0,0,0.6618192195892334,0.0,2.0695149898529053,1,0,0,0,0.6615707278251648,0.0,SEVERE
19
+ 2026-05-11/episode_008,13705,462.2995111942291,87.34776306152344,10.651823043823242,12.331182479858398,10.890426635742188,0.16666666666666666,0.13333333333333333,1.6160932779312134,1.6228556632995605,1.2217721939086914,0,0,0,0,0.6075038313865662,0.0,1.1698782444000244,0,0,0,0,0.6313921809196472,0.0,SEVERE
20
+ 2026-05-11/episode_009,4366,147.5375587940216,12.528318405151367,10.757293701171875,10.269855499267578,10.174468040466309,0.13333333333333333,0.13333333333333333,1.4370399713516235,1.433455467224121,1.58137845993042,0,0,0,0,0.4563417136669159,0.0,1.6967236995697021,0,0,0,0,0.5322908759117126,0.03333333333333333,OK
21
+ 2026-05-11/episode_010,6829,230.4571714401245,6.838210105895996,6.482601165771484,10.263664245605469,10.12047290802002,0.13333333333333333,0.13333333333333333,1.4687319993972778,1.5896480083465576,0.8748554587364197,0,0,0,0,0.4127434492111206,0.0,1.208195686340332,0,0,0,0,0.4724993407726288,0.0,OK
22
+ 2026-05-11/episode_011,7300,261.1779751777649,41.304351806640625,7.981741905212402,9.543685913085938,8.67159366607666,0.13333333333333333,0.13333333333333333,1.3645423650741577,1.4325497150421143,2.1780731678009033,1,0,0,0,0.3773902356624603,0.03333333333333333,4.200778961181641,2,0,0,0,0.35125088691711426,0.03333333333333333,WARN
23
+ 2026-05-11/episode_012,24547,827.8351356983185,9.838443756103516,8.420958518981934,11.970463752746582,10.537294387817383,0.13333333333333333,0.13333333333333333,1.5424200296401978,1.629311203956604,1.3143892288208008,0,0,0,0,0.4577551484107971,323.93333333333334,1.4685096740722656,0,0,0,0,0.49013078212738037,323.9,WARN
24
+ 2026-05-11/episode_013,12724,434.4760947227478,3.957453966140747,3.7450618743896484,4.789372444152832,4.407420635223389,0.16666666666666666,0.13333333333333333,1.3697080612182617,1.4694222211837769,4.977523326873779,4,0,0,0,0.7816130518913269,2.2666666666666666,3.7432634830474854,2,0,0,0,0.9271408319473267,0.43333333333333335,WARN
25
+ 2026-05-11/episode_014,2417,81.90661072731018,2.9458799362182617,2.904794454574585,4.263512134552002,4.156301975250244,0.13333333333333333,0.13333333333333333,1.4061338901519775,1.7620596885681152,3.0673017501831055,1,0,0,0,0.7665240168571472,0.9666666666666667,2.029221534729004,1,0,0,0,0.796659529209137,0.43333333333333335,WARN
26
+ 2026-05-11/episode_015,5960,202.26362705230713,4.010581016540527,3.8817710876464844,6.448170185089111,6.446977615356445,0.13333333333333333,0.13333333333333333,1.3719329833984375,1.35837984085083,3.437119245529175,6,0,0,0,0.7733251452445984,1.2,11.945237159729004,3,0,0,0,0.802914023399353,0.5666666666666667,SEVERE
27
+ 2026-05-11/episode_016,11896,413.4506289958954,4.921731472015381,4.104878902435303,6.035311698913574,5.689062595367432,0.13333333333333333,0.13333333333333333,1.4356489181518555,1.904976725578308,3.8403403759002686,9,0,0,0,0.774321436882019,1.3666666666666667,1.4276446104049683,0,0,0,0,0.7977728843688965,0.16666666666666666,WARN
28
+ 2026-05-11/episode_017,33739,1157.4502182006836,82.49847412109375,18.224760055541992,23.14059829711914,21.998035430908203,0.13333333333333333,0.13333333333333333,1.6930094957351685,2.150322437286377,1.0966863632202148,0,0,0,0,0.46664488315582275,640.8333333333334,1.7117093801498413,0,0,0,0,0.5299887657165527,640.9,SEVERE
 
 
 
figures/dataset_figures/pose_teleport_samples/2026-03-23/episode_001_teleport.png DELETED

Git LFS Details

  • SHA256: a1809f8f57bb980594df425b03df3794ae075a5635553f27553662a92ca62a32
  • Pointer size: 131 Bytes
  • Size of remote file: 489 kB
figures/dataset_figures/pose_teleport_samples/2026-03-23/episode_002_teleport.png DELETED

Git LFS Details

  • SHA256: d414a71b58b1c3bcab267666237c05ec026bb0447d28d52e39151553d3f84d87
  • Pointer size: 131 Bytes
  • Size of remote file: 481 kB
figures/episode_previews/motherboard/2026-03-23/episode_000.gif DELETED

Git LFS Details

  • SHA256: 7e445f0c8b138a58bb1bbb2f62fa90f2abada187425c27e39299fa7120ecd63d
  • Pointer size: 131 Bytes
  • Size of remote file: 871 kB
figures/episode_previews/motherboard/2026-03-23/episode_001.gif DELETED

Git LFS Details

  • SHA256: 409b6788475f1dd921696a2e04839ccb0980ad97ab2f2e353d60a0bc44657558
  • Pointer size: 131 Bytes
  • Size of remote file: 861 kB
figures/episode_previews/motherboard/2026-03-23/episode_002.gif DELETED

Git LFS Details

  • SHA256: dbe421aa9f708a045c82c70b00a36a9fd0ee002ff4894c19ac85d0d5f21f7920
  • Pointer size: 131 Bytes
  • Size of remote file: 897 kB
processed/mode1_v1/motherboard/2026-03-23/episode_000.contact.json DELETED
@@ -1,182 +0,0 @@
1
- {
2
- "tau": 8.0,
3
- "drift_left": 1.2982226610183716,
4
- "drift_right": 1.349454641342163,
5
- "drift_warning": false,
6
- "ref_p01_idx_left": 1161,
7
- "ref_p01_idx_right": 1072,
8
- "tactile_left_intensity": {
9
- "min": 0.0,
10
- "max": 1.2985996007919312,
11
- "mean": 1.267616629600525,
12
- "median": 1.2670005559921265,
13
- "fraction_above_tau": 0.0,
14
- "histogram": [
15
- 2,
16
- 0,
17
- 0,
18
- 0,
19
- 0,
20
- 0,
21
- 0,
22
- 0,
23
- 0,
24
- 0,
25
- 0,
26
- 0,
27
- 0,
28
- 0,
29
- 0,
30
- 0,
31
- 0,
32
- 0,
33
- 0,
34
- 2139
35
- ]
36
- },
37
- "tactile_left_area": {
38
- "min": 0.0,
39
- "max": 0.0,
40
- "mean": 0.0,
41
- "median": 0.0,
42
- "fraction_above_tau": 0.0,
43
- "histogram": [
44
- 0,
45
- 0,
46
- 0,
47
- 0,
48
- 0,
49
- 0,
50
- 0,
51
- 0,
52
- 0,
53
- 0,
54
- 2141,
55
- 0,
56
- 0,
57
- 0,
58
- 0,
59
- 0,
60
- 0,
61
- 0,
62
- 0,
63
- 0
64
- ]
65
- },
66
- "tactile_left_mixed": {
67
- "min": 0.0,
68
- "max": 0.0,
69
- "mean": 0.0,
70
- "median": 0.0,
71
- "fraction_above_tau": 0.0,
72
- "histogram": [
73
- 0,
74
- 0,
75
- 0,
76
- 0,
77
- 0,
78
- 0,
79
- 0,
80
- 0,
81
- 0,
82
- 0,
83
- 2141,
84
- 0,
85
- 0,
86
- 0,
87
- 0,
88
- 0,
89
- 0,
90
- 0,
91
- 0,
92
- 0
93
- ]
94
- },
95
- "tactile_right_intensity": {
96
- "min": 0.0,
97
- "max": 12.399946212768555,
98
- "mean": 3.7102715969085693,
99
- "median": 2.620025157928467,
100
- "fraction_above_tau": 0.0639887902849136,
101
- "histogram": [
102
- 2,
103
- 14,
104
- 999,
105
- 35,
106
- 57,
107
- 74,
108
- 76,
109
- 129,
110
- 111,
111
- 120,
112
- 151,
113
- 119,
114
- 123,
115
- 70,
116
- 39,
117
- 8,
118
- 9,
119
- 0,
120
- 3,
121
- 2
122
- ]
123
- },
124
- "tactile_right_area": {
125
- "min": 0.0,
126
- "max": 0.4599609375,
127
- "mean": 0.1108097955584526,
128
- "median": 0.0518798828125,
129
- "fraction_above_tau": 0.0,
130
- "histogram": [
131
- 1031,
132
- 28,
133
- 44,
134
- 47,
135
- 45,
136
- 72,
137
- 125,
138
- 104,
139
- 101,
140
- 109,
141
- 107,
142
- 38,
143
- 63,
144
- 55,
145
- 45,
146
- 32,
147
- 45,
148
- 30,
149
- 13,
150
- 7
151
- ]
152
- },
153
- "tactile_right_mixed": {
154
- "min": 0.0,
155
- "max": 10.599336624145508,
156
- "mean": 1.9674044847488403,
157
- "median": 0.5857799649238586,
158
- "fraction_above_tau": 0.007473143390938813,
159
- "histogram": [
160
- 1055,
161
- 76,
162
- 73,
163
- 74,
164
- 86,
165
- 93,
166
- 117,
167
- 112,
168
- 103,
169
- 96,
170
- 97,
171
- 57,
172
- 50,
173
- 28,
174
- 8,
175
- 6,
176
- 5,
177
- 3,
178
- 0,
179
- 2
180
- ]
181
- }
182
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
processed/mode1_v1/motherboard/2026-03-23/episode_000.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:af64fc6cb7eda00e127e348900ff040741977f16d6ca03112c1a2e816a201a52
3
- size 316093899
 
 
 
 
processed/mode1_v1/motherboard/2026-03-23/episode_001.contact.json DELETED
@@ -1,182 +0,0 @@
1
- {
2
- "tau": 8.0,
3
- "drift_left": 1.3048450946807861,
4
- "drift_right": 1.3419986963272095,
5
- "drift_warning": false,
6
- "ref_p01_idx_left": 2666,
7
- "ref_p01_idx_right": 6408,
8
- "tactile_left_intensity": {
9
- "min": 0.0,
10
- "max": 1.4029279947280884,
11
- "mean": 1.2842105627059937,
12
- "median": 1.276340126991272,
13
- "fraction_above_tau": 0.0,
14
- "histogram": [
15
- 2,
16
- 0,
17
- 0,
18
- 0,
19
- 0,
20
- 0,
21
- 0,
22
- 0,
23
- 0,
24
- 0,
25
- 0,
26
- 0,
27
- 0,
28
- 0,
29
- 0,
30
- 0,
31
- 0,
32
- 868,
33
- 5839,
34
- 433
35
- ]
36
- },
37
- "tactile_left_area": {
38
- "min": 0.0,
39
- "max": 0.0,
40
- "mean": 0.0,
41
- "median": 0.0,
42
- "fraction_above_tau": 0.0,
43
- "histogram": [
44
- 0,
45
- 0,
46
- 0,
47
- 0,
48
- 0,
49
- 0,
50
- 0,
51
- 0,
52
- 0,
53
- 0,
54
- 7142,
55
- 0,
56
- 0,
57
- 0,
58
- 0,
59
- 0,
60
- 0,
61
- 0,
62
- 0,
63
- 0
64
- ]
65
- },
66
- "tactile_left_mixed": {
67
- "min": 0.0,
68
- "max": 0.0,
69
- "mean": 0.0,
70
- "median": 0.0,
71
- "fraction_above_tau": 0.0,
72
- "histogram": [
73
- 0,
74
- 0,
75
- 0,
76
- 0,
77
- 0,
78
- 0,
79
- 0,
80
- 0,
81
- 0,
82
- 0,
83
- 7142,
84
- 0,
85
- 0,
86
- 0,
87
- 0,
88
- 0,
89
- 0,
90
- 0,
91
- 0,
92
- 0
93
- ]
94
- },
95
- "tactile_right_intensity": {
96
- "min": 0.0,
97
- "max": 15.767017364501953,
98
- "mean": 4.869727611541748,
99
- "median": 4.737385272979736,
100
- "fraction_above_tau": 0.178241388966676,
101
- "histogram": [
102
- 3,
103
- 2038,
104
- 213,
105
- 269,
106
- 431,
107
- 612,
108
- 642,
109
- 557,
110
- 526,
111
- 508,
112
- 457,
113
- 304,
114
- 228,
115
- 120,
116
- 138,
117
- 69,
118
- 10,
119
- 6,
120
- 4,
121
- 7
122
- ]
123
- },
124
- "tactile_right_area": {
125
- "min": 0.0,
126
- "max": 0.54974365234375,
127
- "mean": 0.1418549120426178,
128
- "median": 0.153076171875,
129
- "fraction_above_tau": 0.0,
130
- "histogram": [
131
- 2208,
132
- 161,
133
- 224,
134
- 321,
135
- 383,
136
- 503,
137
- 603,
138
- 641,
139
- 628,
140
- 534,
141
- 385,
142
- 166,
143
- 175,
144
- 105,
145
- 73,
146
- 12,
147
- 6,
148
- 5,
149
- 5,
150
- 4
151
- ]
152
- },
153
- "tactile_right_mixed": {
154
- "min": 0.0,
155
- "max": 14.083170890808105,
156
- "mean": 3.102226972579956,
157
- "median": 2.7012224197387695,
158
- "fraction_above_tau": 0.06412769532343882,
159
- "histogram": [
160
- 2337,
161
- 402,
162
- 459,
163
- 492,
164
- 520,
165
- 439,
166
- 484,
167
- 474,
168
- 369,
169
- 358,
170
- 248,
171
- 205,
172
- 122,
173
- 118,
174
- 73,
175
- 25,
176
- 6,
177
- 2,
178
- 2,
179
- 7
180
- ]
181
- }
182
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
processed/mode1_v1/motherboard/2026-03-23/episode_001.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:d8de0b4107d5f6b1f0957be5d9793d9759c3344d78459f9875ba79dcf1848355
3
- size 1053961611
 
 
 
 
processed/mode1_v1/motherboard/2026-03-23/episode_002.contact.json DELETED
@@ -1,182 +0,0 @@
1
- {
2
- "tau": 8.0,
3
- "drift_left": 1.5297207832336426,
4
- "drift_right": 1.4312639236450195,
5
- "drift_warning": false,
6
- "ref_p01_idx_left": 4597,
7
- "ref_p01_idx_right": 3967,
8
- "tactile_left_intensity": {
9
- "min": 0.0,
10
- "max": 1.5370174646377563,
11
- "mean": 1.3338154554367065,
12
- "median": 1.3196828365325928,
13
- "fraction_above_tau": 0.0,
14
- "histogram": [
15
- 2,
16
- 0,
17
- 0,
18
- 0,
19
- 0,
20
- 0,
21
- 0,
22
- 0,
23
- 0,
24
- 0,
25
- 0,
26
- 0,
27
- 0,
28
- 0,
29
- 0,
30
- 0,
31
- 3747,
32
- 3425,
33
- 982,
34
- 699
35
- ]
36
- },
37
- "tactile_left_area": {
38
- "min": 0.0,
39
- "max": 0.0,
40
- "mean": 0.0,
41
- "median": 0.0,
42
- "fraction_above_tau": 0.0,
43
- "histogram": [
44
- 0,
45
- 0,
46
- 0,
47
- 0,
48
- 0,
49
- 0,
50
- 0,
51
- 0,
52
- 0,
53
- 0,
54
- 8855,
55
- 0,
56
- 0,
57
- 0,
58
- 0,
59
- 0,
60
- 0,
61
- 0,
62
- 0,
63
- 0
64
- ]
65
- },
66
- "tactile_left_mixed": {
67
- "min": 0.0,
68
- "max": 0.0,
69
- "mean": 0.0,
70
- "median": 0.0,
71
- "fraction_above_tau": 0.0,
72
- "histogram": [
73
- 0,
74
- 0,
75
- 0,
76
- 0,
77
- 0,
78
- 0,
79
- 0,
80
- 0,
81
- 0,
82
- 0,
83
- 8855,
84
- 0,
85
- 0,
86
- 0,
87
- 0,
88
- 0,
89
- 0,
90
- 0,
91
- 0,
92
- 0
93
- ]
94
- },
95
- "tactile_right_intensity": {
96
- "min": 0.0,
97
- "max": 18.894174575805664,
98
- "mean": 4.354608535766602,
99
- "median": 4.317556858062744,
100
- "fraction_above_tau": 0.09994353472614342,
101
- "histogram": [
102
- 2,
103
- 3029,
104
- 366,
105
- 535,
106
- 1013,
107
- 1263,
108
- 903,
109
- 615,
110
- 436,
111
- 248,
112
- 138,
113
- 97,
114
- 69,
115
- 40,
116
- 36,
117
- 24,
118
- 20,
119
- 10,
120
- 6,
121
- 5
122
- ]
123
- },
124
- "tactile_right_area": {
125
- "min": 0.0,
126
- "max": 0.6431884765625,
127
- "mean": 0.1313725858926773,
128
- "median": 0.13128662109375,
129
- "fraction_above_tau": 0.0,
130
- "histogram": [
131
- 3121,
132
- 290,
133
- 384,
134
- 590,
135
- 760,
136
- 866,
137
- 866,
138
- 765,
139
- 475,
140
- 282,
141
- 116,
142
- 84,
143
- 59,
144
- 30,
145
- 47,
146
- 31,
147
- 28,
148
- 28,
149
- 11,
150
- 22
151
- ]
152
- },
153
- "tactile_right_mixed": {
154
- "min": 0.0,
155
- "max": 17.843576431274414,
156
- "mean": 2.5875444412231445,
157
- "median": 2.1701927185058594,
158
- "fraction_above_tau": 0.042687747035573126,
159
- "histogram": [
160
- 3351,
161
- 660,
162
- 1038,
163
- 1050,
164
- 889,
165
- 586,
166
- 436,
167
- 307,
168
- 170,
169
- 109,
170
- 80,
171
- 44,
172
- 32,
173
- 38,
174
- 22,
175
- 20,
176
- 12,
177
- 4,
178
- 2,
179
- 5
180
- ]
181
- }
182
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
processed/mode1_v1/motherboard/2026-03-23/episode_002.pt DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:76de4045d3c76bdeed5cf7590e50a1ef8abd02ca6aeae5a211f0187b6f60fd33
3
- size 1306704523
 
 
 
 
tasks.json CHANGED
@@ -5,28 +5,32 @@
5
  "purpose": "Dense, contact-rich, synchronized multimodal interaction data for tactile-visual dynamics / world-model learning. Not a policy-learning / demonstration dataset.",
6
  "operator": "human hands (handheld GelSight Mini sensors with motion-capture rigid bodies; no robot arm involved)",
7
  "embodiment": "human",
8
- "dates": ["2026-03-23", "2026-05-10", "2026-05-11"],
 
 
 
9
  "per_date_notes": {
10
- "2026-03-23": {
11
- "kind": "pilot",
12
- "active_sensors": ["right"],
13
- "note": "Early trials. ONLY the right GelSight + right rigid body were used. The left tactile stream is essentially flat (peak intensity ≤ 1.4 vs ≥ 12 when active) and the left OptiTrack pose is not informative — treat as right-sensor-only data."
14
- },
15
  "2026-05-10": {
16
  "kind": "session",
17
- "active_sensors": ["left", "right"],
 
 
 
18
  "note": "First full bimanual recording session."
19
  },
20
  "2026-05-11": {
21
  "kind": "session",
22
- "active_sensors": ["left", "right"],
23
- "note": "Largest session. A few GelSight LED-flicker frames and one mocap teleport event — see bad_frames.json."
 
 
 
24
  }
25
  },
26
- "n_episode_files": 30,
27
  "notes": "2026-05-11/episode_001 was lost at recording time (HDF5 superblock not finalized). 2026-05-11/episodes 000 and 002 were short test recordings with no tactile contact and were dropped."
28
  }
29
  },
30
  "schema_version": "mode1_v1",
31
  "layout": "processed/mode1_v1/<task>/<date>/episode_*.{pt,contact.json}"
32
- }
 
5
  "purpose": "Dense, contact-rich, synchronized multimodal interaction data for tactile-visual dynamics / world-model learning. Not a policy-learning / demonstration dataset.",
6
  "operator": "human hands (handheld GelSight Mini sensors with motion-capture rigid bodies; no robot arm involved)",
7
  "embodiment": "human",
8
+ "dates": [
9
+ "2026-05-10",
10
+ "2026-05-11"
11
+ ],
12
  "per_date_notes": {
 
 
 
 
 
13
  "2026-05-10": {
14
  "kind": "session",
15
+ "active_sensors": [
16
+ "left",
17
+ "right"
18
+ ],
19
  "note": "First full bimanual recording session."
20
  },
21
  "2026-05-11": {
22
  "kind": "session",
23
+ "active_sensors": [
24
+ "left",
25
+ "right"
26
+ ],
27
+ "note": "Largest session. A few GelSight LED-flicker frames and one mocap teleport event \u2014 see bad_frames.json."
28
  }
29
  },
30
+ "n_episode_files": 27,
31
  "notes": "2026-05-11/episode_001 was lost at recording time (HDF5 superblock not finalized). 2026-05-11/episodes 000 and 002 were short test recordings with no tactile contact and were dropped."
32
  }
33
  },
34
  "schema_version": "mode1_v1",
35
  "layout": "processed/mode1_v1/<task>/<date>/episode_*.{pt,contact.json}"
36
+ }