Upload 510 files
Browse filesAdd the cleaned D2 dataset under quest_d2_pilot_merged_cleaned_2/.
Includes episode metadata, trajectories, and both camera videos.
Please review before merging.
This view is limited to 50 files because it contains too many changes. See raw diff
- quest_d2_pilot_merged_cleaned_2/README.md +267 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000000.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000001.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000002.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000003.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000004.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000005.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000006.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000007.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000008.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000009.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000010.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000011.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000012.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000013.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000014.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000015.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000016.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000018.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000019.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000020.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000021.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000022.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000023.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000024.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000025.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000026.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000027.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000028.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000029.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000030.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000031.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000032.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000033.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000034.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000035.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000036.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000037.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000038.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000039.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000040.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000041.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000042.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000043.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000044.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000045.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000046.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000047.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000048.parquet +3 -0
- quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000049.parquet +3 -0
quest_d2_pilot_merged_cleaned_2/README.md
ADDED
|
@@ -0,0 +1,267 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
task_categories:
|
| 3 |
+
- robotics
|
| 4 |
+
tags:
|
| 5 |
+
- LeRobot
|
| 6 |
+
- franka
|
| 7 |
+
- isaac-lab
|
| 8 |
+
- teleoperation
|
| 9 |
+
- language-grounding
|
| 10 |
+
- compositional-generalization
|
| 11 |
+
size_categories:
|
| 12 |
+
- n<1K
|
| 13 |
+
pretty_name: Quest D2 Pilot (compositional pairing teleoperation demonstrations)
|
| 14 |
+
configs:
|
| 15 |
+
- config_name: default
|
| 16 |
+
data_files: data/*/*.parquet
|
| 17 |
+
---
|
| 18 |
+
|
| 19 |
+
# Quest D2 Pilot
|
| 20 |
+
|
| 21 |
+
Human teleoperation demonstrations for the **D2 compositional-pairing** axis
|
| 22 |
+
of the GroundProbe diagnostic benchmark. A simulated Franka Panda picks a
|
| 23 |
+
named source object and places it in or on a named container; every episode
|
| 24 |
+
in this release ended in task success.
|
| 25 |
+
|
| 26 |
+
- **168 episodes, 17,265 frames** (about 9.2 minutes at the nominal 31.25 Hz)
|
| 27 |
+
- **12 language instructions**, all at the `cluttered` complexity level
|
| 28 |
+
- **LeRobot v2.0** layout: one parquet per episode, one mp4 per camera per episode
|
| 29 |
+
- Two 640×640 RGB cameras: third-person and wrist (336 videos)
|
| 30 |
+
- Three parallel action encodings per step
|
| 31 |
+
- About 59 MB in total
|
| 32 |
+
|
| 33 |
+
Every number on this card was measured from the files in this repository
|
| 34 |
+
(see [Validation](#validation)), not copied from `meta/info.json`.
|
| 35 |
+
|
| 36 |
+
## D2 setup
|
| 37 |
+
|
| 38 |
+
D2 tests whether a policy composes a familiar pick with a familiar place when
|
| 39 |
+
the specific (source, container) pair is new.
|
| 40 |
+
|
| 41 |
+
- **Robot:** Franka Panda (`robot_type: franka_panda`), driven by a
|
| 42 |
+
differential-IK controller.
|
| 43 |
+
- **Simulator:** NVIDIA Isaac Lab-Arena, scene `d2_pairing_env`.
|
| 44 |
+
- **Operator interface:** Meta Quest 2 controller teleoperation.
|
| 45 |
+
- **Source objects (5):** cup, apple, orange, red block, Rubik's cube
|
| 46 |
+
(`obj_src_cup`, `obj_src_apple`, `obj_src_orange`, `obj_src_block`,
|
| 47 |
+
`obj_src_rubik`).
|
| 48 |
+
- **Containers (3):** plate, bowl, tray (`obj_tgt_plate`, `obj_tgt_bowl`,
|
| 49 |
+
`obj_tgt_tray`).
|
| 50 |
+
- **Instruction template:** `Place the <source> on|in the <container>.`
|
| 51 |
+
- **Layouts:** object and container positions are re-sampled per episode from
|
| 52 |
+
`groundlab.layout_seed` (master seed 42). All 168 episodes have distinct
|
| 53 |
+
layout seeds, and each (instruction, `layout_idx`) cell appears once;
|
| 54 |
+
`layout_idx` ranges over 0–19.
|
| 55 |
+
|
| 56 |
+
The 5 × 3 grid has 15 cells. This release contains demonstrations for 12 of
|
| 57 |
+
them; `cup→tray`, `apple→plate` and `orange→bowl` have no episodes.
|
| 58 |
+
|
| 59 |
+
## Per-task coverage
|
| 60 |
+
|
| 61 |
+
| `task_index` | `instruction_id` | instruction | episodes | frames | length range (frames) |
|
| 62 |
+
|---:|---|---|---:|---:|---|
|
| 63 |
+
| 0 | `T_cup_plate` | Place the cup on the plate. | 20 | 1,987 | 77–138 |
|
| 64 |
+
| 1 | `T_cup_bowl` | Place the cup in the bowl. | 18 | 1,760 | 74–130 |
|
| 65 |
+
| 2 | `T_apple_bowl` | Place the apple in the bowl. | 18 | 2,020 | 76–150 |
|
| 66 |
+
| 3 | `T_apple_tray` | Place the apple on the tray. | 19 | 2,065 | 75–141 |
|
| 67 |
+
| 4 | `T_orange_plate` | Place the orange on the plate. | 18 | 1,711 | 77–113 |
|
| 68 |
+
| 5 | `T_orange_tray` | Place the orange on the tray. | 16 | 1,539 | 64–123 |
|
| 69 |
+
| 6 | `T_block_plate` | Place the red block on the plate. | 17 | 1,770 | 76–130 |
|
| 70 |
+
| 7 | `T_block_bowl` | Place the red block in the bowl. | 17 | 1,840 | 90–152 |
|
| 71 |
+
| 8 | `T_block_tray` | Place the red block on the tray. | 1 | 114 | 114 |
|
| 72 |
+
| 9 | `T_rubik_bowl` | Place the rubik's cube in the bowl. | 12 | 1,248 | 75–125 |
|
| 73 |
+
| 10 | `T_rubik_tray` | Place the rubik's cube on the tray. | 11 | 1,129 | 79–128 |
|
| 74 |
+
| 11 | `T_rubik_plate` | Place the rubik's cube on the plate. | 1 | 82 | 82 |
|
| 75 |
+
| | | **total** | **168** | **17,265** | 64–152 |
|
| 76 |
+
|
| 77 |
+
Episode length: min 64, median 102.5, mean 102.8, max 152 frames.
|
| 78 |
+
|
| 79 |
+
Episodes per source object: cup 38, apple 37, red block 35, orange 34,
|
| 80 |
+
Rubik's cube 24. Episodes per container, for the 146 episodes that record
|
| 81 |
+
it: bowl 57, plate 46, tray 43.
|
| 82 |
+
|
| 83 |
+
## Data format
|
| 84 |
+
|
| 85 |
+
```
|
| 86 |
+
meta/info.json dataset-level schema, fps, control timing
|
| 87 |
+
meta/episodes.jsonl one line per episode (168)
|
| 88 |
+
meta/tasks.jsonl task_index → instruction (12)
|
| 89 |
+
meta/bad_seeds.jsonl layout seeds rejected during collection (161)
|
| 90 |
+
data/chunk-000/episode_{index:06d}.parquet 168 files
|
| 91 |
+
videos/chunk-000/observation.images.third_camera/episode_{index:06d}.mp4 168 files
|
| 92 |
+
videos/chunk-000/observation.images.wrist_camera/episode_{index:06d}.mp4 168 files
|
| 93 |
+
prune_manifest.json provenance of the final pruning step
|
| 94 |
+
provenance/ manifests from the earlier merge and prune steps
|
| 95 |
+
```
|
| 96 |
+
|
| 97 |
+
### Per-frame features (parquet)
|
| 98 |
+
|
| 99 |
+
| key | dtype | shape | meaning |
|
| 100 |
+
|---|---|---|---|
|
| 101 |
+
| `observation.state` | float32 | 9 | 7 arm joint positions + 2 finger positions |
|
| 102 |
+
| `observation.tcp_pose` | float32 | 7 | end-effector pose, `[x, y, z, qw, qx, qy, qz]` |
|
| 103 |
+
| `action` | float32 | 7 | `[dx, dy, dz, droll, dpitch, dyaw, gripper]` |
|
| 104 |
+
| `action.tcp_disp` | float32 | 7 | same layout as `action` |
|
| 105 |
+
| `action.joint` | float32 | 8 | 7 arm joint targets + gripper |
|
| 106 |
+
| `episode_index`, `frame_index`, `task_index` | int64 | 1 | indices |
|
| 107 |
+
| `timestamp` | float32 | 1 | seconds from episode start |
|
| 108 |
+
| `next.done`, `next.success` | bool | 1 | termination / success flags |
|
| 109 |
+
|
| 110 |
+
Videos are H.264, yuv420p, 640×640, with exactly one video frame per parquet
|
| 111 |
+
row.
|
| 112 |
+
|
| 113 |
+
The parent project's dataset documentation describes the three action
|
| 114 |
+
encodings as follows; this was not re-derived for this release:
|
| 115 |
+
|
| 116 |
+
- `action` is the IK target the operator commanded minus the measured TCP
|
| 117 |
+
pose, i.e. an aim point rather than a motion.
|
| 118 |
+
- `action.tcp_disp` is the realised displacement `tcp_pose[k+1] − tcp_pose[k]`.
|
| 119 |
+
- `action.joint` is the absolute joint-position target issued by the IK term.
|
| 120 |
+
|
| 121 |
+
Rotations are documented there as XYZ-extrinsic Euler angles in radians, and
|
| 122 |
+
the gripper command as ±1.
|
| 123 |
+
|
| 124 |
+
### Per-episode metadata (`meta/episodes.jsonl`)
|
| 125 |
+
|
| 126 |
+
`episode_index`, `tasks`, `length`, plus `groundlab.*` fields:
|
| 127 |
+
`instruction_id`, `task_index`, `alt_task_index`, `split`, `complexity`,
|
| 128 |
+
`layout_idx`, `layout_seed`, `master_random_seed`, `attempt`, `salt`,
|
| 129 |
+
`success`, `target_asset`, and `tray_asset` (the container; present on 146
|
| 130 |
+
of 168 episodes).
|
| 131 |
+
|
| 132 |
+
In this release every episode has `split = train`,
|
| 133 |
+
`complexity = cluttered`, `success = true`, `master_random_seed = 42`.
|
| 134 |
+
|
| 135 |
+
### Control rate
|
| 136 |
+
|
| 137 |
+
`meta/info.json` records `fps: 31` and a `control` block giving the nominal
|
| 138 |
+
recording rate as 31.25 Hz (`sim_dt` 0.004 s × `decimation` 4 × 2 env steps
|
| 139 |
+
per recorded frame). See the limitation on episodes 0–16 below.
|
| 140 |
+
|
| 141 |
+
## Download
|
| 142 |
+
|
| 143 |
+
The repository is private, so authenticate first.
|
| 144 |
+
|
| 145 |
+
```bash
|
| 146 |
+
pip install -U huggingface_hub
|
| 147 |
+
hf auth login
|
| 148 |
+
hf download idkWhatToUse/quest-d2-pilot --repo-type dataset --local-dir quest-d2-pilot
|
| 149 |
+
```
|
| 150 |
+
|
| 151 |
+
Metadata and trajectories only, without videos:
|
| 152 |
+
|
| 153 |
+
```bash
|
| 154 |
+
hf download idkWhatToUse/quest-d2-pilot --repo-type dataset \
|
| 155 |
+
--include "meta/*" "data/*" --local-dir quest-d2-pilot
|
| 156 |
+
```
|
| 157 |
+
|
| 158 |
+
Reading one episode directly:
|
| 159 |
+
|
| 160 |
+
```python
|
| 161 |
+
import json
|
| 162 |
+
import pyarrow.parquet as pq
|
| 163 |
+
|
| 164 |
+
root = "quest-d2-pilot"
|
| 165 |
+
episodes = [json.loads(line) for line in open(f"{root}/meta/episodes.jsonl")]
|
| 166 |
+
ep = episodes[0]
|
| 167 |
+
frames = pq.read_table(
|
| 168 |
+
f"{root}/data/chunk-000/episode_{ep['episode_index']:06d}.parquet"
|
| 169 |
+
).to_pandas()
|
| 170 |
+
print(ep["tasks"][0], len(frames))
|
| 171 |
+
```
|
| 172 |
+
|
| 173 |
+
Iterate over `meta/episodes.jsonl` rather than over `range(n)`: episode
|
| 174 |
+
indices are not contiguous (see below).
|
| 175 |
+
|
| 176 |
+
## Provenance
|
| 177 |
+
|
| 178 |
+
The demonstrations were collected with the Quest teleoperation collector of
|
| 179 |
+
the GroundProbe benchmark codebase
|
| 180 |
+
(<https://github.com/AndersonYu7/Benchmark>, `policy/QuestTeleop/collect.sh`
|
| 181 |
+
with `SCENE=d2`). `meta/bad_seeds.jsonl` carries rejection timestamps from
|
| 182 |
+
2026-07-29 to 2026-10-03.
|
| 183 |
+
|
| 184 |
+
This release is the local dataset `quest_d2_pilot_merged_cleaned_2`. The
|
| 185 |
+
manifests record these processing steps:
|
| 186 |
+
|
| 187 |
+
1. **Merge** (`provenance/merge_manifest.json`, 2026-10-02,
|
| 188 |
+
`script/merge_quest_dataset.py`): 5 episodes from `quest_d2_10train_v2`
|
| 189 |
+
were appended to `quest_d2_pilot` as episodes 96–100, producing
|
| 190 |
+
`quest_d2_pilot_merged`. Existing episode indices were kept.
|
| 191 |
+
2. **Prune 1** (`provenance/prune_manifest_quest_d2_pilot_merged_cleaned.json`,
|
| 192 |
+
2026-10-02, `script/prune_quest_dataset.py`): 162 → 160 episodes
|
| 193 |
+
(16,358 → 16,329 frames). Dropped episodes 140 and 157, which were 14 and
|
| 194 |
+
15 frames long.
|
| 195 |
+
3. **Prune 2** (`prune_manifest.json`, 2026-10-02): 169 → 168 episodes
|
| 196 |
+
(17,279 → 17,265 frames). Dropped episode 172, which was 14 frames long.
|
| 197 |
+
|
| 198 |
+
The episode counts between steps do not chain (99 after the merge, 162
|
| 199 |
+
before prune 1; 160 after prune 1, 169 before prune 2). Further episodes were
|
| 200 |
+
evidently collected into the dataset between those steps; no manifest
|
| 201 |
+
describes those additions.
|
| 202 |
+
|
| 203 |
+
Both prune steps preserved original episode indices, so the manifests'
|
| 204 |
+
old-to-new mapping is the identity. The manifests contain absolute paths
|
| 205 |
+
from the collection machine.
|
| 206 |
+
|
| 207 |
+
## Limitations
|
| 208 |
+
|
| 209 |
+
- **Episode indices are not contiguous.** 168 episodes span indices 0–176.
|
| 210 |
+
Indices 17, 68, 111, 120, 128, 140, 153, 157 and 172 are absent.
|
| 211 |
+
- **`meta/info.json` is inconsistent with stock LeRobot v2.0 loaders.**
|
| 212 |
+
`chunks_size` is 168, but all files are in `chunk-000`, including episodes
|
| 213 |
+
168–171 and 173–176, which the usual `episode_index // chunks_size` rule
|
| 214 |
+
would look for in a non-existent `chunk-001`. `splits.train` is `"0:177"`,
|
| 215 |
+
an index range rather than an episode count. There is no `meta/stats.json`,
|
| 216 |
+
so normalisation statistics must be computed by the user. Reading the files
|
| 217 |
+
directly, as shown above, avoids all three issues.
|
| 218 |
+
- **Episodes 0–16 differ from the rest.** These 17 episodes (1,491 frames)
|
| 219 |
+
have `timestamp` steps of 0.040 s and videos encoded at 25 fps, while the
|
| 220 |
+
other 151 episodes (15,774 frames) have steps of 1/31 s and videos at
|
| 221 |
+
31 fps. `meta/info.json` states a single rate for the whole dataset.
|
| 222 |
+
Whether the underlying control rate of episodes 0–16 actually differed, or
|
| 223 |
+
only the recorded timestamps and encoding, has not been established. Their
|
| 224 |
+
parquet files also order the columns differently (`action.tcp_disp` last),
|
| 225 |
+
which matters only for code that reads columns by position.
|
| 226 |
+
- **Two tasks have a single episode.** `T_block_tray` and
|
| 227 |
+
`T_rubik_plate` each have one demonstration, both from the episode 0–16
|
| 228 |
+
group. The current benchmark
|
| 229 |
+
specification (`instructions.py`) designates those two pairs, together with
|
| 230 |
+
the three pairs absent from this release, as held-out test cells. All
|
| 231 |
+
episodes here are nonetheless labelled `split = train`. Exclude
|
| 232 |
+
`task_index` 8 and 11 if you intend to evaluate composition on those pairs.
|
| 233 |
+
- **Uneven coverage.** The ten main tasks range from 11 to 20 episodes; the
|
| 234 |
+
Rubik's cube tasks have the fewest.
|
| 235 |
+
- **Single complexity level, successes only.** Every episode is `cluttered`
|
| 236 |
+
and successful. There are no failed or recovery demonstrations.
|
| 237 |
+
- **`groundlab.tray_asset` is missing on 22 episodes** (0–16 and 96–100).
|
| 238 |
+
The container can be recovered from `groundlab.instruction_id`.
|
| 239 |
+
- **Timestamps use the integer rate.** Outside episodes 0–16 the `timestamp`
|
| 240 |
+
column advances by 1/31 s, not the nominal 1/31.25 s.
|
| 241 |
+
- **Short trajectories.** Episodes are 64–152 frames, roughly 2–5 seconds.
|
| 242 |
+
- **Simulation only, single embodiment.** No real-robot data.
|
| 243 |
+
- **Layout reproduction needs the simulator.** `layout_seed` reproduces an
|
| 244 |
+
arrangement only through the scene code in the benchmark repository.
|
| 245 |
+
- **Pilot scale.** This is a pilot collection, not a final benchmark split.
|
| 246 |
+
|
| 247 |
+
## Validation
|
| 248 |
+
|
| 249 |
+
Checked on the local copy before upload (2026-10-11):
|
| 250 |
+
|
| 251 |
+
- 168 lines in `meta/episodes.jsonl`, 168 parquet files, 168 + 168 mp4
|
| 252 |
+
files, with no missing or orphaned files.
|
| 253 |
+
- Every parquet has the columns declared in `meta/info.json`, a row count
|
| 254 |
+
equal to the episode `length`, a correct `episode_index`, `frame_index`
|
| 255 |
+
running 0…n−1, a `task_index` matching `meta/tasks.jsonl`, strictly
|
| 256 |
+
increasing timestamps, the declared vector shapes, no non-finite values,
|
| 257 |
+
and unit-norm TCP quaternions.
|
| 258 |
+
- `next.done` is true only on the last frame of each episode, and
|
| 259 |
+
`next.success` is true at least once in every episode.
|
| 260 |
+
- All 336 videos were fully decoded: H.264, 640×640, and a decoded frame
|
| 261 |
+
count equal to the parquet row count for every episode (17,265 frames per
|
| 262 |
+
camera).
|
| 263 |
+
|
| 264 |
+
## License
|
| 265 |
+
|
| 266 |
+
No license has been declared for this dataset. Scene and object assets used
|
| 267 |
+
to render the episodes come from upstream projects under their own terms.
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000000.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fd7e5d179bdb742f776eb7fee481367d69bee30a45047ab8a6b5b47a0d6f245a
|
| 3 |
+
size 18973
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000001.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4a4e9bca67b595a4869b9a76c93bb0c4e9573598d0e14495fc3a2d4c3c76662a
|
| 3 |
+
size 26525
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000002.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4441127f9947022f4507a4eb65923f95948063147d0cca2eddc17cf90a9fb71e
|
| 3 |
+
size 19557
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000003.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6a4fb98698415e6b0d40fc2fe85075aeb350336b8e48555bf2d6f83361ba0400
|
| 3 |
+
size 18809
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000004.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:253c18cab602cba73d2eccf16ac29fa77a21eec4acac34d2ff51e14d6e77ac32
|
| 3 |
+
size 18589
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000005.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5cb983dbaba5c543cdde92c9a110f1892173529dc3be549f0949ae8aa76a3e08
|
| 3 |
+
size 19938
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000006.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:547bea28506575835619c7f4f0ba745757f95b673181f6d9127dc296ca5b59d7
|
| 3 |
+
size 16264
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000007.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aef6766e60f3cd4fc87198624df3970e6f2910fb216d46e424296f25dcb7589d
|
| 3 |
+
size 20155
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000008.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:521c70a2f1a42e59518ccc77ec4c4b6df26ef52a3c4d1538459e28207dccc144
|
| 3 |
+
size 18817
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000009.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a4207511bb42f6be987924d2ff9b5e198c37f6ae00a6bbef823a541740dbbdd
|
| 3 |
+
size 21661
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000010.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:093293ed2fb3512bc8308f48cd8a5c19afd998e994b1ec18682a0de208b70646
|
| 3 |
+
size 22054
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000011.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2ae28e04e148b682e871d8dc36cc5a0dd9f3dfbacbc75426540db9f8c4f6bee
|
| 3 |
+
size 26446
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000012.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a54fc4ec8e0ed2d1f204bf60724595d575c7fb3e333ac55d2d4e9db89c3c8a54
|
| 3 |
+
size 24354
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000013.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47540b979019d77daf4efe0878ff95fe97f5be27dcbee7017bd8010477efe1ca
|
| 3 |
+
size 18609
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000014.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2c5f40f3a8fcade85191b5ea2e2d22c04e977522204ce5ed20884d110721f548
|
| 3 |
+
size 22647
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000015.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8c2c3db24eddadd0e07900b65bf8415baa5304ab47bf440458738441651c0e5e
|
| 3 |
+
size 25794
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000016.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0dd02f587d221670b66925b11dfce54ccdb60549faf7c105b2249f660bbc8393
|
| 3 |
+
size 19910
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000018.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6305f1b4f830bdbfb3ab234b9720e16c496dab25f935a134e6787b57f76ce146
|
| 3 |
+
size 19508
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000019.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a520254daabf42902b7b07f1154bea57286a81d52894839268a47dbaadd32721
|
| 3 |
+
size 21047
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000020.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0842768ca170a4dc8c0d3715af01e7fb607a71803af445a0a15afb567c94ec27
|
| 3 |
+
size 24986
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000021.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:332285dc6f89f7a3a0719b094afb8bae8a349f246aa362434753ac2ee05362d1
|
| 3 |
+
size 20178
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000022.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:af790e47c9985c6e134e08f7e5dd7a99a33aca27390587fe1c53fd880b37c26d
|
| 3 |
+
size 22462
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000023.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb03ab2639a8b21af36d12325c0faf60a5c1017b0ad71fa183d7181d401efa06
|
| 3 |
+
size 19371
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000024.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3360cc6c4a7355cba365c542d5d3eb7e30f5d4ada1026751d959c3d48f5ca9e3
|
| 3 |
+
size 23440
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000025.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:790a0d6ae3bad174bfd2750267718997beb43b8509742079a0af44b4f2d61497
|
| 3 |
+
size 23658
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000026.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1938db7ab93238177ea1afede03479dee5e19ea4ae2118e86b0937a666b6fd4e
|
| 3 |
+
size 21296
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000027.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:56e46607c53cc45b5d6c4af0ea93bb50ce79d0176023d50e2659bc0c5595ab85
|
| 3 |
+
size 23452
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000028.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2c9d687571a38e81bbe13cc6bb71752ea2ad9726b32afb8949c362af60cdef6a
|
| 3 |
+
size 20502
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000029.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40819debec8b3bb4e5a308980148f0df765c6010d5ce140b840bc21ad42f24b7
|
| 3 |
+
size 20902
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000030.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6074330dd1f16803c5c5c6a96ca342387c219892383b79627eb8f56a94af6ccb
|
| 3 |
+
size 23460
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000031.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cf72355757a3bdeb249addd9c9416cee16fc927faa16fbbe3eb5266e05886704
|
| 3 |
+
size 18379
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000032.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dfdb7b380a8c9583da9c43230387af23d096561d77bab49acc55fedfa02e1ade
|
| 3 |
+
size 21240
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000033.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:05d4f0e9ea7f3a62a5b365730cbe2ea127d24776df1c85ba9db7a3c020bd0358
|
| 3 |
+
size 27494
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000034.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aaddc1d003545c3a1a30e2c9e960c73c23803738247e28a85aefa35caab19fcb
|
| 3 |
+
size 22454
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000035.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3e857b0b56146b8f7a195336e57c249fa4e0b0b7b939400983b666d502837cb6
|
| 3 |
+
size 19524
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000036.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d048e414e7d2b4312b67f53c2bf163ab42f2024a788faeacbeb454959f5d7175
|
| 3 |
+
size 21455
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000037.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:20558312401d43f6dedcd0d32e0f24f2fee851a4ca20ce0ff126af10a64e60af
|
| 3 |
+
size 20502
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000038.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9fed16e8d90d6ac88d56bbfd0e7f3d9c597d35522035ea4a7d62a21f47752726
|
| 3 |
+
size 20118
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000039.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:afb9d0b9d675b9c96d7330005f2580a016a3f81f9344c912ee93c6148995e2cc
|
| 3 |
+
size 22220
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000040.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6cb9cbf641128eed630db3ef9e4b75c02a33cbc7ebcdc6838c49650d804dfbe4
|
| 3 |
+
size 25148
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000041.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:83003af73a78f71473e5af556643a6d05194890a59b3efbc2080b12ab5f61746
|
| 3 |
+
size 23990
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000042.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e17893e61372ef05affa006ce7017ddad0f45fc3be8e25dde9f8837444d927ec
|
| 3 |
+
size 28700
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000043.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2d2639b3d9fff6799364593d1ed4e310dae9caa7d40e12ef69cbf1299843aebd
|
| 3 |
+
size 26956
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000044.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1d3b6b9858b346d972297cbfaf411e0451e07ef2d4f1f8009a89524e41992145
|
| 3 |
+
size 24225
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000045.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9df1c8305ee0f00e1babc69e95280bbb4eb10cb12459e31a297633ee090d6419
|
| 3 |
+
size 27906
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000046.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6034307110f2f07ec05c73a83a64e5607d027e79d427c2dc10ff185d005fa8c2
|
| 3 |
+
size 20854
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000047.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5ee2a1f1d48df9a816592d8e3e204d62ec47d2cd7ff20d30767869fcbdd237a0
|
| 3 |
+
size 28437
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000048.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8484b68e0802c251a1c203fbbaed2ba311df76536be4d3204251d80f24ddc55
|
| 3 |
+
size 29229
|
quest_d2_pilot_merged_cleaned_2/data/chunk-000/episode_000049.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c4b203302dacd5dd132778253f60982d5a510014067f4817eeed1651aeb36f56
|
| 3 |
+
size 25062
|