File size: 4,519 Bytes
7a6460c
 
 
 
 
 
4611395
 
 
 
 
 
7a6460c
 
 
 
 
016fefe
7a6460c
016fefe
 
 
 
 
 
4611395
 
50e2d46
016fefe
7a6460c
4611395
7a6460c
 
4611395
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7a6460c
016fefe
 
 
 
4611395
016fefe
4611395
 
 
 
016fefe
 
4611395
 
 
 
 
 
016fefe
4611395
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
016fefe
4611395
 
7a6460c
 
4611395
 
 
 
7a6460c
4611395
7a6460c
016fefe
 
 
 
 
 
 
 
 
 
7a6460c
4611395
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
---
license: apache-2.0
task_categories:
- robotics
tags:
- LeRobot
- robotics
- isaac-sim
- assembly
- manipulation
- imitation-learning
- contact-rich
configs:
- config_name: default
  data_files: data/*/*.parquet
---

# AssembleBench

Multi-task contact-rich assembly demonstrations for imitation learning / VLA fine-tuning,
and the training set behind the [AssembleBench](https://github.com/hud-evals/assemble-bench)
benchmark: 14 tasks on the NIST ATB-1 taskboard (insert round and square pegs, mesh gears,
thread nuts) on the DROID robot platform.

Read the writeup: [Benchmarking Robot Models on Contact-Rich Assembly](https://www.hud.ai/blog/assemble-benchmark).

This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).

<a class="flex" href="https://huggingface.co/spaces/lerobot/visualize_dataset?path=hud-evals/AssembleBench">
<img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl.svg"/>
<img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-dark-xl.svg"/>
</a>

## Summary

| | |
|---|---|
| Episodes | **1355** |
| Frames | **444651** |
| FPS | 15 |
| Tasks | 14 (pegs, gears, nuts; **no M20**) |
| Action | DROID 8-D joint position (7 arm + binary gripper) |
| Cameras | `observation.images.front`, `observation.images.wrist` at 180x320 |
| Proprio | `observation.state` (8), `joint_vel` (7), `eef_pos` (3), `eef_quat` (4) |

## Tasks (episode counts)

Language-conditioned pick / insert / mesh / thread:

| family | task | n |
|--------|------|---|
| peg_round | 4 / 8 / 12 / 16 mm | 96 / 99 / 100 / 100 |
| peg_square | 4 / 8 / 12 / 16 mm | 98 / 98 / 99 / 100 |
| gear_mesh | small / medium / large | 92 / 97 / 92 |
| nut_thread | M8 / M12 / M16 | 90 / 95 / 99 |

**M20 removed:** embedding analysis (PCA / t-SNE) showed M20 as a separate mode that split the nut cluster; it is not included.

## How it was built

Built from a cleaned and topped-up revision of
[`lukasskellijs/assembly_bench_2`](https://huggingface.co/datasets/lukasskellijs/assembly_bench_2),
with a second QC pass for duplicates, idle chunks, length outliers, and within-family
Mahalanobis outliers. Pipeline:

1. Start from the cleaned v2 revision (911 eps: dups + long outliers already removed from the original 1500).
2. Drop all **M20** nut episodes.
3. Append scripted-expert top-up HDF5 demos that refill the gaps left by the clean pass (success-filtered Isaac Lab recordings).
4. QC blocklist (45 trajectories by MD5 of full `action || state`):
   - exact-duplicate extras (none remained after clean + ingest dedup)
   - idle runs >= 15 frames (action-chunk size used in training)
   - per-task length >3 sigma and clear within-family **20-PC Mahalanobis** outliers (retry / thrash / bad seat)
5. Final exact-duplicate scan must be zero or the build aborts.

Quality over quantity: mild multivariate peg shape outliers that look like valid successes were kept.

## Schema

DROID joint-position contract:

| key | shape | notes |
|-----|-------|--------|
| `action` | (8,) | absolute joint targets + gripper |
| `observation.state` | (8,) | joint pos (7) + gripper |
| `observation.joint_vel` | (7,) | rad/s |
| `observation.eef_pos` | (3,) | base-frame XYZ (m) |
| `observation.eef_quat` | (4,) | world WXYZ |
| `observation.images.front` | video HxWx3 | 180x320 |
| `observation.images.wrist` | video HxWx3 | 180x320 |

## Load

```python
from lerobot.datasets.lerobot_dataset import LeRobotDataset

ds = LeRobotDataset("hud-evals/AssembleBench")
print(ds.num_episodes, ds.num_frames)
print(ds[0]["observation.images.wrist"].shape)
```

## Intended use

- Supervised fine-tuning of VLA / diffusion / ACT policies with **action chunk size 15**
- Multi-task assembly benchmarking (peg / gear / nut families)

Not a teleop human dataset. Demonstrations are from a privileged scripted expert in Isaac Sim / Isaac Lab; visuals and proprio match the policy observation contract.

## Reference checkpoints

pi0.5 finetunes trained on this dataset:

| Checkpoint | What it is |
|---|---|
| [`pi05-AssembleBench-12k`](https://huggingface.co/hud-evals/pi05-AssembleBench-12k) | behavior-cloning baseline, all 14 tasks |
| [`pi05-AssembleBench-cgdagger-r3`](https://huggingface.co/hud-evals/pi05-AssembleBench-cgdagger-r3) | BC + 3 rounds of code-gated DAgger, the final checkpoint |

## Lineage

- Simulator: NVIDIA Isaac Sim 6 / Isaac Lab Arena
- Format: LeRobot v3.0