Datasets:
File size: 7,228 Bytes
0380f4e 3eda8ae 0380f4e 3eda8ae 0380f4e 3eda8ae 0380f4e 3eda8ae 0380f4e 59f2038 0380f4e 3eda8ae 0380f4e 305348c 0380f4e 3eda8ae 0380f4e 305348c 3eda8ae 305348c 3eda8ae 305348c 3eda8ae 0380f4e 0c7435c d00d9c1 0c7435c 7e9a9d4 0380f4e 3eda8ae 0380f4e | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 | ---
license: apache-2.0
task_categories:
- robotics
tags:
- LeRobot
- so101
- robotics
configs:
- config_name: default
data_files: data/*/*.parquet
---
# so101_coffee_subtask
431 teleoperated SO-101 demonstrations of a coffee-making routine, recorded as **12 separately instructed steps**, each captured with two cup variants — a **red** cup and a **white** one.
Every step is a short, self-contained manipulation with its own natural-language instruction, so the set can be used either as 24 single-skill datasets or as one long-horizon routine.
This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).
<a class="flex" href="https://huggingface.co/spaces/lerobot/visualize_dataset?path=nevertmr/so101_coffee_subtask">
<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-xl-dark.svg"/>
</a>
## Overview
| | |
|---|---|
| Episodes / frames | 431 / 99,472 (0.92 h) |
| Composition | 12 steps × 2 cup colours = 24 contiguous blocks |
| Episode length | mean 231 frames (7.7 s), median 223, range 135–434 |
| Distinct instructions | 10 |
| Robot | so_follower, 6-DoF — `shoulder_pan`, `shoulder_lift`, `elbow_flex`, `wrist_flex`, `wrist_roll`, `gripper` |
| Cameras | `observation.images.front`, `observation.images.wrist` — 640×480, av1 |
| fps | 30 |
| Format | LeRobotDataset v3.0 |
## Trajectories
End-effector (pinch-point) paths of every episode, obtained by forward kinematics from the recorded joint states. Top panel is a top-down X–Y view with the table, the two coffee machines and the two circular markers drawn in; bottom panel is the side X–Z view. The robot base is at the origin (▲). Red lines are red-cup episodes, blue lines white-cup ones.
Three steps to a row — the full 12-step routine in order:
**Steps 1–3** — `pick up the cup on the blue circle` → `place the cup to the left coffee machine` → `press the blue button of the left coffee machine`

**Steps 4–6** — `pick up the cup on the blue circle` → `place the cup to the right coffee machine` → `press the blue button of the right coffee machine`

**Steps 7–9** — `pick up the cup on the left coffee machine` → `place the cup to the left pink circle` → `pick up the cup on the right coffee machine`

**Steps 10–12** — `place the cup to the left pink circle` → `pick up the cup near the blue circle` → `place the cup on the blue circle`

All four 3-step rows are in [`viz/`](https://huggingface.co/datasets/nevertmr/so101_coffee_subtask/tree/main/viz).
## Trajectory video
[](https://youtu.be/ey8GLF_dZ2I)
▶ Watch on YouTube (4K): https://youtu.be/ey8GLF_dZ2I
## Episode ranges
The cup colour is not part of the instruction string, so this table is what maps an episode to its step **and** cup. Every combination is one contiguous block of `episode_index`:
| `episode_index` | Step | Cup | Episodes | Frames | Instruction |
|---|---:|---|---:|---:|---|
| **0–19** | 1 | red | 20 | 6,058 | `pick up the cup on the blue circle` |
| **20–39** | 1 | white | 20 | 5,204 | `pick up the cup on the blue circle` |
| **40–59** | 2 | red | 20 | 5,652 | `place the cup to the left coffee machine` |
| **60–79** | 2 | white | 20 | 4,883 | `place the cup to the left coffee machine` |
| **80–99** | 3 | red | 20 | 5,231 | `press the blue button of the left coffee machine` |
| **100–119** | 3 | white | 20 | 4,838 | `press the blue button of the left coffee machine` |
| **120–139** | 4 | red | 20 | 4,323 | `pick up the cup on the blue circle` |
| **140–159** | 4 | white | 20 | 4,093 | `pick up the cup on the blue circle` |
| **160–175** | 5 | red | 16 | 2,886 | `place the cup to the right coffee machine` |
| **176–192** | 5 | white | 17 | 3,461 | `place the cup to the right coffee machine` |
| **193–209** | 6 | red | 17 | 3,859 | `press the blue button of the right coffee machine` |
| **210–226** | 6 | white | 17 | 3,989 | `press the blue button of the right coffee machine` |
| **227–243** | 7 | red | 17 | 4,020 | `pick up the cup on the left coffee machine` |
| **244–260** | 7 | white | 17 | 3,638 | `pick up the cup on the left coffee machine` |
| **261–277** | 8 | red | 17 | 4,171 | `place the cup to the left pink circle` |
| **278–294** | 8 | white | 17 | 4,133 | `place the cup to the left pink circle` |
| **295–311** | 9 | red | 17 | 4,022 | `pick up the cup on the right coffee machine` |
| **312–328** | 9 | white | 17 | 3,490 | `pick up the cup on the right coffee machine` |
| **329–345** | 10 | red | 17 | 3,920 | `place the cup to the left pink circle` |
| **346–363** | 10 | white | 18 | 3,999 | `place the cup to the left pink circle` |
| **364–380** | 11 | red | 17 | 3,488 | `pick up the cup near the blue circle` |
| **381–397** | 11 | white | 17 | 3,576 | `pick up the cup near the blue circle` |
| **398–413** | 12 | red | 16 | 3,058 | `place the cup on the blue circle` |
| **414–430** | 12 | white | 17 | 3,480 | `place the cup on the blue circle` |
| | | | **431** | **99,472** | |
```python
from lerobot.datasets.lerobot_dataset import LeRobotDataset
ds = LeRobotDataset("nevertmr/so101_coffee_subtask")
step3_red = LeRobotDataset("nevertmr/so101_coffee_subtask", episodes=list(range(80, 100)))
```
## Steps and instructions
Steps 1 and 4 / 8 and 10 reuse an instruction: the routine makes two cups, so two of the instructions occur at two different points in the sequence. `task_index` is the index of the instruction, not of the step — there are 12 steps but 10 distinct instructions.
| Step | `task_index` | Instruction | Episodes | Frames |
|---:|---:|---|---:|---:|
| 1 | 0 | `pick up the cup on the blue circle` | 40 | 11,262 |
| 2 | 1 | `place the cup to the left coffee machine` | 40 | 10,535 |
| 3 | 2 | `press the blue button of the left coffee machine` | 40 | 10,069 |
| 4 | 0 | `pick up the cup on the blue circle` *(same as step 1)* | 40 | 8,416 |
| 5 | 3 | `place the cup to the right coffee machine` | 33 | 6,347 |
| 6 | 4 | `press the blue button of the right coffee machine` | 34 | 7,848 |
| 7 | 5 | `pick up the cup on the left coffee machine` | 34 | 7,658 |
| 8 | 6 | `place the cup to the left pink circle` | 34 | 8,304 |
| 9 | 7 | `pick up the cup on the right coffee machine` | 34 | 7,512 |
| 10 | 6 | `place the cup to the left pink circle` *(same as step 8)* | 35 | 7,919 |
| 11 | 8 | `pick up the cup near the blue circle` | 34 | 7,064 |
| 12 | 9 | `place the cup on the blue circle` | 33 | 6,538 |
|