dunnolab/so-combined-eng

---
license: apache-2.0
task_categories:
- robotics
tags:
- LeRobot
configs:
- config_name: default
data_files: data/*/*.parquet
language:
- en
size_categories:
- 1M<n<10M
---

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

## Dataset Description

The English version of this dataset integrates **598 open-source community datasets** into a single unified corpus, comprising **22,709 episodes** and approximately **9.4 million frames** across **563 distinct tasks**. Several transformations were applied to ensure standardization and data quality:

1. **Camera view normalization**
Because community datasets do not follow a consistent naming scheme for camera viewpoints, we used the `Qwen3-VL-8B-Instruct` model to categorize all images into one of three groups: **TOP**, **GRIPPER**, or **SIDE**. All datasets include TOP and GRIPPER viewpoints. Datasets lacking a SIDE viewpoint were padded with a zero-valued image.

2. **Task re-annotation**
We used `Qwen3-VL-8B-Instruct` to refine task annotations where necessary. The re-labeling process considered both the video inputs and the original task descriptions when available. For the Russian version of the dataset, all task descriptions were additionally translated into Russian.

3. **Video standardization**
To ensure reliable merging, all videos were re-encoded using the same codec (**H.264**), frame rate (**30 FPS**), and resolution (**480×640**).

4. **No-op removal**
We applied episodic-level trimming to remove no-op segments at the beginning and end of episodes and removed episodes composed entirely of no-op actions. Corresponding video and parquet files were cropped accordingly. This process eliminated **12.7%** of no-op data from the raw dataset.

- **License:** Apache-2.0

## Community Contributors Included

We acknowledge all community contributors whose datasets served as sources for this repository:

- 00ri
- 1zzx23
- 356c
- AndrejOrsula
- Askel1419
- BobBobbson
- CSCSXX
- CnLori
- Congying1112
- DGEs
- Daiki127
- Dangvi
- DanqingZ
- DorayakiLin
- EGLima
- Evelynix
- EverNorif
- Gano007
- Haribot099
- HarrisonLee24
- Hennadiy
- Jiangeng
- Kazu1232
- KeerthanKrish
- Killian74
- Kimz1
- LeRobot-worldwide-hackathon
- LemonadeDai
- LightwheelAI
- LittleFire99
- Loki0929
- Mazytomo
- Micksavioz
- Mwuqiu
- NeilKim
- Odog16
- Pi-robot
- Qiushuang
- RASPIAUDIO
- RaulSaya
- Rayenghali
- ReubenLim
- Revilo7
- RickRain
- Rorschach4153
- SahilChande
- SeanLMH
- SharkDan
- ShockleyWong
- Stevenyu8888
- SurajChess
- Thorns07
- Trelis
- TzuShian
- UN-kk
- VoicAndrei
- Xiewei1211
- YSanYi
- Yanis7777
- ZGGZZG
- Zak-Y
- ZibinDong
- aaron-ser
- aaronsu11
- abhiloiwal2
- abhisb
- abokinala
- acyanbird
- aiden-li
- alexis779
- allenchienxxx
- amrltqt
- andy309
- apayan
- aractingi
- arulloomba
- avea-robotics
- badwolf256
- bap13
- bensprenger
- boyangs235
- brcg3
- budinggou
- cHemingway
- cezarsolo
- cjlqwe
- cyoung96
- danaaubakirova
- davidgoss
- dc2ac
- demon-zozo
- desroziers
- dleon23
- dongseon
- drjaisree25
- dsfsg
- duthvik
- easonjcc
- edgarkim
- emmanuel-v
- enpeicv
- fbeltrao
- francescocrivelli
- frk2
- ganondorofu
- gmm0820
- guanfengliu
- gxy1111
- haijunsu-osu
- hannb
- hoon-shin
- howld
- hrhraj
- huyouare
- jchun
- jcsux
- jiajun001
- jlesein
- jmendoza-10
- jpizarrom
- juni3227
- jyang-ca
- k1000dai
- kagyvro48
- kaiserbuffle
- kaiyuwu010
- karimnihal
- kivod
- kkurzweil
- kristaqp
- legion1581
- leolin6
- lerobot
- lerobot-edinburgh-white-team
- liamlau
- lijinghai
- lime66
- littledragon
- liyitenga
- ljw1105
- love3165303
- lucasfv
- luriss
- maitereo
- masakinoda
- masato-ka
- mathieutk
- nbirukov
- northhycao
- nuoyihan
- omkarmayekar555
- opan08
- oretti
- orsoromeo
- pandaRQ
- paultr
- pbvr
- pdd46465
- pr0tos
- pranavsaroha
- psavnani5
- ptizzza
- puneetpanwar
- reeced
- ricky0526
- roboticshack
- rowb1
- rs545837
- ryanpennings
- s-higurashi
- samanthalhy
- samsam0510
- samsitol
- seonixx
- seunghoney
- shylee
- slowturtle99
- sshh11
- strainflow
- suessmann
- sunq
- szfforever
- taiobot
- targabor
- tfoldi
- therarelab
- thimble
- tinkhireeva
- tkc79
- tlf123
- tobdeu
- triton7777
- un1c0rnio
- uuysi
- vednot25t
- wangranryan
- weblucas
- weiye11
- wvangils
- y1y2y3
- yingliu-data
- yinxinyuchen
- yo-michi22
- youliangtan
- yuk6ra
- yunhezhui123
- yuto083
- yuz1wan
- zacapa
- zaringleb
- zheng6677
- zlj666
- zonglin1104

## Dataset Structure

[meta/info.json](meta/info.json):
```json
{
"codebase_version": "v2.1",
"robot_type": "so100",
"total_episodes": 22709,
"total_frames": 9443507,
"total_tasks": 563,
"total_videos": 68127,
"total_chunks": 23,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:22709"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4",
"features": {
"action": {
"dtype": "float32",
"shape": [
6
],
"names": [
"main_shoulder_pan",
"main_shoulder_lift",
"main_elbow_flex",
"main_wrist_flex",
"main_wrist_roll",
"main_gripper"
]
},
"observation.state": {
"dtype": "float32",
"shape": [
6
],
"names": [
"main_shoulder_pan",
"main_shoulder_lift",
"main_elbow_flex",
"main_wrist_flex",
"main_wrist_roll",
"main_gripper"
]
},
"observation.images.gripper": {
"dtype": "video",
"shape": [
480,
640,
3
],
"names": [
"height",
"width",
"channels"
],
"info": {
"video.height": 480,
"video.width": 640,
"video.codec": "h264",
"video.pix_fmt": "yuv420p",
"video.is_depth_map": false,
"video.fps": 30,
"video.channels": 3,
"has_audio": false
}
},
"observation.images.top": {
"dtype": "video",
"shape": [
480,
640,
3
],
"names": [
"height",
"width",
"channels"
],
"info": {
"video.height": 480,
"video.width": 640,
"video.codec": "h264",
"video.pix_fmt": "yuv420p",
"video.is_depth_map": false,
"video.fps": 30,
"video.channels": 3,
"has_audio": false
}
},
"observation.images.side": {
"dtype": "video",
"shape": [
480,
640,
3
],
"names": [
"height",
"width",
"channels"
],
"info": {
"video.height": 480,
"video.width": 640,
"video.codec": "h264",
"video.pix_fmt": "yuv420p",
"video.is_depth_map": false,
"video.fps": 30,
"video.channels": 3,
"has_audio": false
}
},
"timestamp": {
"dtype": "float32",
"shape": [
1
],
"names": null
},
"frame_index": {
"dtype": "int64",
"shape": [
1
],
"names": null
},
"episode_index": {
"dtype": "int64",
"shape": [
1
],
"names": null
},
"index": {
"dtype": "int64",
"shape": [
1
],
"names": null
},
"task_index": {
"dtype": "int64",
"shape": [
1
],
"names": null
}
},
"repo_id": "dunno/merged"
}
```