The dataset could not be loaded because the splits use different data file formats, which is not supported. Read more about the splits configuration. Click for more details.
Error code: FileFormatMismatchBetweenSplitsError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
IROS-2025-Challenge-Manip
Dataset Summary π
This dataset contains the IROS Challenge - Manipulation Track benchmark, organized into pretrain, train, and validation splits.
- Pretrain split: ~20,000 single pick-and-place trajectories, packaged into tar files (each containing ~1,000 trajectories).
- Train split: task-specific demonstrations, with ~100 trajectories provided per task.
- Validation split: includes the test-time scenes and object assets in USD format.
Each trajectory in the pretrain and train splits contains:
- Multi-view video recordings (three perspectives: head-mounted camera and two wrist cameras)
- Robot states (joint positions, gripper states, etc.)
- Actions corresponding to the task execution
This dataset is designed to support pretraining, task-specific fine-tuning, and evaluation for robotic manipulation in the IROS Challenge setting.
Get started π₯
Download the Dataset
To download the full dataset, you can use the following code. If you encounter any issues, please refer to the official Hugging Face documentation.
from huggingface_hub import snapshot_download
dataset_path = snapshot_download("InternRobotics/IROS-2025-Challenge-Manip", repo_type="dataset")
Please execute this Python file to post-process the validation set.
cd IROS-2025-Challenge-Manip
python dataset_post_processing.py validation
Unzip the pretrain dataset
cd pretrain
for i in {1..20}; do
echo "Extracting $i.tar.gz ..."
tar -xzf "$i.tar.gz"
done
Dataset Structure
pretrain Folder hierarchy
pretrain
βββ 1.tar.gz
β βββ 1/
β βββ data/
β βββ meta/
β βββ videos/
βββ 2.tar.gz
β βββ 2/
β βββ data/
β βββ meta/
β βββ videos/
...
βββ 20.tar.gz
βββ 20/
βββ data/
βββ meta/
βββ videos/
train Folder hierarchy
train
βββ collect_three_glues
β βββ data/
β βββ meta/
β βββ videos/
βββ collect_two_alarm_clocks/
βββ collect_two_shoes/
βββ gather_three_teaboxes/
βββ make_sandwich/
βββ oil_painting_recognition/
βββ organize_colorful_cups/
βββ purchase_gift_box/
βββ put_drink_on_basket/
βββ sort_waste/
validation Folder hierarchy
validation
βββ IROS_C_V3_Aloha_seen
β βββ collect_three_glues
β β βββ 000
β β β βββ meta_info.pkl
β β β βββ scene.usd
β β β βββ SubUSDs -> ../SubUSDs
β β βββ 001/
β β βββ 002/
β β βββ 003/
β β βββ 004/
β β βββ 005/
β β βββ 006/
β β βββ 007/
β β βββ 008/
β β βββ 009/
β β βββ SubUSDs
β β βββ materials/
β β βββ textures/
β βββ collect_two_alarm_clocks/
β βββ collect_two_shoes/
β βββ gather_three_teaboxes/
β βββ make_sandwich/
β βββ oil_painting_recognition/
β βββ organize_colorful_cups/
β βββ purchase_gift_box/
β βββ put_drink_on_basket/
β βββ sort_waste/
βββ IROS_C_V3_Aloha_unseen
βββ collect_three_glues/
βββ collect_two_alarm_clocks/
βββ collect_two_shoes/
βββ gather_three_teaboxes/
βββ make_sandwich/
βββ oil_painting_recognition/
βββ organize_colorful_cups/
βββ purchase_gift_box/
βββ put_drink_on_basket/
βββ sort_waste/
License and Citation
All the data and code within this repo are under CC BY-NC-SA 4.0. Please consider citing our project if it helps your research.
@misc{contributors2025internroboticsrepo,
title={IROS-2025-Challenge-Manip Colosseum},
author={IROS-2025-Challenge-Manip Colosseum contributors},
howpublished={\url{https://github.com/internrobotics/IROS-2025-Challenge-Manip}},
year={2025}
}
- Downloads last month
- 8,423