datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
fish_datasets_real_electrodyn_expertsys_twodim_fourierDexora_Real-World_Dataset
Dexora: Open-Source VLA for High-DoF Bimanual Dexterity
🔥 News & Updates
2025-12-03: Released the full Real-World Dataset (12.2K episodes) on Hugging Face.
2025-12-12: Released the task-level view of the Real-World Dataset (one folder per high-level task) on Hugging Face.
Coming soon: We will open-source the full 100K-episode simulation dataset.
--
📊 Dataset Overview
The Dexora corpus combines high-fidelity real-world teleoperation data with… See the full description on the dataset page: https://huggingface.co/datasets/Dexora/Dexora_Real-World_Dataset.real_robot_data
Real Robot Manipulation Dataset
Tabletop manipulation tasks recorded on a UR5e with a Robotiq 2F-140
gripper. Most tasks use two Intel RealSense cameras (fixed base + wrist-mounted);
the newest hide_block_with_cup task adds a third (opposite-side counter view).
All recordings at 20 Hz.
Tasks
Folder
Eps
Cameras
Instruction
put_blocks_in_plate
100
2
Pick two blocks from the table and place them on the plate.
remove_blocks_from_plate
100
2
Pick two… See the full description on the dataset page: https://huggingface.co/datasets/Hoshipu/real_robot_data.SWE-Lego-Real-Data-Verified
SWE-Lego-Real-Data-Verified
Gold-patch-validated subset of
PrimeIntellect/SWE-Lego-Real-Data
(itself a fixed fork of SWE-Lego's real-data split). The
resolved split contains 4,323 / 4,432 rows (97.54%) verified scoreable end-to-end: apply
test_patch, apply the gold patch, run the row's test_cmd in its image, require every
F2P/P2P test to report PASSED.
Changes vs upstream
Validation-only subset — our passes: one full pass at concurrency 200, then a 10× retry… See the full description on the dataset page: https://huggingface.co/datasets/PrimeIntellect/SWE-Lego-Real-Data-Verified.real-infrared-maritime-vessel-dataset
Real Infrared Maritime Vessel Dataset
Real infrared imagery of maritime vessels.
The dataset is provided in three forms — full-frame detection images, per-object classification crops, and a hand-curated subset.
Classes (7): liner, bulk carrier, warship, sailboat, canoe, container ship, fishing boat.
Layout
real-infrared-maritime-vessel-dataset/
├── original/ Full-frame IR images + XML bounding-box labels (detection)
│ ├── images/{train,test}/*.jpg… See the full description on the dataset page: https://huggingface.co/datasets/hanchong/real-infrared-maritime-vessel-dataset.old-data-store-realreal_world_datalibero_filtered_noops_rlds_dataset
Modified LIBERO RLDS Datasets
This repository contains the four modified LIBERO datasets used in the OpenVLA fine-tuning experiments, as well as a dataset generated from the Libero-90 task suite, stored in RLDS data format. See Appendix E in the
OpenVLA paper for details about the specific dataset modifications.
Originally provided with the v2 release of the OpenVLA paper, a script to generate the modified datasets in raw HDF5 format is available here, and the code used to convert… See the full description on the dataset page: https://huggingface.co/datasets/real-lab/libero_filtered_noops_rlds_dataset.fish_datasets_real_fowlers_expertsys_twodim_fourier_v2fish_datasets_real_fowlers_twodim_left_to_right_recovery_v2AI-Generated-vs-Real-Images-Datasets
Dataset Card for "AI-Generated-vs-Real-Images-Datasets"
More Information needed
fish_datasets_real_electrodyn_twodim_left_to_rightForceVLA-real-datafish_datasets_real_electrodyn_expertsys_twodim_fourier_v2Processed_real_dataSWE-Lego-Real-Data
Dataset Summary
Paper | Github | HF Collection
SWE-Lego-Real-Data contains 18k real github issues (Python language) and their multi-turn agent trajectories. The column named messages is collected using Qwen/Qwen3-Coder-480B-A35B-Instruct with OpenHands (v0.53.0) agent scaffolding, which can be directly used for SFT training.
Dataset Structure
.
└── data
├── resolved-00000-of-00001.parquet (5k github issues with resolved trajectories)
└──… See the full description on the dataset page: https://huggingface.co/datasets/Lego-X/SWE-Lego-Real-Data.KIT_IRL_Real_Robot_Kitchen_Dataset_Vis_OnlyDataset with joint state play data for a toy kitchen. Does not contain language annotations.
The dataset contains 20 different tasks with two static cameras.
@inproceedings{
reuss2024multimodal,
title={Multimodal Diffusion Transformer: Learning Versatile Behavior from Multimodal Goals},
author={Moritz Reuss and {\"O}mer Erdin{\c{c}} Ya{\u{g}}murlu and Fabian Wenzel and Rudolf Lioutikov},
booktitle={Robotics: Science and Systems},
year={2024}
}
imda_dataset_clean_real
Dataset Card for "imda_dataset_clean_real"
More Information needed
fish_datasets_real_electrodyn_twodim_2024-08-07-23-55-23real-infrared-maritime-vessel-dataset
Real Infrared Maritime Vessel Dataset
Real infrared imagery of maritime vessels.
The dataset is provided in three forms — full-frame detection images, per-object classification crops, and a hand-curated subset.
Classes (7): liner, bulk carrier, warship, sailboat, canoe, container ship, fishing boat.
Layout
real-infrared-maritime-vessel-dataset/
├── original/ Full-frame IR images + XML bounding-box labels (detection)
│ ├── images/{train,test}/*.jpg… See the full description on the dataset page: https://huggingface.co/datasets/Abin0008/real-infrared-maritime-vessel-dataset.Dataset-Real-World
4DCodeBench-RealWorld
This dataset is part of 4DCodeBench, a benchmark for evaluating coding agents on inverse graphics of dynamic scenes. It contains 100 real-world cases, each with a video of a physical event and a hand-annotated mask of the pixels that move in it. 23 of the videos cannot be redistributed; one script rebuilds them from their public sources.
Resources:
Paper: 4DCodeBench: Benchmarking Agents on Inverse Graphics of Dynamic Scenes
Project page:… See the full description on the dataset page: https://huggingface.co/datasets/4DCodeBench/Dataset-Real-World.fish_datasets_real_german_expertsys_twodim_fourier_v2KIT_IRL_Real_Robot_Kitchen_Dataset_LangDataset with joint state play data for a toy kitchen.
The dataset contains 20 different tasks with two static cameras.
@inproceedings{
reuss2024multimodal,
title={Multimodal Diffusion Transformer: Learning Versatile Behavior from Multimodal Goals},
author={Moritz Reuss and {\"O}mer Erdin{\c{c}} Ya{\u{g}}murlu and Fabian Wenzel and Rudolf Lioutikov},
booktitle={Robotics: Science and Systems},
year={2024}
}
real_layout_dataset_20260930
Real Layout Dataset (2026-09-30)
This snapshot contains 47,767 source images and 24,301 JSON annotations. The
images include 2,618 KBG images and 3,306 NeurIPS paper-figure images. Every
JSON annotation has a source image with the same filename stem; not every image
has an annotation.
Files
images_20260930.zip.part-000 through images_20260930.zip.part-022:
consecutive pieces of one ZIP archive containing the images/ directory.
images_20260930.zip.parts.sha256:… See the full description on the dataset page: https://huggingface.co/datasets/EditFigure/real_layout_dataset_20260930.gemma2_9b_it_taboo_salt_oracle_v1-training-dataUS-Real-time-gun-detection-in-CCTV-An-open-problem-datasetgemma2_9b_it_taboo_song_oracle_v1-training-datagemma2_9b_it_taboo_flag_oracle_v1-training-datareal-infrared-maritime-vessel-dataset-super-resolved
Real Infrared Maritime Vessel Dataset (Super-Resolved, HYPIR)
HYPIR-restored crops of the real infrared maritime vessel dataset (cropped/), with vs without a fixed class-name text prompt guiding restoration.
Layout
real-infrared-maritime-vessel-dataset-super-resolved/
├── w_cat_prompt/ restored with prompt: "infrared image of a <class-name> on sea water"
│ ├── labels.txt, {train,val,test}.csv
│ ├── train/{0..6}/*.png 25,045
│ ├── val/{0..6}/*.png… See the full description on the dataset page: https://huggingface.co/datasets/hanchong/real-infrared-maritime-vessel-dataset-super-resolved.real-infrared-maritime-vessel-dataset
Real Infrared Maritime Vessel Dataset
Real infrared imagery of maritime vessels.
The dataset is provided in three forms — full-frame detection images, per-object classification crops, and a hand-curated subset.
Classes (7): liner, bulk carrier, warship, sailboat, canoe, container ship, fishing boat.
Layout
real-infrared-maritime-vessel-dataset/
├── original/ Full-frame IR images + XML bounding-box labels (detection)
│ ├── images/{train,test}/*.jpg… See the full description on the dataset page: https://huggingface.co/datasets/liujinfansjtu89/real-infrared-maritime-vessel-dataset.