datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
SLAM-EVALReplica-SLAMTUM_RGBD-SLAMTUM_RGBD-SLAMhilti-trimble-slam-challenge-2026
Hilti x Trimble SLAM Challenge 2026
The Hilti x Trimble SLAM Challenge 2026 dataset is a real-world robotics benchmark for evaluating visual-inertial SLAM and localization systems on active construction sites.
The dataset combines synchronized dual-fisheye imagery and inertial measurements with building floor plan priors and LiDAR-derived reference trajectories. It was created through a collaboration between Hilti, Trimble, and the Dynamic Robot Systems Group at the University… See the full description on the dataset page: https://huggingface.co/datasets/Hilti-Research/hilti-trimble-slam-challenge-2026.rtk-slam-dataset
RTK-SLAM Dataset
An RTK-SLAM Dataset for Absolute Accuracy Evaluation in GNSS-Degraded Environments
Wei Zhang, Vincent Ress, David Skuddis, Uwe Soergel, Norbert HaalaInstitute for Photogrammetry and Geoinformatics, University of Stuttgart, Germany
[Project Page] | [Paper] | [Code]
Overview
This dataset is designed for evaluating the absolute global positioning accuracy of RTK-SLAM systems in GNSS-degraded and GNSS-denied environments. A key design principle is… See the full description on the dataset page: https://huggingface.co/datasets/Willyzw/rtk-slam-dataset.e2e-stream-slam-training-dataset
e2e-stream-slam training assets
Reproducibility bundle for the V4 SLAMFormer ablation suite.
Code: https://github.com/SlamMate/e2e-semantic-SLAM/tree/submap (commit 3195a7a)
Contents
Checkpoints
File
Size
Role
checkpoints/v1_paper_ckpt10.pth
3.6 GB
SLAMFormer paper base ckpt (10 ep on the paper datasets). PRETRAINED init for V3 Scale Token training.
checkpoints/v3_scale_token_ckpt2.pth
3.8 GB
V3 Scale Token epoch-2 (3 ep, 3×A6000… See the full description on the dataset page: https://huggingface.co/datasets/qizhangslam/e2e-stream-slam-training-dataset.Functional-SLAM-datasetWild-SLAMThis repository contains data for WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments.
Paper | Project Page | Code
WildGS-SLAM accurately tracks the camera trajectory and reconstructs a 3D Gaussian map for static elements from a monocular video sequence, effectively removing dynamic components.
Datasets Used
WildGS-SLAM uses data from the following datasets:
Wild-SLAM Mocap Dataset: (Hugging Face) Download instructions are available in the github repository.… See the full description on the dataset page: https://huggingface.co/datasets/gradient-spaces/Wild-SLAM.Unblur_slam_traning_datasetAgentic_SLAMNPU-SLAM
NPU-SLAM Dataset
A LiDAR-Camera-Inertial SLAM dataset built for FAST-LIVO2, with STM32 hardware-triggered synchronization across a Livox Mid360 LiDAR and three cameras.
Sensors
Livox Mid360 (LiDAR) · built-in IMU · 3× UVC camera (cam0 main, cam1/cam2 auxiliary)
Format
ROS1 bag rosbag2.0
Frames
LiDAR 10 Hz, IMU 200 Hz, Cameras ~10 Hz (640×480)
Synchronization
STM32 hardware trigger, shared clock
License
CC BY 4.0
Get the data
git clone… See the full description on the dataset page: https://huggingface.co/datasets/lan374/NPU-SLAM.glide-slam-example
TUM RGB-D freiburg3_long_office_household — RGB only
A verbatim copy of one sequence of the TUM RGB-D benchmark, used as the sample
input for the GLidE-SLAM pixi run demo
task. It is mirrored here only so the demo can fetch a single sequence quickly;
the authoritative copy is the one published by TUM.
Source
https://cvg.cit.tum.de/rgbd/dataset/freiburg3/rgbd_dataset_freiburg3_long_office_household.tgz
Contents… See the full description on the dataset page: https://huggingface.co/datasets/pablovela5620/glide-slam-example.fold_new_slamThis dataset was created using LeRobot.
Dataset Description
Data Distribution Overview
This figure summarizes the data distribution of the ywxia/fold_new_slam dataset, auto-generated after each conversion via analysis/postprocess_with_overview.py. It shows episode-length distribution, the 3-D EEF workspace, per-dimension state histograms, per-arm action magnitudes, and a sample of frames from each camera.
Task: fold the box on the desk
Episodes: 59 | Frames: 21264 |… See the full description on the dataset page: https://huggingface.co/datasets/ywxia/fold_new_slam.fold_box_dual_arm_60_slamThis dataset was created using LeRobot.
Dataset Description
Data Distribution Overview
This figure summarizes the data distribution of the ywxia/fold_box_dual_arm_60_slam dataset, auto-generated after each conversion via analysis/postprocess_with_overview.py. It shows episode-length distribution, the 3-D EEF workspace, per-dimension state histograms, per-arm action magnitudes, and a sample of frames from each camera.
Task: fold the box on the desk
Episodes: 59 |… See the full description on the dataset page: https://huggingface.co/datasets/ywxia/fold_box_dual_arm_60_slam.Replica-SLAMCP-SLAM_datasetvggt_slam_processed_dataSLAM_project
Omni Instrument SLAM Project Dataset
The Omni Instrument SLAM Project Dataset is a compact robotics dataset designed for evaluating stereo, visual-inertial, and visual-inertial odometry (VIO) pipelines.
It provides:
Stereo Image Pairs
Inertial measurements (IMU)
Ground-truth 6 DoF pose (for VIO)
Raw ROS 1 and ROS 2 recordings
Overview
The dataset is structured into three splits:
Split
Description
stereo
Stereo-only (IMU stationary)
stereoinertial… See the full description on the dataset page: https://huggingface.co/datasets/OmniInstrument/SLAM_project.glasgow-extreme-lighting-slam-datasetThis dataset contains 8 stereo camera (+LiDAR not yet uploaded) robot driving sequences with ground truth pose data recorded using a Vive VR trackers and lighthouses. The structure follows that of the KITTI visual odometry sequences, with the sequences directory containing the left and right stereo images under image_2 and image_3 and the poses directory containing the ground truth poses in .txt files as flattened row-major 4x3 transformation matrices
Wild-SLAMThis repository contains data for WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments.
Paper | Project Page | Code
WildGS-SLAM accurately tracks the camera trajectory and reconstructs a 3D Gaussian map for static elements from a monocular video sequence, effectively removing dynamic components.
Datasets Used
WildGS-SLAM uses data from the following datasets:
Wild-SLAM Mocap Dataset: (Hugging Face) Download instructions are available in the github repository.… See the full description on the dataset page: https://huggingface.co/datasets/sfeges/Wild-SLAM.blur-slam-bpn-data
Blur-SLAM BPN: pipeline-intermediate data + sample results (TUM fr1_desk, full-frame)
Self-produced data for the BPN deblur + 3D Gaussian/Triangle-Splatting pipeline in
the companion code repo zhaoshiwen/blur-slam-bpn-code (see that repo's README
for the full pipeline description and reproduction steps). This repo holds the
pipeline-intermediate artifacts (EVSSM-deblurred frames, COLMAP reconstructions,
training-ready scenes, depth maps) — everything needed to go straight to… See the full description on the dataset page: https://huggingface.co/datasets/zhaoshiwen/blur-slam-bpn-data.solar-sdoThis repo contains the dataset used to train models from https://github.com/SLAMPAI/generative-models-for-highres-solar-images/ (paper: https://arxiv.org/abs/2304.07169).
The dataset is based on the SDO dataset pre-processed following the paper https://arxiv.org/abs/2304.07169/ (see Section 3).
It contains 38676 images (AIA, channel 193Å) of size 1024x1024. See the paper for more details.
mouse_brain_slam_imaging
Dataset Name: Mouse Brain SLAM (simultaneous label-free autofluorescence-multiharmonic microscopy) Images
This dataset contains multimodal microscopy imaging data, including 4-channel SLAM images, THG anisotropy data, and corresponding regional annotations/label maps.
📂 Data Hierarchy & Naming Convention
Each dataset folder follows the naming convention:
[mouse model]_[orientation]_[age]_[acquisition date]
Folder Contents
For every acquisition, the… See the full description on the dataset page: https://huggingface.co/datasets/lyyu0926/mouse_brain_slam_imaging.slam_course_data