datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
sole_training_data
This is the training dataset for SOLE-R1-8B
SOLE-R1-8B is a video-language reward reasoning model for robotics. It is designed to estimate task progress from robot video frames and a natural-language task description, producing both per-timestep reasoning traces and scalar progress predictions that can be used as rewards for online robot reinforcement learning.
This dataset accompanies the paper “SOLE-R1: Video-Language Reasoning as the Sole Reward for On-Robot RL” by Philip… See the full description on the dataset page: https://huggingface.co/datasets/Philip-MIT/sole_training_data.SKIPPD
Citation
If you find SKIPP'D useful to your research, please cite:
Nie, Y., Li, X., Scott, A., Sun, Y., Venugopal, V., & Brandt, A. (2023). SKIPP’D: A SKy Images and Photovoltaic Power Generation Dataset for short-term solar forecasting. Solar Energy, 255, 171-179.
or
@article{nie2023skipp,
title={SKIPP’D: A SKy Images and Photovoltaic Power Generation Dataset for short-term solar forecasting},
author={Nie, Yuhao and Li, Xiatong and Scott, Andea and Sun, Yuchi and Venugopal… See the full description on the dataset page: https://huggingface.co/datasets/solarbench/SKIPPD.sololeveling
Bangumi Image Base of Solo Leveling
This is the image base of bangumi Solo Leveling, we detected 73 characters, 4307 images in total. The full dataset is here.
Please note that these image bases are not guaranteed to be 100% cleaned, they may be noisy actual. If you intend to manually train models using this dataset, we recommend performing necessary preprocessing on the downloaded dataset to eliminate potential noisy samples (approximately 1% probability).
Here is the characters'… See the full description on the dataset page: https://huggingface.co/datasets/BangumiBase/sololeveling.solar-pv-detection-brandenburg-dataset
Solar PV Ground Truth Dataset — Brandenburg Orthophotos
Hand-corrected ground-truth masks for photovoltaic detection on 20 cm GSD
aerial orthophotos (1 km × 1 km tiles, 4-band RGBI) of Brandenburg, Germany.
Companion to the model repository
solar-pv-segmentation-brandenburg.
Contents
Folder
Contents
gt_masks_selected/
1432 hand-corrected patch masks (256×256 px, uint8, 0=background / 1=PV), split into training/validation/testing
gt_masks_full_tiles/… See the full description on the dataset page: https://huggingface.co/datasets/Muemmel/solar-pv-detection-brandenburg-dataset.soloface2
SoloFace2: A Large-Scale Single-Face Dataset with Landmarks
SoloFace2 is a large-scale face detection and landmark regression dataset
optimized for resource-constrained deployment (TinyML, edge AI, microcontrollers).
Every image contains exactly one visible human face or none at all,
making it ideal for training lightweight single-face detectors.
Key Features
5 facial landmarks per face (eyes, nose, mouth corners)
Single-face only — no multi-face ambiguity… See the full description on the dataset page: https://huggingface.co/datasets/BidyutSaha/soloface2.stylized-commercial-dataset
Stylized Commercial Dataset
A curated image-caption dataset developed by SOLRICKS for training generative AI models focused on premium commercial imagery, advertising, product photography, lifestyle, fashion, and editorial visuals.
Trigger: stcom
Overview
The dataset contains 53 image-caption pairs built around a consistent stylized commercial aesthetic.
Visual themes include:
Product advertising
Fashion & lifestyle
Food & beverage
Mediterranean environments… See the full description on the dataset page: https://huggingface.co/datasets/SOLRICKS/stylized-commercial-dataset.robodojo-ablation-bundle
RoboDojo L4 Inspect EEF ablation videos + traces
Organized export of general_pickup / ICL probe rollouts (eval MP4 + planner traces).
Dir
Content
01_skill_bimanual_gray_ok
Early gray-block bimanual lift L4 (OK)
02_negate_xyz
XYZ negate + prior L4 Inspect layout0
03_vision_flip
flipud + fliplr
04_jitter_mean10cm
Sphere mean-10cm per-move jitter
05_camera_mask
Wrist / right-wrist / L9×2 masks
06_icl
Image+EEF and text ICL (34 tasks)
See MANIFEST.txt and… See the full description on the dataset page: https://huggingface.co/datasets/Solomonz/robodojo-ablation-bundle.solar-panel-inspectionSolidGeo
SolidGeo: Measuring Multimodal Spatial Math Reasoning in Solid Geometry
[🌐 Homepage] [💻 Github] [🤗 Huggingface Dataset]
[📊 Leaderboard ] [🔍 Visualization] [📖 Paper]
Dataset Description
SolidGeo is the first large-scale benchmark specifically designed to evaluate the performance of MLLMs on mathematical reasoning tasks in solid geometry. SolidGeo consists of 3,113 real-world K–12 and competition-level problems, each paired with visual context and annotated… See the full description on the dataset page: https://huggingface.co/datasets/SolidGeo/SolidGeo.Franka2_scoop_solid_pour_0324sol-h3-promo-assets
Sol-H3 promotional video assets
Media backing the original Sol-H3 promo Space.
spark-premium-detail-v1 contains a 47-second editorial fusion preview, poster, storyboard, reference images and media validation record. The preview uses existing video footage and original sample audio. The proposed new Ref2VA 8-step dialogue clips have not been generated.
Play the preview and read prompts.
SoleilScan
このリポジトリは、高機能フォトグラメトリーソフト「RealityCapture 1.4」を使用して、クライミングジムのソレイユの壁をスキャンしたプロジェクトのデータを含んでいます。
プロジェクトの目的
ソレイユのクライミングウォールを高精度で3Dスキャンする
RealityCapture 1.4の性能を実際のプロジェクトで検証する
スキャンデータを使って、クライミングウォールの分析や可視化を行う
使用したソフトウェアとハードウェア
RealityCapture 1.4
iphone14 pro
データ
このリポジトリには以下のデータが含まれています：
高解像度の写真（RAW形式とJPEG形式）
RealityCapture 1.4で処理された3Dモデル（OBJ形式とFBX形式）
テクスチャ画像
プロジェクトファイル
今後の展望
3Dモデルを使ったクライミングルートの分析
VRやARでのクライミングウォールの可視化
3Dプリントによる模型の作成… See the full description on the dataset page: https://huggingface.co/datasets/MakiAi/SoleilScan.SOLAR
SOLAR — ARCLE trajectories for ARC, written by an LLM
Synthesized Offline Learning dataset for Abstraction and Reasoning
Step-by-step solutions to the 400 tasks of the ARC-AGI-1 training split, on
inputs resampled by RE-ARC rather than
the original pairs, executable in ARCLE.
Every trajectory is a sequence of grid operations — select a region, recolor it,
move an object, paste, submit — recorded with the full environment state at each
step, so it can be replayed, watched… See the full description on the dataset page: https://huggingface.co/datasets/dbsgh797210/SOLAR.AutomotiveUI-Bench-4K
AutomotiveUI-Bench-4K
Dataset Overview: 998 images and 4,208 annotations focusing on interaction with in-vehicle infotainment (IVI) systems.
Key Features:
Serves as a validation benchmark for automotive UI.
Scope: Covers 15 automotive brands/OEMs, model years 2018-2025.
Image Source: Primarily photographs of IVI displays (due to screenshot limitations in most vehicles), with some direct screenshots (e.g., Android Auto).
Annotation Classes:
Test Action: Bounding box + imperative… See the full description on the dataset page: https://huggingface.co/datasets/sparks-solutions/AutomotiveUI-Bench-4K.solar-panels-taxvb
Dataset Card for solar-panels-taxvb
** The original COCO dataset is stored at dataset.tar.gz**
Dataset Summary
solar-panels-taxvb
Supported Tasks and Leaderboards
object-detection: The dataset can be used to train a model for Object Detection.
Languages
English
Dataset Structure
Data Instances
A data point comprises an image and its object annotations.
{
'image_id': 15,
'image': <PIL.JpegImagePlugin.JpegImageFile image… See the full description on the dataset page: https://huggingface.co/datasets/Francesco/solar-panels-taxvb.sole-r1-oxe-annot-1026maze-solving-for-gemma-4solar-MFI-imagesSOLAR-extras
SOLAR extras — the handcraft control, and rendered previews
Companion to dbsgh797210/SOLAR,
which holds the release itself and nothing else. Everything here is secondary to
it, which is why it is not there.
Browse the trajectories →
config
rows
tasks
what it is
handcraft
100
10
trajectories from makers written by hand, as a control
preview
400
400
one rendered picture per RE-ARC task
preview_handcraft
10
10
the same, for handcraft… See the full description on the dataset page: https://huggingface.co/datasets/dbsgh797210/SOLAR-extras.solar-rl
SolarChain-Eval RL Benchmark Data
This dataset contains the release bundle for SolarChain-Eval: A Physics-Constrained Benchmark for Trustworthy Economic Agents in Decentralized Energy Markets.
The benchmark evaluates autonomous economic governors in decentralized solar-energy markets. It combines city-level photovoltaic generation, peer-to-peer demand, market liquidity, token-burn dynamics, physics-constraint checks, baseline policies, trained RL policies, no-physics ablations… See the full description on the dataset page: https://huggingface.co/datasets/global-nomad-nexus/solar-rl.Solar-Panel-Thermal-Drone-UAV-ImagesGit
This is Thermal Images of solar power plant captured using DJI drone in India.
How to read thermal values from Image:
https://github.com/ManishSahu53/read_thermal_temperature
How to do automated hotspot detection:
https://github.com/ManishSahu53/solarHotspotAnalysis
Sample Images -
)
solar-panel-orientationAnneStokesThis is a dataset that is based off of the works of Anne Stokes, it's made using Pirsus Artstation which is trained off of SD 1.5 ...the images have been cropped, touched up, and resized to SD 1.5's base resolutions...512x768 and 768x512.
...you should be able to use kohya or dreambooth to train a lora using this.
Annotated_Medical_pill_for_Object_Detectionsolar-plants-brazil
🛰️ Solar Plants Brazil
Solar Plants Brazil is a geospatial dataset for binary semantic segmentation of photovoltaic (PV) solar power stations in satellite imagery. It consists of multi-spectral image tiles (including near-infrared) with pixel-level annotations indicating the presence of solar panels. This dataset enables training and evaluating deep learning models that automatically detect solar farm installations from overhead imagery, supporting applications in renewable energy… See the full description on the dataset page: https://huggingface.co/datasets/FederCO23/solar-plants-brazil.mathvision_with_solutionssolar-dust-data-combinedsolar-panel-fault-datasetsolarirradiancedatasetImages taken from the Sage Waggle Node's top camera and the solar irradiance values were taken from the Argonne National Laboratory
tower readings. We made sure to exclude night time photos since there is no sun and we exclusively used summer-time photos as we wanted
to stick to a seasonal model that would be able to make estimates more consistently. Furthermore we also eventually downsized the images
original 2000x2000 images to 500x500 images since the training was taking a bit too long when the images were larger.solar-panels-IGN-bdorthoA YOLO large model was trained with this dataset on IGN BdOrtho imagery.
The resulted detection has very good results which can be see on the MapRoulette project https://maproulette.org/browse/projects/62887.
The trained model is there : https://huggingface.co/Cyrille37/solar-panels-IGN-bdortho