datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
3dfront_render_views3dfront-render-viewshabitat-views-20kai2thor-random-views-20kViewSpatial-Bench
ViewSpatial-Bench: Evaluating Multi-perspective Spatial Localization in Vision-Language Models
Dataset Description
We introduce ViewSpatial-Bench, a comprehensive benchmark with over 5,700 question-answer pairs across 1,000+ 3D scenes from ScanNet and MS-COCO validation sets. This benchmark evaluates VLMs' spatial localization capabilities from multiple perspectives, specifically testing both egocentric (camera) and allocentric (human subject) viewpoints across… See the full description on the dataset page: https://huggingface.co/datasets/lidingm/ViewSpatial-Bench.Awesome_Spatial_VQA_Benchmarks_ViewSpatial-BenchViewSpatial_lmmsevalasset-alignment-reference-views
Asset Alignment Reference Views
Companion dataset for the paper "Rigid 3D Object Alignment: Optimization vs. Feed-Forward Prediction".
Multi-view renderings of correctly assembled source–target pairs: each row
shows one asset already aligned onto its target object, rendered from 12
orbiting viewpoints with RGB and depth.
Where asset-alignment-pairs-905k
shows the asset misaligned and supplies the transformation that fixes it, this
dataset shows the ground-truth assembled result.… See the full description on the dataset page: https://huggingface.co/datasets/macpaw-research/asset-alignment-reference-views.ai2thor-random-views-20k-3obj-filteredego4d-random-views-20k
Ego4D Random Views Dataset
This dataset contains 20,000 random view frames sampled from the Ego4D dataset using a high-performance multi-process generation system.
Dataset Overview
Total Images: 20,000 high-quality frames
Image Format: PNG (1024×1024 resolution)
Source: Ego4D v2 dataset (52,665+ video files)
Sampling Method: Multi-process random sampling with maximum diversity
Generation Time: 797.57 seconds (~13 minutes)
Generation Speed: 25.08 frames/second… See the full description on the dataset page: https://huggingface.co/datasets/weikaih/ego4d-random-views-20k.deepfake-frame-views-v3
Precomputed Deepfake Frame Views
This repo stores sampled frame-view .npz shards for fast training. Extract shards locally and pass --precomputed-manifest manifest.jsonl --precomputed-root <extract_dir> to the trainer.
viewsuite
Contents
Every archive extracts to a directory of the same name (viewagent15k_<scene source>[_<renderer>][_<variant>]/ for task sets; viewagent150k_scannet_open3d for the full ScanNet set; viewagent_scannet and viewagent_mindcube for source data).
Task archives hold {path_to_view,view_to_path,interactive_view_planning}_{train,dev|eval,test}.jsonl
plus the rendered views they reference. SHA256SUMS lists the checksum of every archive.
ScanNet scenes
File… See the full description on the dataset page: https://huggingface.co/datasets/MLL-Lab/viewsuite.record-up-views-1This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so101_follower",
"total_episodes": 10,
"total_frames": 3763,
"total_tasks": 1,
"total_videos": 10,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:10"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/jayyucippg/record-up-views-1.auxiliary-views-knowledge-acquisition
Auxiliary Views Knowledge Acquisition
This repository contains the cleaned source documents and evaluation
probes used in Knowledge Acquisition During Pre-training? Large Language Models
Learn Better With Auxiliary Views (arXiv:2609.04180).
News
August 21, 2026: Our paper was accepted to Findings of EMNLP 2026.
Configurations
Configuration
Split
Rows
documents
train
30
factual_cloze
test
6,435
factual_mcqa_5shot
test
4,515… See the full description on the dataset page: https://huggingface.co/datasets/jiosephlee/auxiliary-views-knowledge-acquisition.deepfake-celebvhq-frame-views-v1
CelebV-HQ Frame Views
This dataset contains precomputed frame-view .npz files generated from CelebV-HQ metadata. It is intended for training the real class / hard negatives in the temporal deepfake MoE pipeline.
Labels use 0 = real, 1 = fake. Every record in this repo is labeled real.
Format
shards/*.tar: contains views/*.npz
manifest_parts/*.jsonl: manifest rows for each uploaded shard
worker_manifests/*.jsonl: per-worker aggregate manifests
metadata/*.json:… See the full description on the dataset page: https://huggingface.co/datasets/gonnerthetooner/deepfake-celebvhq-frame-views-v1.View-Spatial-Benchmv_viewsdeepfake-real-frame-views-hardneg-v1
Real YouTube Hard-Negative Frame Views
Prepared frame-view shards for real YouTube/social-media videos. Rows are labeled 0 = real and are intended as hard negatives for a deepfake detector.
Only use this dataset if you have rights/permission for the source videos or derived artifacts.
{
"candidate_count": 494,
"clip_seconds": 180,
"created_at": "2026-06-14T05:23:49Z",
"dataset_name": "youtube_real_hardneg",
"decode_failed": 0,
"download_failed": 1,
"finished_at":… See the full description on the dataset page: https://huggingface.co/datasets/gonnerthetooner/deepfake-real-frame-views-hardneg-v1.record-side-views-updown-lThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so101_follower",
"total_episodes": 10,
"total_frames": 2839,
"total_tasks": 1,
"total_videos": 10,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:10"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/jayyucippg/record-side-views-updown-l.robocasa-30demos-5chosen-tasks_params_virtual_views_splattedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 104,
"total_frames": 26745,
"total_tasks": 45,
"total_videos": 1456,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:104"},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/tt1225/robocasa-30demos-5chosen-tasks_params_virtual_views_splatted.UniCAD-Synthetic-Views
UniCAD Synthetic Multi-View CAD Images
This dataset contains the synthetic multi-view image data created for training
UniCAD. It provides four 1024 x 1024 RGB views of CAD objects in two visual
styles:
sketch: hand-drawn-style images.
realistic: photorealistic-style images.
The release contains only the UniCAD team's generated images and lightweight
provenance metadata. It does not redistribute Text2CAD descriptions,
Text2CAD CadQuery programs, or the original DeepCAD renders.… See the full description on the dataset page: https://huggingface.co/datasets/Chen8566/UniCAD-Synthetic-Views.robocasa-100demos-5chosen-tasks_params_virtual_views_splattedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 364,
"total_frames": 92882,
"total_tasks": 96,
"total_videos": 5096,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:364"},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/tt1225/robocasa-100demos-5chosen-tasks_params_virtual_views_splatted.ViewSpatial-Bench-vlmevalkoch_kichen_2_viewsThis dataset was created using LeRobot.
deepfake-frame-views-balanced-fusion-v1
Balanced Deepfake Frame Views
Balanced .npz frame-view shards for fusion retraining. Records are copied from source precomputed frame-view repos without decoding videos.
target real ratio: 0.5
records per shard: 512
planned records: 8192
labels: {"0": 4096, "1": 4096}
Use this repo as PRECOMPUTED_VIEWS_REPO when retraining fusion.
africa-botswana-botswana-views-conflict-forecasts-4aba2a12
Botswana - VIEWS conflict forecasts | Africa (Botswana official open data)
36 rows - 1 Africa country - 2026-2027 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Botswana as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Botswana - VIEWS conflict forecasts
Publisher:… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-botswana-botswana-views-conflict-forecasts-4aba2a12.africa-namibia-namibia-views-conflict-forecasts-1b2e686f
Namibia - VIEWS conflict forecasts | Africa (Namibia official open data)
36 rows - 1 Africa country - 2026-2027 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Namibia as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Namibia - VIEWS conflict forecasts
Publisher:… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-namibia-namibia-views-conflict-forecasts-1b2e686f.africa-botswana-botswana-views-conflict-forecasts-d48b1a3b
Botswana - VIEWS conflict forecasts | Africa (Botswana official open data)
1,000 rows - 1 Africa country - 2026 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Botswana as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Botswana - VIEWS conflict forecasts
Publisher:… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-botswana-botswana-views-conflict-forecasts-d48b1a3b.africa-guinea-bissau-guinea-bissau-views-conflict-forecasts-32aeb43f
Guinea-Bissau - VIEWS conflict forecasts | Africa (Guinea-Bissau official open data)
36 rows - 1 Africa country - 2026-2027 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Guinea-Bissau as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Guinea-Bissau - VIEWS conflict… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-guinea-bissau-guinea-bissau-views-conflict-forecasts-32aeb43f.africa-algeria-algeria-views-conflict-forecasts-3308a8b4
Algeria - VIEWS conflict forecasts | Africa (Algeria official open data)
36 rows - 1 Africa country - 2026-2027 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Algeria as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Algeria - VIEWS conflict forecasts
Publisher:… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-algeria-algeria-views-conflict-forecasts-3308a8b4.