datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
bakkhali-river-high-low-tide
Bakkhali River — High Tide vs Low Tide, Bangladesh
517 photographs of the Bakkhali River near Cox's Bazar, Bangladesh, documenting the same general stretch of river at high tide (264 images) and low tide (253 images). Captured across 10 separate sessions between 2 July and 15 August 2026.
This is not a frame-by-frame matched pair set — sessions were shot on different dates and the camera position varies within each session — but high- and low-tide frames come from the same short… See the full description on the dataset page: https://huggingface.co/datasets/golamrob/bakkhali-river-high-low-tide.srt-cxr14-frozen-probe
Frozen general-purpose features beat a fine-tuned baseline on ChestX-ray14
A linear probe on frozen google/gemma-4-31B-it hidden states, on all 112,120
images of NIH ChestX-ray14, using the official test_list.txt. No
fine-tuning, no radiology pretraining, no augmentation. One Linear(d, 14)
under BCE, which is fourteen logistic regressions.
ChestX-ray14, official split
mean AUROC
method
Wang et al. 2017 (dataset authors)
0.7451
ResNet-50, fine-tuned end to end
this… See the full description on the dataset page: https://huggingface.co/datasets/RiverRider/srt-cxr14-frozen-probe.Pixiv-2.6M
Pixiv 2.6M specific
This dataset contain 2.6M images from Pixiv.
Introduction
This dataset aims to add some concepts which is lacked or banned in danbooru.
Images Content
The images is collected with search cralwer made with selenium and python
I use some tags which may lack in danbooru for example "scenery" or some tags for "explicit contents" to do the search and collect the url of result than download all the image.
Since there are lot of posts with multiple… See the full description on the dataset page: https://huggingface.co/datasets/Amber-River/Pixiv-2.6M.riverside-sunset-golden-hour
Riverside Sunset at Golden Hour — Bakkhali River, Bangladesh
100 photographs of sunset over the Bakkhali River at low tide, near Cox's Bazar, Bangladesh. All frames were captured in a single evening session on 13 July 2026 (17:59–18:24 local time, UTC+6) during golden hour.
Contents
100 JPG images, 2000×1333 px, filenames riverside-sunset-golden-hour-004.JPG through -108.JPG. Numbers are not fully sequential — 100 of 110 frames captured during the session were… See the full description on the dataset page: https://huggingface.co/datasets/golamrob/riverside-sunset-golden-hour.pitt-rivers-igbo-collection
Pitt Rivers Museum (Igbo Collection) Dataset
This dataset is a curated collection of archival records from the Pitt Rivers Museum, specifically focusing on the Igbo people of Nigeria.
💾 Dataset Structure
The dataset is organized as a standard ImageFolder with metadata stored in data.jsonl.
/images: Contains high-resolution JPEG images sourced directly from the IIIF server.
data.jsonl: Contains rich metadata (Accession Number, Photographer, Dimensions, Acquisition Info).… See the full description on the dataset page: https://huggingface.co/datasets/nwokikeonyeka/pitt-rivers-igbo-collection.doodles-datasetA complete doodles dataset!
pitt-rivers-wellcome-collection
Dataset Card for Pitt Rivers Museum - Sir Henry Wellcome Collection
Dataset Summary
This dataset contains approximately 14,000 object records from the Pitt Rivers Museum (University of Oxford) that are associated with the Sir Henry Wellcome collection. Following Wellcome's death, his vast historical medical and ethnographic museum was dispersed to institutions worldwide; the Pitt Rivers Museum was a major recipient of this dispersal.
The dataset includes deeply… See the full description on the dataset page: https://huggingface.co/datasets/JCatesWellcome/pitt-rivers-wellcome-collection.