datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
hagrid-sample-30k-384pThis dataset contains 31,833 images from HaGRID (HAnd Gesture Recognition Image Dataset) downscaled to 384p. The original dataset is 716GB and contains 552,992 1080p images. I created this sample for a tutorial so readers can use the dataset in the free tiers of Google Colab and Kaggle Notebooks.
Original Authors:
Alexander Kapitanov
Andrey Makhlyarchuk
Karina Kvanchiani
Original Dataset Links
GitHub
Kaggle Datasets Page
Object Classes
['call'… See the full description on the dataset page: https://huggingface.co/datasets/cj-mills/hagrid-sample-30k-384p.hagrid-sample-500k-384pThis dataset contains 509,323 images from HaGRID (HAnd Gesture Recognition Image Dataset) downscaled to 384p. The original dataset is 716GB and contains 552,992 1080p images. I created this sample for a tutorial so readers can use the dataset in the free tiers of Google Colab and Kaggle Notebooks.
Original Authors:
Alexander Kapitanov
Andrey Makhlyarchuk
Karina Kvanchiani
Original Dataset Links
GitHub
Kaggle Datasets Page
Object Classes
['call'… See the full description on the dataset page: https://huggingface.co/datasets/cj-mills/hagrid-sample-500k-384p.hagrid-30k-384p-select-gesturesRefined version of https://huggingface.co/datasets/cj-mills/hagrid-sample-500k-384p used for Conjure
tongue-images-384-segmented-augmentedtongue-images-384inet1k_compressed_384fashion-product-images-small-384x512tongue-images-384-augmentedmtg-scryfall-cropped-art-embeddings-siglip-so400m-patch14-384tongue-images-384-segmentedgelbooru-webp-subset-384px-dedupSubset of gelbooru_full dowscaled to 384px
Pruned using ViT-SO400M-16-SigLIP2-384 cosine similarity threshold 0.95
hagrid-sample-120k-384pThis dataset contains 127,331 images from HaGRID (HAnd Gesture Recognition Image Dataset) downscaled to 384p. The original dataset is 716GB and contains 552,992 1080p images. I created this sample for a tutorial so readers can use the dataset in the free tiers of Google Colab and Kaggle Notebooks.
Original Authors:
Alexander Kapitanov
Andrey Makhlyarchuk
Karina Kvanchiani
Original Dataset Links
GitHub
Kaggle Datasets Page
Object Classes
['call'… See the full description on the dataset page: https://huggingface.co/datasets/cj-mills/hagrid-sample-120k-384p.uavdt_384
UAVDT (384)
Vehicle detection from UAV (drone) aerial imagery.
This is a 384×384 resized version of the UAVDT dataset, with annotations in COCO format (images, annotations, categories).
Contents
images.zip — all images, resized to 384×384 (JPEG).
annotations.zip — COCO-format JSON annotation files.
Extract
unzip images.zip
unzip annotations.zip
Splits
Split
Images
Annotations
Categories
train
1,266
35,007
1
val
271
7… See the full description on the dataset page: https://huggingface.co/datasets/hieupth/uavdt_384.mtg-scryfall-cropped-art-embeddings-open-clip-ViT-SO400M-14-SigLIP-384hagrid-sample-250k-384pThis dataset contains 254,661 images from HaGRID (HAnd Gesture Recognition Image Dataset) downscaled to 384p. The original dataset is 716GB and contains 552,992 1080p images. I created this sample for a tutorial so readers can use the dataset in the free tiers of Google Colab and Kaggle Notebooks.
Original Authors:
Alexander Kapitanov
Andrey Makhlyarchuk
Karina Kvanchiani
Original Dataset Links
GitHub
Kaggle Datasets Page
Object Classes
['call'… See the full description on the dataset page: https://huggingface.co/datasets/cj-mills/hagrid-sample-250k-384p.LaTeX_OCR_384x384scut_head_384
SCUT-HEAD (384)
Head detection dataset.
This is a 384×384 resized version of the SCUT-HEAD dataset, with annotations in COCO format (images, annotations, categories).
Contents
images.zip — all images, resized to 384×384 (JPEG).
annotations.zip — COCO-format JSON annotation files.
Extract
unzip images.zip
unzip annotations.zip
Splits
Split
Images
Annotations
Categories
train
2,543
61,943
1
val
862
22,642
1… See the full description on the dataset page: https://huggingface.co/datasets/hieupth/scut_head_384.query_image_anchor_positive_large_384recap-datacomp-384-1Mdark_face_384
DarkFace (384)
Face detection in extremely low-light (dark) images.
This is a 384×384 resized version of the DarkFace dataset, with annotations in COCO format (images, annotations, categories).
Contents
images.zip — all images, resized to 384×384 (JPEG).
annotations.zip — COCO-format JSON annotation files.
Extract
unzip images.zip
unzip annotations.zip
Splits
Split
Images
Annotations
Categories
train
5,400
45,273
1… See the full description on the dataset page: https://huggingface.co/datasets/hieupth/dark_face_384.laion2b-mixed-384px-human-dedupImages from laion2b-mixed-1024px-human dowscaled to 384px
Pruned using ViT-SO400M-16-SigLIP2-384 cosine similarity threshold 0.95
tongue-images-384tongue-images-384-augmentedfisheye8k_384
FishEye8K (384)
Object detection in fisheye (360 deg. camera) images.
This is a 384×384 resized version of the FishEye8K dataset, with annotations in COCO format (images, annotations, categories).
Contents
images.zip — all images, resized to 384×384 (JPEG).
annotations.zip — COCO-format JSON annotation files.
Extract
unzip images.zip
unzip annotations.zip
Splits
Split
Images
Annotations
Categories
train
5,255
112,077
5… See the full description on the dataset page: https://huggingface.co/datasets/hieupth/fisheye8k_384.cache_detailed_C384dataset_side384_mlt_eddataset_side384dataset_side384_mlt500_empty_staged_384px
