datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
somethings_missing_here
Something's Missing Here
A dataset of viewpoint-aligned before/after image pairs for training and
evaluating missing-object detection: each pair shows the same scene from
(nearly) the same viewpoint, once with one or more objects present and once
after they were removed, with a label naming what went missing.
The dataset originated from pairs derived from the
simkoc/Remove360 dataset
(training/Remove360_based/), extended with original photographs collected
by the dataset author… See the full description on the dataset page: https://huggingface.co/datasets/Shay85Gil/somethings_missing_here.SoMeData
README for SoMeData
This is the official dataset of the ACL 2024 paper SoMeLVLM: A Large Vision Language Model for Social Media Processing.
Important Notice
Before using data in SoMeData, you agree that the use of the data is only restricted to research or education purposes and that all copyright and license restrictions associated with the dataset/code will be followed.
If you find our dataset useful, we would greatly appreciate it if you could consider citing our… See the full description on the dataset page: https://huggingface.co/datasets/Lishi0905/SoMeData.GTSRB_224x224_balanced
Balanced GTSRB (224x224)
This is a balanced GTSRB dataset containing 43 classes, with 1,000 training samples per class and the same number of test samples as in the original dataset.All images have been resized to 224×224 using interpolation and padding, maintaining aspect ratio.
For classes with fewer than 1,000 training samples, data augmentation was used to supplement the dataset (Note: without any flip transforms, thanks to this post).
For details on how the dataset was… See the full description on the dataset page: https://huggingface.co/datasets/SomeBottle/GTSRB_224x224_balanced.imagenet_50_224x224
ImageNet 50 (224x224)
This is a randomly selected subset of ImageNet containing 50 categories, with 1,000 training samples and 100 validation samples per class.All images have been resized to 224×224 using the CenterCrop method.
Each class has 50 validation samples taken from the original training set, which do not overlap with the 1,000 selected training samples.
For details on how the dataset was generated, you can refer to the flow.ipynb notebook in the repository.
The… See the full description on the dataset page: https://huggingface.co/datasets/SomeBottle/imagenet_50_224x224.