Datasets:

taesiri
/

imagenet-hard

Tasks:

Image Classification

Modalities:

Formats:

Languages:

Size:

ArXiv:

Tags:

Out Of Distribution

Libraries:

License:

Dataset card Viewer Files Files and versions Community

taesiri commited on Apr 15, 2023

Commit

331b128

•

1 Parent(s): a500be3

Update README.md

Files changed (1) hide show

README.md +4 -0

README.md CHANGED Viewed

@@ -40,6 +40,10 @@ The **ImageNet-Hard** is a new benchmark that comprises 13,630 images, collected
 This dataset is challenging to state-of-the-art vision models, as merely zooming in often fails to enhance their ability to classify images correctly. Consequently, even the most advanced models, such as `CLIP-ViT-L/14@336px`, struggle to perform well on this dataset, achieving only `2.02%` accuracy.
 ### Classifiers Performance

 This dataset is challenging to state-of-the-art vision models, as merely zooming in often fails to enhance their ability to classify images correctly. Consequently, even the most advanced models, such as `CLIP-ViT-L/14@336px`, struggle to perform well on this dataset, achieving only `2.02%` accuracy.
+### Dataset Distribution
+![Dataset Distribution](https://taesiri.github.io/ZoomIsAllYouNeed/static/svg/imagenet_hard_distribution.svg)
 ### Classifiers Performance