Skip to main content
BenchmarkComputer VisionvILSVRC 2012

ImageNet

by Deng et al. / Stanford / Princeton · open-source · Last verified 2026-03-17

ImageNet (ILSVRC) is the foundational large-scale visual recognition benchmark with 1.2 million training images across 1,000 object categories. Top-1 and Top-5 accuracy on the validation set have been the standard measure of progress in image classification for over a decade.

https://www.image-net.org
A
AGreat
Adoption: A+Quality: AFreshness: C+Citations: A+Engagement: F

Specifications

License
Custom (research only)
Pricing
open-source
Capabilities
evaluation, image-classification, transfer-learning-baseline
Integrations
Use Cases
model-evaluation, computer-vision, transfer-learning
API Available
No
Evaluated Models
vit-huge, convnext-xxlarge, dinov2, eva-giant
Metrics
top-1-accuracy, top-5-accuracy
Methodology
1,000-class image classification evaluated on the 50,000-image validation set. Models predict the top-1 and top-5 class labels; accuracy is computed as the fraction of images where the ground-truth label appears in the prediction.
Last Run
2025-12-15
Tags
image-classification, vision, top-1-accuracy, ilsvrc, foundational
Added
2026-03-17
Completeness
100%

Index Score

81.2
Adoption
97
Quality
88
Freshness
55
Citations
99
Engagement
0

Explore the full AI ecosystem on Agents as a Service