awesome-computer-vision-models

Why we included this project

This reference collects the numbers that matter when you are choosing a backbone for an image task: parameter counts, FLOPs, and the top-1/top-5 error rates reported in the papers, all laid out in tables. The classification section covers the usual families, AlexNet, VGG, ResNet, DenseNet, Inception, MobileNet, EfficientNet and others, with separate tables for segmentation and detection models. What makes it useful is the side-by-side comparison, which shows at a glance how a lightweight mobile model trades accuracy for compute against a heavier SENet or NASNet variant. If you are scoping a new vision project or writing a literature review, it is a fast way to shortlist candidate architectures before you read the original papers.