weiaicunzai/awesome-image-classification

Organizes landmark papers chronologically from 2014 onward with ImageNet performance benchmarks (top-1/top-5 accuracy), enabling researchers to track architectural evolution from VGG through Vision Transformers. Includes links to official implementations and accompanying PyTorch reference code, targeting practitioners who need both theoretical foundations and reproducible baselines for CNN and transformer-based classifiers.

3,053 stars. No commits in the last 6 months.