CIFAR-100

CIFAR-100 is a dataset that contains 60,000 32x32 color images across 100 different classes, used to evaluate the performance of machine learning models in image classification tasks.

Papers using CIFAR-100 (3)

Explaining CLIP Zero-shot Predictions Through Concepts2026Preserving Cross-Modal Consistency for CLIP-based Class-Incremental Learning2025Topological Alignment Of Shared Vision-language Embedding Space2025