AI Challenger is a large-scale image understanding dataset that can serve as an effective benchmark for evaluating and improving various algorithms. In addition, researchers can also use this dataset for pre-training models. The dataset includes: • 3 subsets: Human Keypoint Detection (HKD), Large Scale Attribute Dataset (LAD) and Image Chinese Captioning (ICC). • Rich annotations that bridge the semantic gap between low-level images and high-level concepts. • It is an effective benchmark to evaluate and improve different algorithms.