The ELEVATER benchmark is a collection of resources for training, evaluating, and analyzing language-image models on image classification and object detection. ELEVATER consists of:
23 PAPERS • 2 BENCHMARKS