Adaptive Parameter Selection for Tuning Vision-Language Models

Benchmark Model Rank Results
few-shot-image-classification-on-dtd-16-shotCLIP-AST (ViT-L/14)–Accuracy: 78.6
few-shot-image-classification-on-eurosat-16-shotCLIP-AST (ViT-L/14)–Accuracy: 94.45
image-classification-on-flowers-102CLIP-AST (ViT-L/14)–Accuracy: 99.06