Prompt Engineering On Imagenet 21K

Accuracy

Results

Performance results of various models on this benchmark

Comparison Table

Model Name	Accuracy
visual-prompt-tuning	24.8
prompt-pre-training-with-twenty-thousand-1	25.3