PETR: Prompt Ensembling with Training-free Routing for Vision-Language Models
Open paper
Benchmark
Model
Rank
Results
prompt-engineering-on-imagenet-a
PETR
–
Top-1 accuracy %: 71.52
prompt-engineering-on-imagenet-r
PETR
–
Top-1 accuracy %: 87.77
prompt-engineering-on-imagenet-v2
PETR
–
Top-1 accuracy %: 70.39
Rank counts only results with a code link.