Fine-grained Image Classification on Stanford Cars (Calibration)
67.65AccuracyDynaPrompt (Paper)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DynaPrompt (Paper)Backbone=CLIP-ViT/B16, Framework=DynaPrompt, Hard Prompt=a photo of a2026.04 | 67.65 | — | — | |
| DynaPrompt (Replication)Backbone=CLIP-ViT/B16, Framework=DynaPrompt, Hard Prompt=a photo of a2026.04 | 66.88 | 5.08 | 21 | |
| DynaPrompt+FPPBackbone=CLIP-ViT/B16, Framework=DynaPrompt, Hard Prompt=a photo of a2026.04 | 66.26 | 3.7 | 23 | |
| DynaPrompt+C-TPTBackbone=CLIP-ViT/B16, Framework=DynaPrompt, Hard Prompt=a photo of a2026.04 | 66.04 | 2.17 | 22 | |
| DynaPrompt+O-TPTBackbone=CLIP-ViT/B16, Framework=DynaPrompt, Hard Prompt=a photo of a2026.04 | 65.58 | 2.13 | 23 |