Image Captioning on Flickr30k Adversarial (epsilon_a = 4/255)
37.1CIDErCLIP
Evaluation Results
| Method | Links | |
|---|---|---|
| CLIPET3 Defense=true2026.03 | 37.1 | |
| FAREAdversarial Training Strength (epsilon_t)=4/255, ET3 Defense=true2026.03 | 33.5 | |
| TeCoAAdversarial Training Strength (epsilon_t)=2/255, ET3 Defense=true2026.03 | 32.1 | |
| FAREAdversarial Training Strength (epsilon_t)=2/255, ET3 Defense=true2026.03 | 31.9 | |
| TeCoAAdversarial Training Strength (epsilon_t)=4/255, ET3 Defense=true2026.03 | 30.2 | |
| FAREAdversarial Training Strength (epsilon_t)=4/255, ET3 Defense=false2026.03 | 23.1 | |
| TeCoAAdversarial Training Strength (epsilon_t)=4/255, ET3 Defense=false2026.03 | 19.5 | |
| FAREAdversarial Training Strength (epsilon_t)=2/255, ET3 Defense=false2026.03 | 17.5 | |
| TeCoAAdversarial Training Strength (epsilon_t)=2/255, ET3 Defense=false2026.03 | 14.8 | |
| CLIPET3 Defense=false2026.03 | 1.1 |