| Backbone | Decoder | Detector | Box | Ratio | Beam | K | Seed | SPICE | METEOR | CIDEr | CHAIRs | CHAIRi | BLEU | Log CIDEr | Num of Samples | Max Tokens |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| minigpt4 | halc | dino | 0.45 | 0.2 | 1 | 4 | 0 | 18.4 | 21.9 | 0.6 | 16.0 | 7.1 | 0.154 | 11.76 | 64 | 50 |
| minigpt4 | halc | dino | 0.45 | 0.4 | 1 | 4 | 0 | 18.4 | 21.9 | 4.5 | 18.0 | 7.8 | 0.160 | 11.91 | 64 | 50 |
| minigpt4 | halc | dino | 0.45 | 0.6 | 1 | 4 | 0 | 18.4 | 21.5 | 4.5 | 20.0 | 7.8 | 0.160 | 11.87 | 64 | 50 |
| minigpt4 | halc | dino | 0.45 | 0.8 | 1 | 4 | 0 | 16.9 | 21.2 | 0.1 | 24.0 | 10.2 | 0.147 | 11.82 | 64 | 50 |
| minigpt4 | halc | dino | 0.45 | 1.0 | 1 | 4 | 0 | 15.6 | 19.8 | 0.0 | 23.1 | 12.2 | 0.139 | 12.31 | 64 | 50 |
| minigpt4 | halc | dino | 0.45 | 0.6 | 2 | 4 | 0 | 18.3 | 22.1 | 4.9 | 24.0 | 9.6 | 0.165 | 12.43 | 64 | 50 |
| minigpt4 | halc | dino | 0.45 | 0.6 | 3 | 4 | 0 | 18.1 | 22.1 | 3.0 | 26.3 | 7.7 | 0.168 | 11.95 | 64 | 50 |
| minigpt4 | halc | dino | 0.45 | 0.6 | 5 | 4 | 0 | 12.5 | 19.5 | 0.7 | 28.6 | 9.5 | 0.135 | 13.77 | 64 | 50 |
| minigpt4 | halc | dino | 0.45 | 0.6 | 8 | 4 | 0 | 21.9 | 22.7 | 0.0 | 25.0 | 15.4 | 0.140 | 12.98 | 64 | 50 |