Image Captioning on ArtEmis (test)
70.1BLEU-1GRIT
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GRITV.E. Type=R+G2022.07 | 70.1 | 40.1 | 20.9 | 11.3 | 16.8 | 33.3 | |
| GRITEmotion Grounded=No, Visual Encoder Type=R+G2022.07 | 70.1 | 40.1 | 20.9 | 11.3 | 16.8 | 33.3 | |
| GRITV.E. Type=R+G2026.05 | 70.1 | 40.1 | 20.9 | 11.3 | 16.8 | 33.3 | |
| GRITEmotion Grounded=Yes, Visual Encoder Type=R+G2022.07 | 69.3 | 39.4 | 19.2 | 11.1 | 16.5 | 33 | |
| SATV.E. Type=G2022.07 | 53.6 | 29 | 15.5 | 8.7 | 14.2 | 29.7 | |
| SATEmotion Grounded=No, Visual Encoder Type=G2022.07 | 53.6 | 29 | 15.5 | 8.7 | 14.2 | 29.7 | |
| SATV.E. Type=G2026.05 | 53.6 | 29 | 15.5 | 8.7 | 14.2 | 29.7 | |
| SATEmotion Grounded=Yes, Visual Encoder Type=G2022.07 | 52 | 28 | 14.6 | 7.9 | 13.4 | 29.4 | |
| M²Trans.Emotion Grounded=Yes, Visual Encoder Type=R2022.07 | 51.1 | 28.2 | 15.4 | 9 | 13.7 | 28.6 | |
| M²Trans.V.E. Type=R2022.07 | 50.7 | 28.2 | 15.9 | 9.5 | 13.7 | 28 | |
| M²Trans.Emotion Grounded=No, Visual Encoder Type=R2022.07 | 50.7 | 28.2 | 15.9 | 9.5 | 14 | 28 | |
| M2Trans.V.E. Type=R2026.05 | 50.7 | 28.2 | 15.9 | 9.5 | 13.7 | 28 | |
| ANPV.E. Type=G2022.07 | 39.6 | 13.4 | 4.2 | 1.4 | 8.8 | 20.2 | |
| ANPEmotion Grounded=No, Visual Encoder Type=G2022.07 | 39.6 | 13.4 | 4.2 | 1.4 | 8.8 | 20.2 | |
| ANPV.E. Type=G2026.05 | 39.6 | 13.4 | 4.2 | 1.4 | 8.8 | 20.2 | |
| NNV.E. Type=H2022.07 | 36.4 | 13.9 | 5.4 | 2.2 | 10.2 | 21 | |
| NNEmotion Grounded=No, Visual Encoder Type=H2022.07 | 36.4 | 13.9 | 5.4 | 2.2 | 10.2 | 21 | |
| NNV.E. Type=H2026.05 | 36.4 | 13.9 | 5.4 | 2.2 | 10.2 | 21 |