Region Captioning on Visual Genome
19.7METEORGLaMM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GLaMMEvaluation protocol=Fine-Tuned (FT)2023.11 | 19.7 | 180.5 | |
| GLaMMRegion Feat. Extractor=true2025.12 | 19.7 | 180.5 | |
| GLaMMtask-specific finetuning=true2024.04 | 19 | 163.9 | |
| Groma2024.04 | 19 | 158.4 | |
| GromaRegion Feat. Extractor=true2025.12 | 19 | 158.4 | |
| GETok-SFTRegion Feat. Extractor=false2025.12 | 19 | 165.9 | |
| ASM-FTtype=fine-tuned2024.02 | 18.3 | 148.7 | |
| ASMv2-13Bparameters=13B2024.02 | 17.9 | 153.5 | |
| GPT4RoI-13Bparameters=13B2024.02 | 17.6 | 146.8 | |
| GPT4RoI2024.04 | 17.4 | 145.2 | |
| GPT4RoI2023.11 | 17.4 | 145.2 | |
| GPT4RoI-7Bparameters=7B2024.02 | 17.4 | 145.2 | |
| GPT4RoIRegion Feat. Extractor=true2025.12 | 17.4 | 145.2 | |
| GRIT2024.04 | 17.1 | 142 | |
| GRiT2023.11 | 17.1 | 142 | |
| GRIT2024.02 | 17.1 | 142 | |
| GRITRegion Feat. Extractor=true2025.12 | 17.1 | 142 | |
| GLaMMEvaluation protocol=Zero-Shot (ZS)2023.11 | 17 | 127 |