Image Captioning on MS-COCO (val)
90.9CIDErBYORn
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| BYORnLVLM=LLaVA2026.06 | 90.9 | — | — | 0.4 | |
| BYELVLM=LLaVA2026.06 | 80.6 | — | — | 54.8 | |
| SFTLVLM=LLaVA2026.06 | 74.9 | — | — | 48 | |
| ONIONLVLM=LLaVA2026.06 | 74.8 | — | — | 20.7 | |
| SFTLVLM=Qwen-VL2026.06 | 66.9 | — | — | 89.8 | |
| ONIONLVLM=Qwen-VL2026.06 | 66.9 | — | — | 45.9 | |
| BYORn-FLVLM=InternVL2026.06 | 66.8 | — | — | 2.6 | |
| BYELVLM=InternVL2026.06 | 66.5 | — | — | 64 | |
| BYORnLVLM=InternVL2026.06 | 65.1 | — | — | 2.8 | |
| SFTLVLM=InternVL2026.06 | 64.9 | — | — | 71.6 | |
| ONIONLVLM=InternVL2026.06 | 64.9 | — | — | 35.6 | |
| SFTLVLM=Otter2026.06 | 64 | — | — | 99.4 | |
| ONIONLVLM=Otter2026.06 | 63.8 | — | — | 52.2 | |
| BYELVLM=Qwen-VL2026.06 | 63.4 | — | — | 91.8 | |
| BYORn-FLVLM=Qwen-VL2026.06 | 62.9 | — | — | 2.3 | |
| BYELVLM=Otter2026.06 | 62.4 | — | — | 100 | |
| BYORn-FLVLM=LLaVA2026.06 | 60.6 | — | — | 1.6 | |
| BYORnLVLM=Qwen-VL2026.06 | 59.6 | — | — | 2.7 | |
| BYORn-FLVLM=Otter2026.06 | 58.2 | — | — | 1.2 | |
| BYORnLVLM=Otter2026.06 | 57.8 | — | — | 0.6 | |
| NICBackbone=ResNet-50, Augmentation=Baseline2019.05 | — | 61.4 | 22.9 | — | |
| NICBackbone=ResNet-50, Augmentation=Mixup2019.05 | — | 61.6 | 23.2 | — | |
| NICBackbone=ResNet-50, Augmentation=Cutout2019.05 | — | 63 | 24 | — | |
| NICBackbone=ResNet-50, Augmentation=CutMix2019.05 | — | 64.2 | 24.9 | — |