Open-ended Visual Question Answering on LLS48-VQA
0.5245BLEU-1MIS-DINOv2
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MIS-DINOv2Backbone=ViT-L, Pre-training Type=Self-supervised learning, Evaluation Protocol=Fine-tuned2026.02 | 0.5245 | 0.4336 | 0.3725 | 0.327 | 0.4529 | 0.2727 | |
| ZENBackbone=ViT-B, Pre-training Type=Existing pretrained models, Evaluation Protocol=Fine-tuned2026.02 | 0.5244 | 0.434 | 0.3734 | 0.3282 | 0.4519 | 0.2726 | |
| SurgeNetBackbone=CAFormer18, Pre-training Type=Existing pretrained models, Evaluation Protocol=Fine-tuned2026.02 | 0.5186 | 0.4284 | 0.3678 | 0.3228 | 0.4445 | 0.2672 | |
| MIS-DINOv2Backbone=ViT-L, Backbone status=frozen2026.02 | 0.5175 | 0.4238 | 0.3618 | 0.3159 | 0.4419 | 0.2639 | |
| MIS-DINOv2Backbone=ViT-B, Pre-training Type=Self-supervised learning, Evaluation Protocol=Fine-tuned2026.02 | 0.5173 | 0.425 | 0.3629 | 0.3167 | 0.4432 | 0.2657 | |
| ZENBackbone=ViT-B, Backbone status=frozen2026.02 | 0.5148 | 0.4227 | 0.3615 | 0.3162 | 0.4402 | 0.2648 | |
| MIS-MAEBackbone=ViT-B, Pre-training Type=Self-supervised learning, Evaluation Protocol=Fine-tuned2026.02 | 0.5131 | 0.4212 | 0.3601 | 0.3149 | 0.4391 | 0.2616 | |
| MIS-DINOBackbone=ViT-B, Pre-training Type=Self-supervised learning, Evaluation Protocol=Fine-tuned2026.02 | 0.5116 | 0.4197 | 0.3582 | 0.3126 | 0.4384 | 0.2614 | |
| MIS-DINOv2Backbone=ViT-B, Backbone status=frozen2026.02 | 0.511 | 0.4167 | 0.3542 | 0.3081 | 0.4347 | 0.2596 | |
| MIS-DINOBackbone=ViT-B, Backbone status=frozen2026.02 | 0.5062 | 0.4138 | 0.3524 | 0.3071 | 0.4345 | 0.2594 | |
| SurgeNetBackbone=CAFormer18, Backbone status=frozen2026.02 | 0.503 | 0.4097 | 0.3479 | 0.3023 | 0.4285 | 0.255 | |
| EndoViTBackbone=ViT-B, Pre-training Type=Existing pretrained models, Evaluation Protocol=Fine-tuned2026.02 | 0.5009 | 0.4086 | 0.3476 | 0.3028 | 0.4275 | 0.2537 | |
| ImageNetBackbone=ViT-B, Pre-training Type=Existing pretrained models, Evaluation Protocol=Fine-tuned2026.02 | 0.5006 | 0.4066 | 0.3443 | 0.2985 | 0.4237 | 0.2505 | |
| MIS-MoCoV3Backbone=ViT-B, Backbone status=frozen2026.02 | 0.5002 | 0.4079 | 0.3473 | 0.3027 | 0.4244 | 0.2527 | |
| PeskaVLPBackbone=ResNet50, Pre-training Type=Existing pretrained models, Evaluation Protocol=Fine-tuned2026.02 | 0.5 | 0.4085 | 0.3479 | 0.3034 | 0.43 | 0.2558 | |
| MIS-MoCoV3Backbone=ViT-B, Pre-training Type=Self-supervised learning, Evaluation Protocol=Fine-tuned2026.02 | 0.498 | 0.406 | 0.3455 | 0.301 | 0.4268 | 0.2558 | |
| MIS-MSNBackbone=ViT-B, Pre-training Type=Self-supervised learning, Evaluation Protocol=Fine-tuned2026.02 | 0.4926 | 0.4019 | 0.3423 | 0.2983 | 0.4197 | 0.2515 | |
| EndoFMBackbone=ViT-B, Pre-training Type=Existing pretrained models, Evaluation Protocol=Fine-tuned2026.02 | 0.4917 | 0.3958 | 0.3328 | 0.2869 | 0.4152 | 0.2448 | |
| MIS-MAEBackbone=ViT-B, Backbone status=frozen2026.02 | 0.4873 | 0.3937 | 0.3323 | 0.2873 | 0.4125 | 0.2443 | |
| ImageNetBackbone=ViT-B, Backbone status=frozen2026.02 | 0.4846 | 0.3882 | 0.3248 | 0.2787 | 0.4066 | 0.2401 | |
| EndoViTBackbone=ViT-B, Backbone status=frozen2026.02 | 0.4824 | 0.3857 | 0.3226 | 0.277 | 0.4077 | 0.2403 | |
| EndoFMBackbone=ViT-B, Backbone status=frozen2026.02 | 0.482 | 0.3883 | 0.3267 | 0.2819 | 0.4079 | 0.2399 | |
| PeskaVLPBackbone=ResNet50, Backbone status=frozen2026.02 | 0.4817 | 0.3869 | 0.3256 | 0.2808 | 0.4068 | 0.2417 | |
| GSViTBackbone=EfficientViT, Pre-training Type=Existing pretrained models, Evaluation Protocol=Fine-tuned2026.02 | 0.4167 | 0.3102 | 0.2468 | 0.2022 | 0.3363 | 0.1926 | |
| MIS-MSNBackbone=ViT-B, Backbone status=frozen2026.02 | 0.3992 | 0.2924 | 0.2288 | 0.1842 | 0.3272 | 0.1933 | |
| GSViTBackbone=EfficientViT, Backbone status=frozen2026.02 | 0.3838 | 0.2862 | 0.2261 | 0.1843 | 0.3292 | 0.1861 |