Image Captioning on COCO (test) (ROUGE Metrics)
0.518ROUGE-1VIST2
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| VIST2Scale=8B2026.01 | 0.518 | 0.369 | 0.501 | 0.4867 | |
| VIST2Scale=4B2026.01 | 0.489 | 0.342 | 0.473 | 0.4881 | |
| VIST2Scale=0.6B2026.01 | 0.462 | 0.318 | 0.451 | 0.4901 | |
| GLM-4.5VScale=9B2026.01 | 0.335 | 0.295 | 0.328 | 0.4863 | |
| MiMo-VLScale=7B2026.01 | 0.327 | 0.258 | 0.318 | 0.4923 | |
| InternVL-3.5Scale=8B2026.01 | 0.322 | 0.251 | 0.217 | 0.4863 | |
| Qwen3-VLScale=8B2026.01 | 0.309 | 0.252 | 0.307 | 0.484 | |
| Minicpm-4.5VScale=3B2026.01 | 0.308 | 0.27 | 0.302 | 0.3916 | |
| Ovis2.5Scale=2.2B2026.01 | 0.221 | 0.168 | 0.218 | 0.3921 |