Medical Report Generation on DeepEyeNet
0.447BLEU-1DREAM
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| DREAMNumber of parameters=0.36B, GFLOPs=52026.04 | 0.447 | 0.369 | 0.328 | 0.241 | 0.576 | 0.491 | 0.91 | |
| GCS-M3VLT2026.04 | 0.43 | 0.345 | 0.319 | 0.231 | 0.559 | 0.497 | — | |
| mPLUG-Owl2 AdaptorFine-tuned=true2026.04 | 0.416 | 0.331 | 0.269 | 0.214 | 0.546 | 0.434 | 0.88 | |
| M3 Transformer2026.04 | 0.394 | 0.312 | 0.291 | 0.208 | 0.537 | 0.429 | — | |
| LLaVA-MedFine-tuned=true2026.04 | 0.386 | 0.305 | 0.282 | 0.196 | 0.482 | 0.427 | 0.88 | |
| Expert Transformer2026.04 | 0.382 | 0.291 | 0.237 | 0.186 | 0.472 | 0.413 | — | |
| VisionGPTFine-tuned=true2026.04 | 0.353 | 0.28 | 0.261 | 0.182 | 0.491 | 0.412 | 0.86 | |
| GC Transformer2026.04 | 0.297 | 0.23 | 0.214 | 0.142 | 0.462 | 0.391 | — | |
| Show, Attend & Tell2026.04 | 0.264 | 0.209 | 0.197 | 0.129 | 0.416 | 0.39 | — | |
| Non-local Attention2026.04 | 0.23 | 0.15 | 0.094 | 0.053 | 0.37 | 0.291 | — | |
| Deep Context Model2026.04 | 0.219 | 0.134 | 0.074 | 0.035 | 0.398 | 0.252 | — | |
| Show and Tell2026.04 | 0.207 | 0.124 | 0.094 | 0.039 | 0.397 | 0.339 | — | |
| Contextualized GPT2026.04 | 0.203 | 0.142 | 0.1 | 0.073 | 0.389 | 0.211 | — | |
| DeepOpth2026.04 | 0.184 | 0.114 | 0.068 | 0.032 | 0.361 | 0.232 | — |