Machine Translation on MSCOCO En-Fr
53.2BLEUmBART + MT* w/ adapters
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| mBART + MT* w/ adaptersObjectives=NMT + MLM, trainable_params=12.6M2022.12 | 53.2 | 0.832 | — | |
| VGAMTObjectives=MMT + VMLM, trainable_params=13.2M2022.12 | 51.1 | 0.811 | — | |
| mBART + MT*Objectives=NMT, trainable_params=-2022.12 | 47 | 0.733 | — | |
| VALHALLA*Method Category=Image-free Methods, Ensemble=true2025.07 | 46.5 | — | — | |
| VALHALLAMethod Category=Image-dependent Methods, Ensemble=false2025.07 | 46.4 | — | — | |
| Noise-robustMethod Category=Image-dependent Methods, Ensemble=false2025.07 | 46.34 | — | — | |
| D2P-MMT (R)Method Category=Image-dependent Methods, Inference Image Source=Reconstructed, Ensemble=false2025.07 | 46.23 | — | — | |
| MMT-VQAMethod Category=Image-dependent Methods, Ensemble=false2025.07 | 45.75 | — | — | |
| D2P-MMT (A)Method Category=Image-dependent Methods, Inference Image Source=Authentic, Ensemble=false2025.07 | 45.44 | — | — | |
| DCCNMethod Category=Image-dependent Methods, Ensemble=false2025.07 | 45.4 | — | — | |
| ImagiTMethod Category=Image-free Methods, Ensemble=false2025.07 | 45.3 | — | — | |
| Doubly-ATTMethod Category=Image-dependent Methods, Ensemble=false2025.07 | 45.16 | — | — | |
| Gated Fusion*Method Category=Image-dependent Methods, Ensemble=true2025.07 | 44.86 | — | — | |
| Selective AttentionMethod Category=Image-dependent Methods, Ensemble=false2025.07 | 44.82 | — | — | |
| Transformer-TinyMethod Category=Text-Only Transformer, Ensemble=false2025.07 | 44.81 | — | — | |
| RMMT*Method Category=Image-free Methods, Ensemble=true2025.07 | 44.5 | — | — | |
| TLM + MT*Objectives=NMT, trainable_params=42M2022.12 | 43.6 | 0.542 | — | |
| VTLM + MMT*Objectives=MMT, trainable_params=44M2022.12 | 43.4 | 0.5 | — | |
| Vanilla MT*Objectives=NMT, trainable_params=4.0M2022.12 | 41.2 | 0.403 | — | |
| Graph-MMT*Objectives=MMT, trainable_params=4.0M2022.12 | 41 | 0.387 | — | |
| Gated Fusion*Objectives=MMT, trainable_params=2.8M2022.12 | 40.4 | 0.394 | — | |
| Gated FusionModality=Multimodal2022.12 | — | — | 60.1 | |
| Graph-MMTModality=Multimodal2022.12 | — | — | 61.6 | |
| mBART + MTModality=Text-only, Adapters=false2022.12 | — | — | 66.4 | |
| mBART + MTModality=Text-only, Adapters=true2022.12 | — | — | 69.5 | |
| TLM + MTModality=Text-only2022.12 | — | — | 63.4 | |
| Vanilla MTModality=Text-only2022.12 | — | — | 61.5 | |
| VGAMTModality=Multimodal2022.12 | — | — | 68.9 | |
| VTLM + MMTModality=Multimodal2022.12 | — | — | 63.3 |