Natural Language Understanding on GLUE large tasks (dev)
94.1SST-2 ScoreVid-Voken w/CLIP
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Vid-Voken w/CLIPStudent Backbone=BERT-base, Teacher Visual Backbone=CLIP2026.03 | 94.1 | 89.8 | 89 | 83.9 | 89.2 | |
| Vid-Voken w/ResNetStudent Backbone=BERT-base, Teacher Visual Backbone=ResNet2026.03 | 93.4 | 89.2 | 88.7 | 83 | 88.6 | |
| ARMADAStudent Backbone=BERT-base, Teacher Model=Stable Diffusion2026.03 | 93.1 | 91.3 | 89.5 | 83.5 | 89.4 | |
| X-adapterStudent Backbone=BERT-base2026.03 | 92.7 | 91.4 | 89.3 | 84.4 | 89.4 | |
| Img-VokenStudent Backbone=BERT-base2026.03 | 92.2 | 88.6 | 88.6 | 82.6 | 88 | |
| ViLBERTStudent Backbone=BERT-base2026.03 | 90.3 | 89.6 | 88.4 | 82.4 | 87.7 | |
| VisualBERTStudent Backbone=BERT-base2026.03 | 90.3 | 88.9 | 88.4 | 82.4 | 87.5 | |
| VL-BERTStudent Backbone=BERT-base2026.03 | 90.1 | 89.5 | 88.6 | 82.9 | 87.8 | |
| Visual GuidanceStudent Backbone=BERT-base2026.03 | 89.7 | 88.2 | 84.2 | 79.5 | 85.4 |