image-sentence alignment on VALSE coreference standard
pairwise accuracy
- 2021-12-14 — ViLBERT 12-in-1: pairwise accuracy 75.7
| Rank | Model | pairwise accuracy | Accuracy (%) | Paper | Code | Year |
|---|---|---|---|---|---|---|
| 11 | VisualBERT | 49.5 | 50.0 | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |
| 12 | ViLBERT | 47.2 | 50.0 | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |
| 13 | LXMERT | 46.8 | 49.8 | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |
| 14 | GPT1 | 45.6 | – | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |