image-sentence alignment on VALSE coreference clean
pairwise accuracy
- 2021-12-14 — ViLBERT 12-in-1: pairwise accuracy 69.2
| Rank | Model | pairwise accuracy | Accuracy (%) | Paper | Code | Year |
|---|---|---|---|---|---|---|
| 11 | ViLBERT | 48.1 | 50.0 | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |
| 12 | VisualBERT | 47.6 | 50.0 | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |
| 13 | GPT1 | 45.2 | – | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |
| 14 | LXMERT | 44.2 | 49.0 | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |