image-sentence alignment on VALSE foil-it (noun phrases)
pairwise accuracy
- 2021-12-14 — CLIP: pairwise accuracy 88.8
| Rank | Model | pairwise accuracy | Accuracy (%) | Paper | Code | Year |
|---|---|---|---|---|---|---|
| 10 | ViLBERT | 86.9 | 55.9 | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |
| 12 | GPT2 | 80.7 | – | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |
| 13 | GPT1 | 77.5 | – | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |
| 14 | VisualBERT | 48.5 | 46.6 | VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena | heidelberg-nlp/valse | 2021 |