ArT
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Scene Text Recognition | ArT | Accuracy84.9 | 24 | |
| Scene Text Recognition | ArT (test) | Word Accuracy86.4 | 19 | |
| Scene Text Recognition | ArT 34k samples | Word Accuracy86.4 | 16 | |
| Scene Text Detection | ArT ICDAR2019 (test) | Recall73.7 | 16 | |
| Sequential Recommendation | Art (test) | Hit@50.0922 | 13 | |
| Scene Text Recognition | ArT 36-char (test) | Word Accuracy84.5 | 12 | |
| Nudity Unlearning | ART | ESD0 | 10 | |
| Safety Evaluation | ART | Safety Score99.7 | 7 | |
| Natural Language Generation | ART 10K | BLEU6 | 4 | |
| Abductive Reasoning | ART10 | Average Rating4.45 | 4 | |
| Text Detection | ArT (test) | Recall73.3 | 4 | |
| Text Recognition | ArT 2019 (test) | Accuracy58.82 | 3 | |
| Decision Making | Art UK-based participants | Final Accuracy8.5 | 1 | |
| Decision Making | Art 90% AI Accuracy (UK-based participants) | Final Accuracy- | 0 |
Showing 14 of 14 rows