DiT-L

Reported on 6 benchmarks across 1 task · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision6 results

Document Layout AnalysisonPubLayNet val
Figure· 2022-03-04
0.972
best: 0.975 (DETR)
SOTA
DiT: Self-supervised Pre-training for Document Image Transformer arXiv:2203.02378
Document Layout AnalysisonPubLayNet val
List· 2022-03-04
0.96
best: 0.975 (TRDLU)
SOTA
DiT: Self-supervised Pre-training for Document Image Transformer arXiv:2203.02378
Document Layout AnalysisonPubLayNet val
Overall· 2022-03-04
0.949
best: 0.962 (VGT)
DiT: Self-supervised Pre-training for Document Image Transformer arXiv:2203.02378
Document Layout AnalysisonPubLayNet val
Table· 2022-03-04
0.978
best: 0.981 (VGT)
DiT: Self-supervised Pre-training for Document Image Transformer arXiv:2203.02378
Document Layout AnalysisonPubLayNet val
Text· 2022-03-04
0.944
best: 0.967 (VSR)
DiT: Self-supervised Pre-training for Document Image Transformer arXiv:2203.02378
Document Layout AnalysisonPubLayNet val
Title· 2022-03-04
0.893
best: 0.939 (VGT)
DiT: Self-supervised Pre-training for Document Image Transformer arXiv:2203.02378