LayoutLMv2BASE

Reported on 5 benchmarks across 3 tasks · 1 paper

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

Natural Language Processing5 results

Visual Question Answering (VQA)onDocVQA test
ANLS· 2020-12-29
0.7808
best: 0.9436 (Human)
LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding arXiv:2012.14740
Semantic entity labelingonFUNSD
F1· 2020-12-29
82.76
best: 93.2 (LayoutMask (large))
LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding arXiv:2012.14740
Key Information ExtractiononCORD
F1· 2020-12-29
94.95
best: 98.52 (RORE (GeoLayoutLM))
LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding arXiv:2012.14740
Key Information ExtractiononKleister NDA
F1· 2020-12-29
83.3
best: 85.2 (LayoutLMv2LARGE)
LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding arXiv:2012.14740
Key Information ExtractiononSROIE
F1· 2020-12-29
96.25
best: 97.81 (LayoutLMv2LARGE (Excluding OCR mismatch))
LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding arXiv:2012.14740