MaskOCR-L

Reported on 1 benchmark across 1 task · 1 paper · 1 SOTA

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

Methodology1 result

Optical Character Recognition (OCR)onBenchmarking Chinese Text Recognition: Datasets, Baselines, and an Empirical Study
Accuracy (%)· 2022-06-01
82.6
best: 89.6 (DTrOCR)
SOTA
MaskOCR: Text Recognition with Masked Encoder-Decoder Pretraining arXiv:2206.00311