Information Extraction on CORD (test)
98.94F1 ScoreUDoc
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| UDocModality=V+T+L2022.12 | 98.94 | — | — | — | — | — | — | — | — | |
| DocFormerv2_large#param=750M, modalities=image + text + spatial2023.06 | 97.7 | — | 97.71 | 97.7 | — | — | — | — | — | |
| UDOP-DualModality=V+T+L2022.12 | 97.64 | — | — | — | — | — | — | — | — | |
| UDOPModality=Vision+Text+Layout2022.12 | 97.58 | — | — | — | — | — | — | — | — | |
| UDOPModality=V+T+L2022.12 | 97.58 | — | — | — | — | — | — | — | — | |
| UDOP#param=794M, modalities=image + text + spatial, 2D position source=OCR lines (not word box)2023.06 | 97.58 | — | — | — | — | — | — | — | — | |
| LayoutLMv3largeModality=Vision+Text+Layout2022.12 | 97.46 | — | — | — | — | — | — | — | — | |
| LayoutLMv3largeModality=V+T+L2022.12 | 97.46 | — | — | — | — | — | — | — | — | |
| LayoutLMv3_LARGEParameters=368M + α, Pretrain Data Size=11M2023.03 | 97.46 | — | — | — | — | — | — | — | — | |
| LayoutLMv3_large#param=368M, modalities=image + text + spatial2023.06 | 97.46 | — | — | — | — | — | — | — | — | |
| LayoutLMv3 LargeParameters=368M, Modality=T+L+I2023.05 | 97.46 | — | — | — | — | — | — | — | — | |
| LayoutLMv3Model Category=Fine-tuned PTM, Evaluation Protocol=Fine-tuned2024.04 | 97.46 | — | — | — | — | — | — | — | — | |
| BROSlargeModality=Text+Layout2022.12 | 97.4 | — | — | — | — | — | — | — | — | |
| BROSlargeModality=T+L2022.12 | 97.4 | — | — | — | — | — | — | — | — | |
| BROS LargeParameters=340M, Modality=T+L2023.05 | 97.4 | — | — | — | — | — | — | — | — | |
| BROS_LARGEText block order=With order, Modality=Text + Layout2021.08 | 97.28 | — | — | — | — | — | — | — | — | |
| FormNetModality=Text+Layout2022.12 | 97.28 | — | — | — | — | — | — | — | — | |
| FormNetModality=T+L2022.12 | 97.28 | — | — | — | — | — | — | — | — | |
| BROS_LARGEParameters=340M + α, Pretrain Data Size=11M2023.03 | 97.28 | — | — | — | — | — | — | — | — | |
| FormNet#param=345M, modalities=text + spatial2023.06 | 97.28 | — | 98.02 | 96.55 | — | — | — | — | — | |
| LayoutLMv2_LARGEText block order=With order, Modality=Text + Layout + Image2021.08 | 97.24 | — | — | — | — | — | — | — | — | |
| ERNIE-Layout LargeModality=T+L+I2023.05 | 97.21 | — | — | — | — | — | — | — | — | |
| LayoutMask LargeParameters=404M, Modality=T+L2023.05 | 97.19 | — | — | — | — | — | — | — | — | |
| DocFormerlargeModality=Vision+Text+Layout2022.12 | 96.99 | — | — | — | — | — | — | — | — | |
| DocFormerlargeModality=V+T+L2022.12 | 96.99 | — | — | — | — | — | — | — | — | |
| DocFormer_LARGEParameters=536M + α, Pretrain Data Size=5M2023.03 | 96.99 | — | — | — | — | — | — | — | — | |
| DocFormer_large#param=536M, modalities=image + text + spatial2023.06 | 96.99 | — | 97.25 | 96.74 | — | — | — | — | — | |
| LayoutMask BaseParameters=182M, Modality=T+L2023.05 | 96.99 | — | — | — | — | — | — | — | — | |
| DocFormer LargeParameters=536M, Modality=T+L+I2023.05 | 96.99 | — | — | — | — | — | — | — | — | |
| UniDocModality=Vision+Text+Layout2022.12 | 96.86 | — | — | — | — | — | — | — | — | |
| UDoc#param=272M, modalities=image + text + spatial2023.06 | 96.86 | — | — | — | — | — | — | — | — | |
| DocFormerv2_base#param=232M, modalities=image + text + spatial2023.06 | 96.8 | — | 97.51 | 96.1 | — | — | — | — | — | |
| LayoutLMv3_large#param=368M, modalities=image + text + spatial, 2D position source=word boxes2023.06 | 96.78 | — | 96.78 | 96.78 | — | — | — | — | — | |
| ERNIE-LayoutModel scale=base2022.10 | 96.61 | — | — | — | — | — | — | — | — | |
| LayoutLMv3_BASEParameters=133M + α, Pretrain Data Size=11M2023.03 | 96.56 | — | — | — | — | — | — | — | — | |
| LayoutLMv3_base#param=133M, modalities=image + text + spatial2023.06 | 96.56 | — | — | — | — | — | — | — | — | |
| LayoutLMv3 BaseParameters=133M, Modality=T+L+I2023.05 | 96.56 | — | — | — | — | — | — | — | — | |
| BROS_BASEText block order=With order, Modality=Text + Layout2021.08 | 96.5 | — | — | — | — | — | — | — | — | |
| BROS_BASEParameters=110M + α, Pretrain Data Size=11M2023.03 | 96.5 | — | — | — | — | — | — | — | — | |
| TILTSize variant=Large, Parameters=780M2021.02 | 96.33 | — | — | — | — | — | — | — | — | |
| DocFormerModel scale=base2022.10 | 96.33 | — | — | — | — | — | — | — | — | |
| TILTlargeModality=Vision+Text+Layout2022.12 | 96.33 | — | — | — | — | — | — | — | — | |
| TILTlargeModality=V+T+L2022.12 | 96.33 | — | — | — | — | — | — | — | — | |
| TILT_large#param=780M, modalities=image + text + spatial2023.06 | 96.33 | — | — | — | — | — | — | — | — | |
| DocFormer_base#param=183M, modalities=image + text + spatial2023.06 | 96.33 | — | 96.52 | 96.14 | — | — | — | — | — | |
| DocFormer BaseParameters=183M, Modality=T+L+I2023.05 | 96.33 | — | — | — | — | — | — | — | — | |
| TILT LargeParameters=780M, Modality=T+L+I2023.05 | 96.33 | — | — | — | — | — | — | — | — | |
| LayoutLM_BASEText block order=With order, Modality=Text + Layout2021.08 | 96.26 | — | — | — | — | — | — | — | — | |
| LayoutLM_LARGEText block order=With order, Modality=Text + Layout2021.08 | 96.12 | — | — | — | — | — | — | — | — | |
| LILTModality=Text+Layout2022.12 | 96.07 | — | — | — | — | — | — | — | — | |
| LILTModality=T+L2022.12 | 96.07 | — | — | — | — | — | — | — | — | |
| LayoutLM_BASEParameters=160M + α, Pretrain Data Size=11M2023.03 | 96.07 | — | — | — | — | — | — | — | — | |
| LiLT BaseModality=T+L2023.05 | 96.07 | — | — | — | — | — | — | — | — | |
| LAMBERTSize variant=Base, Parameters=125M2021.02 | 96.06 | — | — | — | — | — | — | — | — | |
| LayoutLMv2_BASEText block order=With order, Modality=Text + Layout + Image2021.08 | 96.05 | — | — | — | — | — | — | — | — | |
| LayoutLMv2Size variant=Large, Parameters=426M2021.02 | 96.01 | — | — | — | — | — | — | — | — | |
| LayoutLMv2largeModality=Vision+Text+Layout2022.12 | 96.01 | — | — | — | — | — | — | — | — | |
| LayoutLMv2largeModality=V+T+L2022.12 | 96.01 | — | — | — | — | — | — | — | — | |
| LayoutLMv2_LARGEParameters=426M + α, Pretrain Data Size=11M2023.03 | 96.01 | — | — | — | — | — | — | — | — | |
| LayoutLMv2_large#param=426M, modalities=image + text + spatial2023.06 | 96.01 | — | 95.65 | 96.37 | — | — | — | — | — | |
| LayoutLMv2 LargeParameters=426M, Modality=T+L+I2023.05 | 96.01 | — | — | — | — | — | — | — | — | |
| BROS BaseParameters=110M, Modality=T+L2023.05 | 95.73 | — | — | — | — | — | — | — | — | |
| ESPParameters=50M, Pretrain Data Size=0.9M2023.03 | 95.65 | — | — | — | — | — | — | — | — | |
| BROS_base#param=139M, modalities=text + spatial2023.06 | 95.36 | — | 95.58 | 95.14 | — | — | — | — | — | |
| TILTSize variant=Base, Parameters=230M2021.02 | 95.11 | — | — | — | — | — | — | — | — | |
| TILTModel scale=base2022.10 | 95.11 | — | — | — | — | — | — | — | — | |
| TILT_base#param=230M, modalities=image + text + spatial2023.06 | 95.11 | — | — | — | — | — | — | — | — | |
| TILT BaseParameters=230M, Modality=T+L+I2023.05 | 95.11 | — | — | — | — | — | — | — | — | |
| LayoutLMv2Size variant=Base, Parameters=200M2021.02 | 94.95 | — | — | — | — | — | — | — | — | |
| LayoutLMv2Model scale=base2022.10 | 94.95 | — | — | — | — | — | — | — | — | |
| LayoutLMv2_BASEParameters=200M + α, Pretrain Data Size=11M2023.03 | 94.95 | — | — | — | — | — | — | — | — | |
| LayoutLMv2_base#param=200M, modalities=image + text + spatial2023.06 | 94.95 | — | 94.53 | 95.39 | — | — | — | — | — | |
| LayoutLMv2 BaseParameters=200M, Modality=T+L+I2023.05 | 94.95 | — | — | — | — | — | — | — | — | |
| LayoutLMSize variant=Large, Parameters=343M2021.02 | 94.93 | — | — | — | — | — | — | — | — | |
| LayoutLMv1_large#param=343M, modalities=text + spatial2023.06 | 94.93 | — | 94.32 | 95.54 | — | — | — | — | — | |
| BERT_LARGEText block order=With order, Modality=Text2021.08 | 94.74 | — | — | — | — | — | — | — | — | |
| LayoutLMSize variant=Base, Parameters=113-160M2021.02 | 94.72 | — | — | — | — | — | — | — | — | |
| LayoutLMModel scale=base2022.10 | 94.72 | — | — | — | — | — | — | — | — | |
| LayoutLMv1_base#param=113M, modalities=text + spatial2023.06 | 94.72 | — | 94.37 | 95.08 | — | — | — | — | — | |
| DocTr2023.07 | 94.4 | — | — | — | — | — | — | — | — | |
| EXTRACTCONF (Isotonic)Calibration Method=Isotonic, Training Dataset=DocILE, Evaluation Protocol=Zero-shot transfer2026.06 | 94.2 | — | 90.9 | — | — | 0.839 | 0.029 | 0.081 | 0.032 | |
| EXTRACTCONF (Lasso)Calibration Method=Lasso, Training Dataset=DocILE, Evaluation Protocol=Zero-shot transfer2026.06 | 94 | — | 89.9 | — | — | 0.854 | 0.025 | 0.078 | 0.029 | |
| RoBERTa_LARGEParameters=355M + α2023.03 | 93.8 | — | — | — | — | — | — | — | — | |
| RoBERTa_large#param=355M, modalities=text + spatial2023.06 | 93.8 | — | — | — | — | — | — | — | — | |
| ROBERTA LargeParameters=355M, Modality=T2023.05 | 93.8 | — | — | — | — | — | — | — | — | |
| LayoutLMv3_base#param=133M, modalities=image + text + spatial, 2D position source=word boxes2023.06 | 93.61 | — | 92.92 | 94.31 | — | — | — | — | — | |
| LayoutLMv3Tagging Formulation=DocTr formulation2023.07 | 93.6 | — | — | — | — | — | — | — | — | |
| RoBERTa_BASEParameters=125M + α2023.03 | 93.54 | — | — | — | — | — | — | — | — | |
| ROBERTA BaseParameters=125M, Modality=T2023.05 | 93.54 | — | — | — | — | — | — | — | — | |
| BERT_BASEText block order=With order, Modality=Text2021.08 | 93.13 | — | — | — | — | — | — | — | — | |
| BROSTagging Formulation=DocTr formulation2023.07 | 92.9 | — | — | — | — | — | — | — | — | |
| LayoutLMv2Tagging Formulation=DocTr formulation2023.07 | 92.7 | — | — | — | — | — | — | — | — | |
| SPADE2023.07 | 92.5 | — | — | — | — | — | — | — | — | |
| LayoutLMv3Tagging Formulation=IOB2023.07 | 92.2 | — | — | — | — | — | — | — | — | |
| UniLMv2_large#param=355M, modalities=text + spatial2023.06 | 92.05 | — | 91.23 | 92.89 | — | — | — | — | — | |
| UniLMv2 LargeParameters=355M, Modality=T2023.05 | 92.05 | — | — | — | — | — | — | — | — | |
| BROSTagging Formulation=IOB2023.07 | 91.8 | — | — | — | — | — | — | — | — | |
| DonutModality=Vision2022.12 | 91.6 | — | — | — | — | — | — | — | — | |
| DonutModality=V2022.12 | 91.6 | — | — | — | — | — | — | — | — | |
| SPADEmodalities=text + spatial2023.06 | 91.5 | — | — | — | — | — | — | — | — |