Information Extraction on SROIE (test)
98.7F1 ScoreStructTexT
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| StructTexTModality=T+L+V2024.02 | 98.7 | — | — | |
| LAMBERTSize variant=Base, Parameters=125M2021.02 | 98.17 | — | — | |
| TILTSize variant=Large, Parameters=780M2021.02 | 98.1 | — | — | |
| TILT LargeParameters=780M, Modality=T+L+I2023.05 | 98.1 | — | — | |
| LayoutLMv2Size variant=Large, Parameters=426M2021.02 | 97.81 | — | — | |
| LayoutLMv2 LargeParameters=426M, Modality=T+L+I2023.05 | 97.81 | — | — | |
| TILTSize variant=Base, Parameters=230M2021.02 | 97.65 | — | — | |
| TILTModel scale=base2022.10 | 97.65 | — | — | |
| TILT BaseParameters=230M, Modality=T+L+I2023.05 | 97.65 | — | — | |
| ERNIE-Layout LargeModality=T+L+I2023.05 | 97.55 | — | — | |
| LayoutMask LargeParameters=404M, Modality=T+L2023.05 | 97.27 | — | — | |
| ERNIE-LayoutModel scale=base2022.10 | 97.19 | — | — | |
| LayoutMask BaseParameters=182M, Modality=T+L2023.05 | 96.87 | — | — | |
| BROS_LARGEText block order=With order, Modality=Text + Layout2021.08 | 96.62 | — | — | |
| LayoutLMv2_LARGEText block order=With order, Modality=Text + Layout + Image2021.08 | 96.39 | — | — | |
| BROS_BASEText block order=With order, Modality=Text + Layout2021.08 | 96.28 | — | — | |
| LayoutLMv2Size variant=Base, Parameters=200M2021.02 | 96.25 | — | — | |
| LayoutLMv2Model scale=base2022.10 | 96.25 | — | — | |
| LayoutLMv2 BaseParameters=200M, Modality=T+L+I2023.05 | 96.25 | — | — | |
| LayoutLMv2_BASEText block order=With order, Modality=Text + Layout + Image2021.08 | 96.09 | — | — | |
| BROS BaseParameters=110M, Modality=T+L2023.05 | 95.48 | — | — | |
| LayoutLM_LARGEText block order=With order, Modality=Text + Layout2021.08 | 95.36 | — | — | |
| LayoutLMSize variant=Large, Parameters=343M2021.02 | 95.24 | — | — | |
| LayoutLM LargeParameters=343M, Modality=T+L2023.05 | 95.24 | — | — | |
| LayoutLM_BASEText block order=With order, Modality=Text + Layout2021.08 | 95.11 | — | — | |
| UniLMv2 LargeParameters=355M, Modality=T2023.05 | 94.88 | — | — | |
| UniLMv2 BaseParameters=125M, Modality=T2023.05 | 94.59 | — | — | |
| LayoutLMModel scale=base2022.10 | 94.38 | — | — | |
| LayoutLM BaseParameters=160M, Modality=T+L+I2023.05 | 94.38 | — | — | |
| BERT_LARGEText block order=With order, Modality=Text2021.08 | 94.25 | — | — | |
| BERT_BASEText block order=With order, Modality=Text2021.08 | 93.67 | — | — | |
| BERT LargeParameters=340M, Modality=T2023.05 | 92 | — | — | |
| DocLLM-7BModel Size=7B, Modality=Text+Layout, Evaluation Setting=SDDS2023.12 | 91.9 | — | — | |
| Proprietary KIE ModelModality=T+L2024.02 | 91.7 | — | — | |
| LayoutLMSize variant=Base, Parameters=113-160M2021.02 | 91.38 | — | — | |
| DocLLM-1BModel Size=1B, Modality=Text+Layout, Evaluation Setting=SDDS2023.12 | 91 | — | — | |
| BERT BaseParameters=110M, Modality=T2023.05 | 90.99 | — | — | |
| GPT-4+OCRModel Size=~1T, Modality=Text, Evaluation Setting=Zero-Shot (ZS)2023.12 | 90.6 | — | — | |
| OriginalBackbone=Qwen2.5-VL-Instruct-3B2025.09 | 88.7 | — | 97.5 | |
| Index-Preserving Lightweight Token PruningBackbone=Qwen2.5-VL-Instruct-3B2025.09 | 87.9 | — | 97.8 | |
| PLAINTEXTModality=T2024.02 | 79.9 | — | — | |
| SPATIAL FORMATYModality=T+L2024.02 | 79 | — | — | |
| SPATIAL FORMATModality=T+L2024.02 | 77 | — | — | |
| BOUNDING BOXModality=T+L2024.02 | 75.4 | — | — | |
| BOUNDINGBOXMARKUPModality=T+L2024.02 | 74.3 | — | — | |
| CENTERPOINTModality=T+L2024.02 | 73.3 | — | — | |
| LayoutLLM-7BModel Category=Ours, Evaluation Protocol=Zero-shot, Backbone Initialization=Vicuna-1.5-7B2024.04 | 72.12 | — | — | |
| LayoutLLM-7BModel Category=Ours, Evaluation Protocol=Zero-shot, Backbone Initialization=Llama2-7B-chat2024.04 | 70.97 | — | — | |
| DocKylinBackbone=Qwen2.5-VL-Instruct-3B2025.09 | 69.6 | — | 83.5 | |
| Llama2-7B-chatModel Category=LLM, Input Representation=Plain Text, Evaluation Protocol=Zero-shot2024.04 | 68.97 | — | — | |
| Vicuna-1.5-7BModel Category=LLM, Input Representation=Plain Text, Evaluation Protocol=Zero-shot2024.04 | 68.2 | — | — | |
| Vicuna-7BModel Category=LLM, Input Representation=Plain Text, Evaluation Protocol=Zero-shot2024.04 | 67.49 | — | — | |
| Vicuna-1.5-7BModel Category=LLM, Input Representation=Layout Text (Text + Box), Evaluation Protocol=Zero-shot2024.04 | 66.2 | — | — | |
| Qwen-VL-7BModel Category=MLLM, Evaluation Protocol=Fine-tuned, Implementation=Original Paper2024.04 | 58.59 | — | — | |
| Llama2+OCRModel Size=7B, Modality=Text, Evaluation Setting=Zero-Shot (ZS)2023.12 | 56.4 | — | — | |
| Llama2-7B-chatModel Category=LLM, Input Representation=Layout Text (Text + Box), Evaluation Protocol=Zero-shot2024.04 | 51.15 | — | — | |
| Vicuna-7BModel Category=LLM, Input Representation=Layout Text (Text + Box), Evaluation Protocol=Zero-shot2024.04 | 45.43 | — | — | |
| Llama2-7BModel Category=LLM, Input Representation=Layout Text (Text + Box), Evaluation Protocol=Zero-shot2024.04 | 34.96 | — | — | |
| Llama2-7BModel Category=LLM, Input Representation=Plain Text, Evaluation Protocol=Zero-shot2024.04 | 15.86 | — | — | |
| LLaVA-1.5-7BModel Category=MLLM, Evaluation Protocol=Zero-shot2024.04 | 3.83 | — | — | |
| LLaVAR-7BModel Category=MLLM, Evaluation Protocol=Zero-shot, Implementation=Original Paper2024.04 | 2.38 | — | — | |
| ToMEBackbone=Qwen2.5-VL-Instruct-3B2025.09 | 0 | — | 9.9 | |
| BERTModel size=BASE2020.12 | — | 90.99 | — | |
| BERTModel size=LARGE2020.12 | — | 92 | — | |
| BERT_largeModel Size=Large2022.10 | — | 92 | — | |
| BROS2020.12 | — | 95.48 | — | |
| ERNIE-Layout_largeModel Size=Large2022.10 | — | 97.55 | — | |
| LayoutLMModel size=BASE2020.12 | — | 94.38 | — | |
| LayoutLMModel size=LARGE2020.12 | — | 95.24 | — | |
| LayoutLM_largeModel Size=Large2022.10 | — | 95.24 | — | |
| LayoutLMv2Model size=BASE2020.12 | — | 96.25 | — | |
| LayoutLMv2Model size=LARGE2020.12 | — | 97.81 | — | |
| LayoutLMv2_largeModel Size=Large2022.10 | — | 97.81 | — | |
| PICK2020.12 | — | 96.12 | — | |
| RoBERTa_largeModel Size=Large2022.10 | — | 92.8 | — | |
| TILT_largeModel Size=Large2022.10 | — | 98.1 | — | |
| Top-1 on SROIE LeaderboardDate range=until 2020-12-242020.12 | — | 97.67 | — | |
| TRIE2020.12 | — | 96.18 | — | |
| UniLMv2Model size=BASE2020.12 | — | 94.59 | — | |
| UniLMv2Model size=LARGE2020.12 | — | 94.88 | — | |
| UniLMv2_largeModel Size=Large2022.10 | — | 94.88 | — |