Document Image Classification On Rvl Cdip
评估指标
Accuracy
Parameters
评测结果
各个模型在此基准测试上的表现结果
比较表格
| 模型名称 | Accuracy | Parameters |
|---|---|---|
| layoutlmv3-pre-training-for-document-ai-with | 95.93% | 368M |
| multimodal-side-tuning-for-document | 92.2% | 12M |
| dit-self-supervised-pre-training-for-document | 92.11% | 87M |
| roberta-a-robustly-optimized-bert-pretraining | 90.06 | 125M |
| docformer-end-to-end-transformer-for-document | 96.17% | 183M |
| improving-accuracy-and-speeding-up-document | 92.31% | - |
| docformer-end-to-end-transformer-for-document | 95.50% | 536M |
| cutting-the-error-by-half-investigation-of | 90.97% | - |
| document-image-classification-with-intra | 92.21% | - |
| layoutxlm-multimodal-pre-training-for | 95.21% | - |
| going-full-tilt-boogie-on-document | 95.25% | - |
| layoutlm-pre-training-of-text-and-layout-for | 94.42% | 160M |
| dit-self-supervised-pre-training-for-document | 92.69% | 304M |
| vlcdoc-vision-language-contrastive-pre | 93.19% | 217M |
| going-full-tilt-boogie-on-document | 95.52% | - |
| structextv2-masked-visual-textual-prediction | 93.4% | 28M |
| docxclassifier-high-performance-explainable | 94.00% | 95.4M |
| eaml-ensemble-self-attention-based-mutual | 97.70% | - |
| analysis-of-convolutional-neural-networks-for | 90.94% | - |
| lilt-a-simple-yet-effective-language | 95.68% | - |
| dopta-improving-document-layout-analysis | 94.12% | 85M |
| training-data-efficient-image-transformers | 90.32% | 87M |
| visual-and-textual-deep-feature-fusion-for | 97.05% | 197M |
| donut-document-understanding-transformer | 95.3% | - |
| layoutlmv2-multi-modal-pre-training-for | 95.64% | - |
| beit-bert-pre-training-of-image-transformers | 91.09% | 87M |
| layoutlmv3-pre-training-for-document-ai-with | 95.44% | 133M |
| multimodal-side-tuning-for-document | 92.7% | 57M |
| layoutlmv2-multi-modal-pre-training-for | 95.25% | 200M |
| structextv2-masked-visual-textual-prediction | 94.62% | 238M |
| transferdoc-a-self-supervised-transferable | 93.18% | 221M |