Visual Question Answering On Docvqa Test

Métriques

ANLS

Résultats

Résultats de performance de divers modèles sur ce benchmark

Tableau comparatif

Nom du modèle	ANLS
matcha-enhancing-visual-language-pretraining	0.742
layout-and-task-aware-instruction-prompt-for	0.884
pali-3-vision-language-models-smaller-faster	0.876
qwen-vl-a-frontier-large-vision-language	0.651
ernie-layout-layout-knowledge-enhanced-pre	0.8486
dublin-document-understanding-by-language	0.782
pix2struct-screenshot-parsing-as-pretraining	0.721
dublin-document-understanding-by-language	0.803
pali-3-vision-language-models-smaller-faster	0.886
qwen-vl-a-frontier-large-vision-language	0.9024
pali-x-on-scaling-up-a-multilingual-vision	0.868
pali-x-on-scaling-up-a-multilingual-vision	0.80
layout-and-task-aware-instruction-prompt-for	0.8336
going-full-tilt-boogie-on-document	0.8705
docvqa-a-dataset-for-vqa-on-document-images	0.665
docformerv2-local-features-for-document	0.8784
multi-label-cluster-discrimination-for-visual	0.916
unifying-vision-text-and-layout-for-universal	0.878
unifying-vision-text-and-layout-for-universal	0.847
omni-smola-boosting-generalist-multimodal	0.906
ernie-layout-layout-knowledge-enhanced-pre	0.8841
going-full-tilt-boogie-on-document	0.8392
layoutlmv2-multi-modal-pre-training-for	0.8672
omni-smola-boosting-generalist-multimodal	0.908
donut-document-understanding-transformer	0.675
end-to-end-document-recognition-and	0.632
docvqa-a-dataset-for-vqa-on-document-images	0.9436
qwen-vl-a-frontier-large-vision-language	0.626
layoutlmv2-multi-modal-pre-training-for	0.7808
pali-x-on-scaling-up-a-multilingual-vision	0.809
layout-and-task-aware-instruction-prompt-for	0.8255
screenai-a-vision-language-model-for-ui-and	0.8988
pix2struct-screenshot-parsing-as-pretraining	0.766