Multimodal Named Entity Recognition on Snap Res. (test)
0.911F1 ScoreMoRe_MoE
Evaluation Results
| Method | Links | |
|---|---|---|
| MoRe_MoERetrieval Modality=Mixture-of-Experts (Text + Image)2022.12 | 0.911 | |
| MoRe_ImageRetrieval Modality=Image-based2022.12 | 0.902 | |
| ITARetrieval Modality=None2022.12 | 0.9015 | |
| MoRe_TextRetrieval Modality=Textual2022.12 | 0.9009 | |
| Ours: BaselineRetrieval Modality=None2022.12 | 0.8935 | |
| biLSTM-CRF + RpBERT-Large_GSbackbone=BERT-Large2021.02 | 0.878 | |
| Sun et al.Retrieval Modality=None2022.12 | 0.878 | |
| ITA-All+CVASource=Our Reproductions2021.12 | 0.8744 | |
| biLSTM-CRF + RpBERT_GSbackbone=BERT-Base2021.02 | 0.874 | |
| RpBERT baseSource=Reported, Note=uses the test set to select the best model2021.12 | 0.874 | |
| RIVA +Source=Reported2021.12 | 0.868 | |
| Sun et al.Retrieval Modality=None2022.12 | 0.868 | |
| biLSTM-CRF + UNITER2021.02 | 0.861 | |
| biLSTM-CRF + VL-BERT2021.02 | 0.86 | |
| BERT-CRFSource=Reproduced2021.12 | 0.8598 | |
| biLSTM-CRF + VILBERT2021.02 | 0.857 | |
| Yu et al. (2020)2021.02 | 0.853 | |
| RpBERT baseSource=Reproduced2021.12 | 0.6214 |