Multi-label Classification on MuReD (5-fold cross-val)
83.7PrecisionQMP-RETNet
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| QMP-RETNetInitialization=Not explicitly stated in Table 1 but used for comparison2026.02 | 83.7 | 68.3 | 95.3 | 75.7 | |
| MedMambaInitialization=ImageNet-1k pretrained, Finetune Epochs=2002026.02 | 80.3 | 54.8 | 94 | 63.3 | |
| LAGNetInitialization=ImageNet-1k pretrained, Finetune Epochs=2002026.02 | 78.6 | 64.3 | 93.6 | 64.4 | |
| C-TranInitialization=ImageNet-1k pretrained, Finetune Epochs=2002026.02 | 76.1 | 65.6 | 94.5 | 74 | |
| RETFoundInitialization=Specifically pretrained weights, Finetune Epochs=202026.02 | 72 | 60.3 | 92.7 | 65.2 | |
| IRECTeInitialization=ImageNet-1k pretrained, Finetune Epochs=2002026.02 | 71.8 | 65.6 | 93.5 | 74.2 | |
| ML-GCNInitialization=ImageNet-1k pretrained, Finetune Epochs=2002026.02 | 69.7 | 58.6 | 91.2 | 64.5 | |
| InceptionV3Initialization=ImageNet-1k pretrained, Finetune Epochs=2002026.02 | 66.1 | 48.9 | 88.1 | 43.5 | |
| ResNet-18Initialization=ImageNet-1k pretrained, Finetune Epochs=2002026.02 | 61.4 | 57.6 | 89.4 | 45.2 | |
| DenseNet121Initialization=ImageNet-1k pretrained, Finetune Epochs=2002026.02 | 61.1 | 58.3 | 92.9 | 54.5 | |
| Tresnet-rInitialization=ImageNet-1k pretrained, Finetune Epochs=2002026.02 | 59.7 | 60.5 | 89.3 | 55.7 |