ReviewKD++(T: faster rcnn(resnet101), S:faster rcnn(mobilenet-v2))

Reported on 3 benchmarks across 1 task · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing3 results

Knowledge DistillationonCOCO 2017 val
AP@0.5· 2023-05-26
55.18
best: 61.8 (ReviewKD++(T: faster rcnn(resnet101), S:faster rcnn(resnet50)))
Improving Knowledge Distillation via Regularizing Feature Norm and Direction arXiv:2305.17007
Knowledge DistillationonCOCO 2017 val
AP@0.75· 2023-05-26
37.21
best: 44.94 (ReviewKD++(T: faster rcnn(resnet101), S:faster rcnn(resnet50)))
Improving Knowledge Distillation via Regularizing Feature Norm and Direction arXiv:2305.17007
Knowledge DistillationonCOCO 2017 val
mAP· 2023-05-26
34.51
best: 41.03 (ReviewKD++(T: faster rcnn(resnet101), S:faster rcnn(resnet50)))
Improving Knowledge Distillation via Regularizing Feature Norm and Direction arXiv:2305.17007