MoVie: Revisiting Modulated Convolutions for Visual Counting and Beyond

Benchmark Model Rank Results
object-counting-on-tallyqa-complexMoVie-ResNeXt#1Accuracy: 56.8RMSE: 1.43
object-counting-on-tallyqa-complexMoVie#3Accuracy: 54.1RMSE: 1.52
object-counting-on-tallyqa-simpleMoVie-ResNeXt#1Accuracy: 74.9RMSE: 1
object-counting-on-tallyqa-simpleMoVie#3Accuracy: 70.8RMSE: 1.09