Closed thanhuitha closed 3 years ago
Hi,
In general, yes, it should work! The only difference is that my model is developed, for now, for relatively small images! (MNIST, CIFAR-10, CIFAR-100, SVHN, TinyImageNet). Also, the VGG architecture is much heavier than the ResNet-34, so the time and memory consumption are increased a lot! Keep me updated if you take this direction,
Guglielmo
Hi Guglielmo Camporese, congratulations to you for your work! I want to ask Can I can apply this model in open-set face recognition problem, for example in LFW dataset, by replace resnet34 by VGG-Face? Can you give me some advice about this ^^ many thanks