aaron-xichen / pytorch-playground

Base pretrained models and datasets in pytorch (MNIST, SVHN, CIFAR10, CIFAR100, STL10, AlexNet, VGG16, VGG19, ResNet, Inception, SqueezeNet)
MIT License
2.62k stars 612 forks source link

ofmap value of LinearQuant layer #50

Closed thehalocline closed 4 years ago

thehalocline commented 4 years ago

when I print all of my ofmap of linear_quantization, the negative values are all zero even if they do not go through relu. So, ReLU input = ReLU output right now. Do you know why?