search

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

error accuracy #15

Closed TangKun1234 closed 4 years ago

TangKun1234 commented 6 years ago

I download the alexnet model from (https://download.pytorch.org/models/alexnet-owt-4df8aa71.pth) and evaluate the model by the val224_compressed.pkl with quantize.py , but the accuracy only 0.5397. Why?

aaron-xichen commented 4 years ago

The default setting will quantize the weights with 8bits and activation with 8bits.