Without binarizing the downsample layer for ResNet on ImageNet

htqin / IR-Net

[CVPR 2020] This project is the PyTorch implementation of our accepted CVPR 2020 paper : forward and backward information retention for accurate binary neural networks.

177 stars 38 forks source link

Without binarizing the downsample layer for ResNet on ImageNet #10

Closed liyunqianggyn closed 4 years ago

liyunqianggyn commented 4 years ago

Nice work!!! Thank you for sharing the code. When i check the code, I found that the downsample layers for ResNet on ImageNet are not binarized to 1-bit values. Is that correct? In Cifar10, the downsample layer is quantized?

htqin commented 4 years ago

Hey @liyunqianggyn , I think the problem started in #7 . We follow the setting of Bi-Real Net for fair comparison.