Ask for the ImageNet pretrained ResNet-152 or ResNext-101/ResNext-152

[PkuRainBow]

@hangzhaomit Really great work!

We find that the provided pretrained models all replace the 7x7 convolutions with three 3x3 convolutions and achieve really good performance.

I just have tried to use other ImageNet pretrained models(https://github.com/Cadene/pretrained-models.pytorch) and most of these models fail to achieve comparable performance with the provided ResNet-101 on the semantic segmentation tasks.

Thus, I am wondering whether could you share with us such modified pretrained models of ResNet-152, ResNext-101 or ResNext-152 over the link http://sceneparsing.csail.mit.edu/model/pretrained_resnet/

[zhangxinyu-xyz]

@hangzhaomit Really great work!

We find that the provided pretrained models all replace the 7x7 convolutions with three 3x3 convolutions and achieve really good performance.

I just have tried to use other ImageNet pretrained models(https://github.com/Cadene/pretrained-models.pytorch) and most of these models fail to achieve comparable performance with the provided ResNext-101 on the semantic segmentation tasks.

Thus, I am wondering whether could you share with us such modified pretrained models of ResNet-152, ResNext-101 or ResNext-152 over the link http://sceneparsing.csail.mit.edu/model/pretrained_resnet/

Hi, have you got the pretrained model of ResNext-152?