Stride for layer4 in resnet
ruotianluo opened this issue · 5 comments
ruotianluo commented
@philokey, I will create a new branch for this so that the current pretrained won't be messed up.
This issue was originally opened in ruotianluo/pytorch-resnet#2
ruotianluo commented
You can check the convert_from_tensorflow branch now.
philokey commented
Thanks, I will try it.
ruotianluo commented
It's now significantly slower (because the stride 1 needs more computation) than tf-faster-rcnn, I will investigate what's the bottleneck. (0.31 -> 0.37 s per iter)
Feynman27 commented
@ruotianluo I'm trying to train a resnet101+faster rcnn model starting from image net weights. Changing the stride in layer4 from 2->1 does indeed increase the training time. Did you ever resolve the bottleneck?
ruotianluo commented
@Feynman27 I haven't profiled the code in detail. But it should be faster after commit df3cc05