graycrown

hi, @Cindy-xdZhang When I try to apply the clean.hpp to calculate the Surface Normal in KITTI， the configuration are: window_size = 15 threshold = 0.1 when the non-zero value occur...

Thanks for the excellent work. In your paper, I found that you train/test your model on NYU_v2 dataset and the delta metric are calculated, for delta(1.25^3), you reach 100%, I...

The consumption of GPU memory seems abnomal

7

hi, @JiaxiongQ, thks for the wonderful work. When I try to implement the training process on Synshtic dataset, it seem abnaomal about GPU memory consumption. The parameters of the model...

1. wider_load generate a error anno.txt for image path 2. wider_load generate a error anno.txt for image name

Loss of ArcFaceLoss became nan after the loss.backward

2

My training code like this, metric is ArcFace,: feature, output = model(images, targets=labels) arc_output = metric(feature, labels) loss = loss_CE(arc_output, labels) print(metric.kernel) loss.backward() optimizer.step() print(metric.kernel) Before the loss.backward, metric.kernel is...

As the title said, since you creat the repo, 7 mon have passed, plz share the code, thks

Knowledge distillation on regression task

2

I noticed the conclusion in your paper - "In contrast, Single -> Multi knowledge distillation improves or matches the performance of the other methods on all tasks except STS, the...

I tested the Fast MTCNN on FDDB dataset, the minisize was set as 20, The max False Positive Value is 152. But most of the result on FDDB can reach...