Bestyunshen

楼主你好，我想问一下对于这种既有分类器又有判别器的架构，可不可以先把所有的loss都backward()之后，再一次执行判别器，分类器，生成器等等模块的optimizer.step() 因为我之前自己实现CVAE-GAN的时候，先执行C的zero_gard,再执行loss_C.backward(),然后再执行optimizer_C.step()，按照同样步骤运行到D的backward就会报错,但是先把所有loss都bankward之后再更新参数就没报错 所以这种backward+step的步骤 顺序是绑定的吗 按照我那种方式统一step可不可以- -

请问预训练的AU-RCNN适用于其他脸部数据集的单个图片吗

1

我在做一个表情迁移模型，希望可以检测更多的面部AU，但我的迁移模型是基于单个图片而不是视频，所以预训练的AU-RCNN可以检测单张图片吗 还是要重新训练

``` AU12 - OF None | Forward | ./Demo\000.jpg : 0.5 AU12 - OF None | Forward | ./Demo\001.jpg : 0.5 AU12 - OF None | Forward | ./Demo\002.jpg :...

The Result

2

Hi The results of inversion seems not good as the paper. Did you know what the qusetion is?

The Result

1

Hi I try to run this code but got bad results. I also try to use vgg loss followed by lpips loss ,however the result is not as good as...

How to obtain a smpl mesh from mutil-view cameras?

1

Could you share some method/repo that predict the smpl mesh from the calibrated multi-view cameras?