Jiahang Wang
Jiahang Wang
1. I didn't find something wrong from your description. 2. You can directly retrain the model following the guidance in the train.sh step by step. 3. The default training epoch...
Hi, maybe the main reason of unreasonable results is the domain difference between the testing images and training dataset(MPV), where the background of person images are more simple(gray or white)....
If you test on the custom dataset, maybe retraining or fine tuning our model on your own dataset would be a choice.
Question about total variation regularization and training strategy of appearance generative network
Hi, the tv loss doesn't make so many differences, if you want to use this criterion, just set a small loss weight. For joint training, using joint loss in appearance...
Currently the pretrained model only supports the women face since the people in the dataset used in our experiments are almost female.
> Hi, I can build nms, but I can't build roi_align. I'm not sure where is the problem, please help me, thanks! hi, have u solved the problem? I have...
> > Hi, I can build nms, but I can't build roi_align. I'm not sure where is the problem, please help me, thanks! > > hi, have u solved the...
> ，想用来识别一下自己，不知道怎么弄，请问能给出这部分代码吗？ 请问训练好的模型可以分享一下吗，谢谢
请问这个链接的model就是使用LFW数据集和siamese、三元组损失（本项目中代码）中训练出来的模型吗？谢谢。
> Group =s , 分组操作使得卷积核为原来的channel/s, 代码里面的裁剪通道还是channel， > 比如（input=256,output=256, group=2）其实只有128个卷积通道， 代码里面的out of range Hi, 请问分组卷积这里有比较好的解决方案吗？