tsing90 pytorch_semantic_human_matting issues

tsing90 / pytorch_semantic_human_matting

This is an unofficial implementation of the paper "Semantic human matting":

https://arxiv.org/pdf/1809.01354.pdf

83 stars 18 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

After T-net training, pth model file getting so big.

#38 LoveU3tHousand2 opened 3 years ago
0
预训练的模型

#37 tomkitty1999 opened 3 years ago
0
The reason why did you pretrain T-net and M-net ？

#36 hanzhangliu closed 4 years ago
0
关于test.py中测试t_net的一个问题

#35 jiusigua opened 4 years ago
0
一个关于数据准备的问题

#34 jiusigua opened 4 years ago
1
about m_net output

#33 sjf18 opened 4 years ago
4
test m_net

#32 Tomhouxin opened 5 years ago
3
DIM数据集从哪可以申请下载？

#31 Tomhouxin closed 5 years ago
1
I feel it is slow when training.

#30 xuehuachunsheng opened 5 years ago
3
Predict alpha result

#29 Jason-xin closed 5 years ago
0
M-Net loss too small

#28 sanshibayuan opened 5 years ago
0
How big is your dataset?

#27 sanshibayuan opened 5 years ago
0
训练T网络效果不好

#26 tobechao opened 5 years ago
8
May be a small bug...

#25 Jiaozrr opened 5 years ago
6
question about `network.py`

#24 liminn closed 5 years ago
1
generate trimap question

#23 ztgameming closed 5 years ago
1
训练M-net的中间结果

#22 songcheng closed 5 years ago
2
One question

#21 Zyttty closed 5 years ago
1
hello tsing90，can you share the dataset about adobe of DIM for study？

#20 l2009312042 closed 5 years ago
1
loading data for e2e training

#19 xinggangw opened 5 years ago
4
How did you construct bg image set

#18 lan2720 closed 5 years ago
2
test M net

#17 Oseira closed 5 years ago
2
Error on the new Train file.

#16 usalexsantos closed 5 years ago
1
Hi,could you upload the pretrained model?

#15 wodsoe closed 5 years ago
2
why test result is very different from train an validation result?

#14 hellozgm closed 5 years ago
8
about load pretrained model when training end to end

#13 wrrJasmine closed 5 years ago
2
结果是怎样生成的呢？

#12 judgeeeeee closed 5 years ago
0
How about MobileNet as T-Net?

#11 pourfard closed 5 years ago
1
Train time and Train loss

#10 kkkmax closed 5 years ago
1
compositing on the fly

#9 Oseira closed 5 years ago
3
Background composition

#8 gdjmck closed 5 years ago
2
how does alpha_r look like?

#7 gdjmck closed 5 years ago
12
M net output

#6 judgeeeeee closed 5 years ago
4
Qualitative results

#5 fahrenheitjo closed 5 years ago
1
predict abnormal results of T-net

#4 kkkmax closed 5 years ago
1
请问效果如何

#3 taasoo closed 5 years ago
5
Error running train.

#2 usalexsantos closed 5 years ago
1
Dataset

#1 usalexsantos closed 5 years ago
7