search

yccyenchicheng / pytorch-SegInpaint

A PyTorch Implementation of Segmentation for Image Inpainting based on SPG-Net: https://arxiv.org/abs/1805.03356
19 stars 2 forks source link

Semantic FCN Seems Lost #4

Open YIYANGCAI opened 4 years ago

YIYANGCAI commented 4 years ago

I think your code indicates that the module consists of SPN and SGN. However, the very beginning module of FCN does not appear in the model's structure. Therefore, the condition GAN is not correct.

yccyenchicheng commented 4 years ago

Thank you for your comments. I implemented this a while ago and modified the backbone based on pix2pixHD, so there might be some discrepancy compared to the original paper. I will update the code as soon as possible.

Sorry for any inconvenience.

YIYANGCAI commented 4 years ago

Thank you for your comments. I implemented this a while ago and modified the backbone based on pix2pixHD, so there might be some discrepancy compared to the original paper. I will update the code as soon as possible.

Sorry for any inconvenience.

thanks for your response. Actucally, this code is really useful, for I can train the SP network seperately, with masks, images and gt of semantics. Your work is great!