search

comzzw / Generative-Image-Inpainting

[AAAI2023, Oral] Official Implementation of Generative Image Inpainting with Segmentation Confusion Adversarial Training and Contrastive Learning
https://arxiv.org/abs/2303.13133
17 stars 3 forks source link

Questions about training on celeba-hq #2

Open Ethereal256 opened 6 months ago

Ethereal256 commented 6 months ago 
 Hi comzzw, when I train on the dataset celeba-hq, it returns "num_feat_layers <= len(feature_gt_final)", do you know why? 
 I printed the value of num_feat_layers during training and it showed 5. And the len(feat_gt_final)=4. What should I do?
Zhangyang9910 commented 4 months ago

I'm also having the same issue, please do you have a solution at the moment？

Ethereal256 commented 1 week ago

I'm also having the same issue, please do you have a solution at the moment？ Nope