wang yilan comments

Results 26 comments of


                                            wang yilan

objects365 预训练模型有了吗

metoo

在RealESRNet基础上finetune RealESRGAN 随着迭代次数的增加出现了模式崩塌？

完全一样

How to use higher resolution images for model training？？

And I have a problem, the ffhq dataset is 1024x1024, how to get the data resolution 512x512 in this paper? the author seems that he dosen't show this .

ModuleNotFoundError: No module named 'tf' in convert_tensorflow.py

I think maybe the code is too old, maybe you can try import tensorflow.campat.v1 as tf

应对图像非bilinear退化问题不理想

感觉没那么复杂,就是模型太小了,这没办法,不可能突破物理极限啊

请问下两个阶段ablation的细节

哇,感谢大佬回复,作者要不啥时候搞个直播或者在b站上发个视频讲解呗,我看别的issue大家对细节的关注还挺多的

请问下两个阶段ablation的细节

thx the author and I close the issue

Question about the cross-antropy loss average?

I'm not the author but you can try the entropy loss in CodeFormer

为什么要用在低清图上finetune的encoder?

sry我好像记错了,finetune的是supir这篇文章.....

推理大一点的图片非常慢，请问作者有推出小模型的计划吗

你问了一个现在全世界都很难解决的问题...