Jing He
Results
4
issues of
Jing He
您好,感谢您的开源代码。 在Readme中您有说明,生成高分辨率时的训练流程是32->64->128->256, 每次训练都基于前一分辨率得到的model进行finetune。 这样的训练策略的确会比直接训练要容易得多,那请问您试过直接训练256分辨率吗,调整训练参数是否也能得到类似的效果?
Hi Professor,thank you for sharing the code! I have extracted features by using the pretrained model and matched them following [this issue](https://github.com/cvlab-epfl/tf-lift/issues/22) ,but the matching result seems not good enough,...
Hi, Thanks for your sharing the code. May I ask you about how to visualize the segmentation/surface normal/depth maps? Like this:
Hi! Thanks the authers sharing the great work! As you mentioned in the paper Sec. 5.1, May I ask which "Stable Diffusion Image Variations Model" you fine-tuned? Could you provide...