menghanxia/disentangledcolorization

A colorization framework that disentangles the color multimodality and the structural consistency via adaptively located anchors, so that both aspects can be achieved effectively. [SIGGRAPH Asia 2022]

disentangledcolorization's Issues

Training issue: nan occurs after totalLoss_idx.backward()

Hi,
I am trying to finetune DISCO model on other color image datasets. The training settings are kept as default as on ./scripts/anchorcolorprob_hint2class-enhanced-h8.sh. After several loops, "nan" value occurs and the training losses become nan too. I checked the intermediate variables. The "nan" starts after the execution of "totalLoss_idx.backward()". Would you please give me some suggestion? This error usually caused by gradient explosion or gradient vanishing. I tried gradient clipping using torch.nn.utils.clip_grad_norm(model.parameters(), args.clip), and didn't work. Looking forward to your reply.

Best.

請教如何針對高解析度照片上色?

你好
我嘗試在disco平台上，將一張1920 x 1080 大小的黑白照片上色，卻一直無法成功
是不是該平台目前最高只能設定在512 x 512?

另外如果想在本地上訓練+上色，是否有規定照片大小?
電腦規格為win10 i9-12900K RTX3080
謝謝

metric is not available

Thank you for your inspiring work.

The link you provided for metric evaluation is dead.

Why AB channel in CIE Lab space scaled by 110

Hi,
Wonderful work!
As stated on L.35 in ./utils/dataset_lab.py, the ab channels are scaled by 110. instead of 128. Would you please explane the reason? In OPENCV, LAB space should be defined as L -> 0100, a -> -128+128, b-> -128~+128. Thanks again for your wonderful work.

How to use model “SPixelNet-s16” in inference scripts?

thank you for your awesome work!
I felt confused about superpixel modle “SPixelNet-s16” and "SPixelNet-s8".
how to use it if i want to test colorize and superpixel performance?