mingqi zhang

You can use function lovasz_softmax(), and set the classes=[1].

I have the same problem, have you solve it? @AlexCHEU

alpha本身像素间就有连续变化过程，用最近邻插值就会不平滑，可以把alpha可视化出来，线性插值会更平滑。

``` def warp(x, flo): """ warp an image/tensor (im2) back to im1, according to the optical flow x: [B, C, H, W] (im2) flo: [B, 2, H, W] flow """...

Hi, Could you share me the dataset link for me? Thank you! My email is [email protected]

@htzheng Thanks for your reply, there's another question. I try download data_large_panoptic.tar from Dropbox many times, but can not success, Only part of the whole dataset(Maybe it's too large? others...

@htzheng Hi，I finetune the co-mod-gan use object-aware masks with pretrain weights of G and D. The learning rate is set to 0.0001. I try to use ls gan loss or...

Have you solve this problem?

Maybe you can reference to https://github.com/huggingface/diffusers/issues/1829 , sd 1.5 is not work well in 3090/A10, some layers in unet can not backpropagation with xofrmer.

I have the same confusion about it,so i visualize the label that i can't tell what is it.