Kyung Sunggu

I am confused about class channel dimention.

5

I am confused about the class channel dimension. I look at the cityspace dataset mask, the shape is (H, W) and there is no class dimension. they mark the class...

I got curious while reading the paper. Why do we use just one positive sample?

1

I got curious while reading the paper. Why do we use just one positive sample? I mean we can generate more positive samples by using strong augmentations. I think many...

I've been wondering, why didn't you freeze the discriminator weights while using it in the generator loss calculation?

Question about skip-connection in Reconstruction Task

2

I think the role of skip-connection is rather disturbing in your paper where reconstruction tasks are usually performed, what do you think? Using skip connection in reconstruction often transfers a...

I already posted my problem before. but I still have a problem with the cuda memory issue. I tried to run your model called "CRNN" that is training CNN and...

gpu memory issue

1

I have a problem with the Encoder model. I wanted to make the CRNN from scratch so I build my custom CNN encoder model like "Efficientnet-b0", it is worked but...

error batch norm

2

When I enter inputs to the CNN model, there is an error in the batch norm layer because the forward() make a batch size of 1. How did you solve...

Why didn't you extract features for every 2d input image??? There is no for loop. I saw your code, and I'm sure you draw a feature on the 2d image...

Question about Auto-Encoder as Feature extractor

1

I read your paper but you said you used auto-encoder as a feature extractor. However, as I analyze your code, you use U-net architecture or V-net architecture. which one is...