Unet-Segmentation-Pytorch-Nest-of-Unets
Unet-Segmentation-Pytorch-Nest-of-Unets copied to clipboard
bigmb
choose network
Hello, I would like to ask a question, when I use unet, it is normal, but switch to other network, the loss of training set is decreased, but the loss of verification set has not decreased, and dice is 0
hello,I get the same problem,could you tell me how you solve this problem?
It can be due to model not learning anything while using any RNN type model. | Does this happen while using nested Unet?
Thanks for your reply. when I using unet and trans-unet model, it take such problem. how can I solve this problem?
Which model is trans-Unet model? Can you explain more about the data and transformation you are using?
I'm sorry,it should be attention unet. I found my error from data_transform,because when I using my own data,I changed this by mistake. Now the dice score isn't 0,but it still very low.