LambdaLi

我已解决test_dice极低的问题 | I have resolved the problem that the test_dice is extremely low

10

  这个代码我跑了四次，最后发现是自己对题主的意思理解错误导致，然后我花了几天时间逐行阅读代码，希望我的遭遇能够帮助到各位：` ①不要预处理‘测试集(test)’！！！不要预处理‘测试集(test)’！！！不要预处理‘测试集(test)’！！！重要的事情说三遍，只需要预处理训练集(train)，dataset_lits_test.py会对测试(test)集图像进行处理。 ②前面前辈说到的ResUNet.py里135-161行的map里的upsample里的scale参数后两位减半。 ③把test.py第25行的注释给取消注释，该行的作用看注释的注释，做完后便！！！不！！！需要像前辈之前说的运行test.py时把②里说到的减半参数->进行恢复。 ------------------------------------------------------------------------------------------------------------------------------------------------- I ran this code four times and finally realized that it was due to my incorrect understanding of the author's meaning. I spent...

Hello, your work is amazing. I have a question about the alpha parameter of the cross attention and self attention fusion module in the decoder. It was 0.5 in version...

For 3D segmentation

1

Your work is breathtaking, I asked a naive question, can it be used in 3D image segmentation now? I only know 1D and 2D in your natten, are you still...

Hello Your work has amazed me, and I cannot express my admiration in a few words. While reading the source code, I noticed that you are using the v2 version...