jcruan519 / malunet Goto Github PK

Hello, take the liberty to interrupt, I downloaded the dataset isic2018 through the official website for the split is 2594 sheets Where should I get the extra 100 data from the 2694 in the paper?
If convenient, can you share your train.info .log?

在做对比实验时，没法达到和论文中对比试验的指标，可以麻烦发一下吗？谢谢！！！

Hello, I would like to ask how long it usually takes you to run 300 epochs on the ISIC2018 dataset. I downloaded the code to run it and found that each epoch took about half an hour. Is it normal?

Hello, how do you get the main and std values for image preprocessing？

How do I set up the code you provide if I just want to run the test collection?

As described in the paper, the dataset is randomly split in experiment.
Can you provide the list for a fair comparison?

I cannot able to access the dataset from the given link. Could you please provide a working link?

Peace be upon you
What are your thoughts when wanting to train and predict on large images such as 768*1024

ISIC2018，MALUNet跑出来的结果为
test of best model, loss: 0.2932,miou: 0.7998552734900135, f1_or_dsc: 0.8887995443534216, accuracy: 0.9469325188362953, specificity: 0.9712784932378719, sensitivity: 0.871260228505311, confusion_matrix: [[38912813 1150684]
[ 1659403 11230188]]

还有作者的EGE-Net实际测试结果为
ISIC2017，test of best model, loss: 0.7335,miou: 0.7868152686906821, f1_or_dsc: 0.8806901110345153, accuracy: 0.9593141526442308, specificity: 0.9718391644287886, sensitivity: 0.8970188284372141, confusion_matrix: [[34468587 998791]
[ 734361 6396661]]

作者能否提供下模型权重以供对比？感谢！

Hello, when I was training, I found that the gpu was not in use, but the cpu utilization rate was close to 100%. What should I do?