How to download datasets from the website?

[JianhuiWei7]

I can not open the website using Browser

[XuanheZhou]

The link seems fine on our computers (with browsers like google). Try it again or download it via curl?

[JianhuiWei7]

I try to open it with Chrome, failed, and try to curl it, and it only returns the html tags. I think maybe the website is restricted to tsinghua students because it is tsinghua cloud. Is it possible if we can apply another channel like Baidu Cloud or others?

[m90113808]

I tried using the data from the link (https://github.com/TsinghuaDatabaseGroup/AI4DBCode/tree/master/PerformancePredictor/pmodel_data/job), which allowed the code to run without errors. However, when I tried to add a breakpoint at line 44 of the file train.py to observe the model's prediction results, I found that after 80 epochs of training, the prediction values for different nodes became increasingly uniform, even though the labels for them had significant differences. I think this may indicate that the model has not learned knowledge related to node performance. Is this behavior normal? Alternatively, could you please provide the complete data and scripts that can reproduce the results mentioned in the paper? @zhouxh19

[XuanheZhou]

Sorry for the confusion. I will check the problem soon. Stay tuned🙏