search

zhouhaoyi / Informer2020

The GitHub repository for the paper "Informer" accepted by AAAI 2021.
Apache License 2.0
5.28k stars 1.1k forks source link

关于预测结果里的real_prediction.npy里的数据是标准化的，我要如何反标准化？（Regarding the real in the prediction results_ The data in prediction.npy is standardized. How can I de standardize it?） #516

Closed 2921113436 closed 1 year ago

2921113436 commented 1 year ago

哈哈，我懂了，属于自问自答了，希望帮到有缘人。parser.add_argument('--inverse', action='store_true', help='inverse output data', default=False)将这行代码的default=False改为True

luxury-jr commented 1 year ago

感恩

Med-Rokaimi commented 1 year ago

哈哈，我懂了，属于自问自答了，希望帮到有缘人。parser.add_argument('--inverse', action='store_true', help='inverse output data', default=False)将这行代码的default=False改为True

Hey, thank you. I want to inverse that ones generated at the training process in preds.npy and trues.npy files. When I use this paramater it pass the data with normalization to the trainin.

at915 commented 1 year ago

十分感谢

at915 commented 1 year ago

改完后mse值会变得很大，请问是什么问题？

MountVoom commented 1 year ago

改完后mse值会变得很大，请问是什么问题？

因为反归一化后，mse就使用真实值在计算，和真实值本身的大小有很大关系。

Seb5Vet commented 1 year ago

改完后mse值会变得很大，请问是什么问题？

我的也是这个问题，请问解决了吗