Xiaojing Zhang

Thank you for your interest in HpyerTS. Currently, the built-in DeepAR only supports univariate (with covariates) time forecasting task.

Thanks for your advice, and now hyperts has the API documentation: https://hyperts.readthedocs.io/en/latest/contents/0900_api.html.

Have you installed TensorFlow?

> Does this package support time series forecasting of intermittent or block demand? Do you mean the time series is not continuous?

> For intermittent or block time series, HyperTS will impute missing values during feature preprocessing.

From version 0.1.4, ``model.get_params`` has been changed to ``model.get_pipeline_params()``，thank you for your warning.

Hello, there are some with saving the model by `pickle`, but which is being solved. At the same time, if you have a good plan, welcome to submit `Pull requests`...

Thank you for your interest in HpyerTS. If you only want to train once without performing the search, you can do it in one of two ways: 1、Set `max_trials=1`, `final_retrain_on_wholedata=False`...

您好，您的数据格式转换成panel形式了吗？ 数据格式：https://hyperts.readthedocs.io/en/latest/contents/0300_dataformat.html

从错误看，网络输入层未识别出您传入了29维特征的数据，识别成了6维，您可以详细说一下您的数据吗？