Alert: No strf_time_format given for ds. Provide strf_time format during "setup" for better results. Error in creating date time derived features. Continuing...

[axisrow]

I ran auto_timeseries locally on a MacBook Air M1 on the airpassengers time series, it doesn't work and gives errors:

Predicting using test dataframe shape = (12, 1) for ML model For large datasets: ML predictions will take time since it has to predict each row and use that for future predictions... Using given input: pandas dataframe... Alert: No strf_time_format given for ds. Provide strf_time format during "setup" for better results. Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... Error in creating date time derived features. Continuing... ML predictions completed

I read on the forum that I need to update the library, but I already downloaded the latest version 0.92. Then I installed it in Colab, according to the instructions. And exactly the same errors occur there. Here is a link to Colab

[AutoViML]

Thanks for sharing the Colab notebook. I tried running it but there are so many errors with dask versions, I couldn't install auto-ts at all.

This whole dask is a huge mess and I am seriously thinking of ripping it out of auto-ts.

What do you think?

Ram