This message was deleted.
# neural-forecasts
Slackbot
01/22/2024, 10:27 AMThis message was deleted.
a
Afiq Johari
01/22/2024, 10:28 AMI think it's because the train dataframe has no 'ds' column, that should be the date column
d
Doyel De Sarkar
01/22/2024, 10:32 AMI have the ds column.. but the dayatype is int..do that mean I have to convert it in date type...
Now I tried to convert it into date object and ran again same error..
c
Cristian (Nixtla)
01/22/2024, 2:58 PMHi @Doyel De Sarkar! There might be an issue with your input data. Do you have the unique_id, ds and y columns?
Cristian (Nixtla)
01/22/2024, 2:59 PMCan you try your data on the normal LSTM as well?
d
Doyel De Sarkar
01/22/2024, 3:20 PMHey Cristian...thanks for replying...normal LSTM is running fine.. previously I was trying to fit the AutoLSTM directly.. but later I tried to use neuralforecast class to train the auto model as mentioned in the example of Auto NHITS...then it ran fine.. please let me know if I am doing wrong..
b
Bryce Davies
02/14/2024, 10:13 AMHey @Doyel De Sarkar did you manage to get this resolved at all? I'm running into a similar issue, my dataframe has the 'ds' column, trying to do model.fit for AutoNHITS