https://github.com/nixtla logo
#general
Title
# general
a

Ajay Bhatt

08/01/2022, 12:15 PM
Hi Everyone! I am trying to train the MQNHITS model on my monthly time-series data whose structure is similar to that of M4 monthly time-series data I tried all the notebooks which are available on GitHub and did all the required configurations(“H” to “M”) but none of them is giving results. the training is automatically stopping in between as there is no proper documentation available please help. here is the screenshot of my data.
👀 1
f

fede (nixtla) (they/them)

08/01/2022, 6:58 PM
Hi @Ajay Bhatt! For the moment, you can’t use integer datestamps to produce forecasts. If you have a map for each int to a monthly date, then you will be able to produce forecasts.
m

Max (Nixtla)

08/02/2022, 12:27 AM
If don’t have a real mapping you could obviously use one artificial one. e.g. 1 -> 01/01/1900 2 -> 02/01/1900 …
a

Ajay Bhatt

08/02/2022, 3:23 AM
@Max (Nixtla) @fede (nixtla) (they/them) yes, I have the mapping of each int to the given datestamp, I convert the datestamp to the numbers so that its structure will be the same as that of M4 but still it’s not working.
m

Max (Nixtla)

08/02/2022, 2:49 PM
I'm very sorry this is not working. Could you share a little bit more info on the error or maybe even share some code so we can reproduce the error ?
a

Ajay Bhatt

08/03/2022, 9:00 AM
@Max (Nixtla) training is not working it is stopping in between although the data is correct, I have shared the data and the notebook with u in DM please check
7 Views