How does chronos2 handle missing values in historical context?

[ChernovAndrey]

Hello!
I've undrerstood that chronos2 does not support imputation. But it does not throw an error, when I have nan values in historical context.

I found this remark in the paper:
"The mask is a binary indicator equal to 1 when the value is observed, and 0 otherwise. It serves two purposes: indicating which values are missing in the historical context and specifying which input dimensions correspond to future-known covariates. After construction of the mask, all missing values in u_d are replaced with zeros."

So the question: what is the best way to handle missing values? Should I interpolate it first or passing nan values is expected behaviour from a user?

Thanks in advance for your answer!

[shchur]

The recommended approach is to keep the missing values as NaN