This repository was archived by the owner on Jun 2, 2023. It is now read-only.

Description
I tried using the log target variable option and found that xr.ufuncs.log is now outdated and we should switch to np.log, as documented here:
Line to change.
Also, I think the val and tst data need to be logged here. Why are only the training data logged? The mean and standard deviation are computed on the logged data and then applied to the val and tst data, which are not logged. Some comments in the preproc_utils need to be updated to describe what log_y_vars and log_vars should be.
And we should subtract off the small constant after transforming from log to real space here.
I can address these edits in a PR