Some questions about the downscaling results from Climax downscaling.

[Teenye]

I followed the code and the description in the paper, using the downscaling code, but I was unable to reproduce the metric results for the five variables in the article (z500, t850, u10, v10, t2m). May I ask if you can provide a checkpoint for me to verify this result?

[Teenye]

Another question is, I want to know if there are only these five input and output variables when downscaling, or if there are other variables involved.

[Teenye]

I tried to conduct a downscaling experiment using these five variables. Regardless of whether a pre trained model was used or not, the lrmse of t2m was only about 2.94, and z500 was around 821.43.