Closed weatherforecasterwhai closed 1 month ago
I tried several times, only 1G data, for most of times, .compute() is really slow, but with no errors.
Today, I got this error again and again. Would you please help me with this:
Traceback (most recent call last):
File "/share/home/njhpc/Wanghai/neuralgcm0805/neuralgcm_new.py", line 80, in
Maybe I was wrong. My local server computer is banned from gcs.
Hi, I can't .compute() .zarr like months ago. It even can't compute() out in one night. Like #112, and https://github.com/google-research/arco-era5/issues/67.
So, I try to download ERA5 .netcdf data first, and then change to .zarr format for neuralgcm input. However, the format always send errors. 1.Do you have better suggestions for changing .netcdf to .zarr for neuralgcm input? 2.Since .netcdf is so popular in weather forecasting area. Will you make a .netcdf input demo in the future? Or, please make the .compute() easier like before. Thank you!