Gb/hr exo #100
Gb/hr exo #100
Conversation
grantbuster
force-pushed
the
gb/hr_exo
branch
from
4fcd4f8
to
34b2882
Compare
| t_exogenous = None | ||
| if exogenous_data is not None: | ||
| s_exogenous = exogenous_data[:len(self.spatial_models)] |
There was a problem hiding this comment.
maybe this is still worth keeping just for readability?
There was a problem hiding this comment.
Im trying to avoid the variable-to-variable pointer here: s_exogenous = exogenous_data which always seems kind of silly. Is it worth the readability? I could go either way.
There was a problem hiding this comment.
Oh yeah thats fair. I could go either way also. Could just pass exo_data[:len(self.spatial_models)] directly into generate.
There was a problem hiding this comment.
Yeah but the reason im not doing that now is so that the spatial models can have access to ALL of the exo_data including the 4km data that is technically assigned to the temporal model.
There was a problem hiding this comment.
Oh yeah I suppose that spatial models need up to len(self.spatial_models) + 1 if theyre using the concat or add layers.
|
Thanks! yeah the key was realizing the fwp can handle multiple exo layers up to a pretty high resolution and then the multi step model can decide on how to distribute those layers |
grantbuster
force-pushed
the
gb/hr_exo
branch
2 times, most recently
from
b482020
to
d8a5e3d
Compare
…h multiple exo data resolutions
grantbuster
force-pushed
the
gb/hr_exo
branch
from
d8a5e3d
to
6dc8b20
Compare
No description provided.