Sketch of dim-ed tensors #407
Conversation
|
Added a test |
| return xtensor_constant(x, **kwargs) | ||
|
|
||
|
|
||
| def as_xtensor_variable(x, name=None): |
There was a problem hiding this comment.
would this have dims?
There was a problem hiding this comment.
This is analogous to as_tensor_variable which doesn't allow you to pass constructor information. Right now only accepts xarrays or things that are already xtensor variables
There was a problem hiding this comment.
There is an Op xtensor_from_tensor that can be used to add dims to a vanilla tensor that only has shape
| return xtensor_constant(x, **kwargs) | ||
|
|
||
|
|
||
| def as_xtensor_variable(x, name=None): |
There was a problem hiding this comment.
would this have dims?
as_xtensor_variable([1, 2, 3], dims=("channel", ))There was a problem hiding this comment.
Yes we can allow that
ricardoV94
force-pushed
the
named_tensor
branch
3 times, most recently
from
592c699 to
a3db755
Compare
ricardoV94
force-pushed
the
named_tensor
branch
2 times, most recently
from
2d908df to
6e88797
Compare
ricardoV94
added
enhancement
ricardoV94
force-pushed
the
named_tensor
branch
6 times, most recently
from
2742594 to
71d31ba
Compare
Co-authored-by: Oriol Abril-Pla <[email protected]>
|
Is this the same checklist as a new backend? |
Slightly different, why? |
|
Curious |
|
@OriolAbril I moved to #1411 a pytensor branch so that PRs show up here instead of on my fork |
Moved to # #1411