Tensor Train Optimiser #31
Conversation
|
|
VolodyaCO
requested review from
AthenaCaesura
and removed request for
mstechly
|
|
||
| from orquestra.opt.api.functions import CallableWithGradient | ||
| from orquestra.opt.optimizers.pso.continuous_pso_optimizer import ( | ||
| Bounds, # TODO: where should these Bounds live? |
There was a problem hiding this comment.
Are these bounds specific to the pso optimizer? if so then I think they're living in the correct place, we might just want to change the name to PSOBounds or something like that. Either way, we should take care of this to avoid having TODO's in our product code. 😆
There was a problem hiding this comment.
I think they are also used here in TTOpt, so I'd think this Bounds class should be in a higher level of the directory tree, but am unsure where they should be.
There was a problem hiding this comment.
Perhaps the bounds type could live in the __init__.py? Seems like a decent place for it.
|
Sorry I left this here for so long! Getting back on it now! |
|
|
||
| from orquestra.opt.api.functions import CallableWithGradient | ||
| from orquestra.opt.optimizers.pso.continuous_pso_optimizer import ( | ||
| Bounds, # TODO: where should these Bounds live? |
There was a problem hiding this comment.
Perhaps the bounds type could live in the __init__.py? Seems like a decent place for it.
|
ok! Once tests pass you should be good to merge! |
Co-authored-by: Athena Caesura <athena.caesura@zapatacomputing.com>
VolodyaCO
force-pushed
the
feature/volodyaco/ttopt
branch
from
15a117b
to
2796ee9
Compare
…mputing/orquestra-opt into feature/volodyaco/ttopt
…ctionWithBestCandidates
Description
Recently, @ArtemStrashko gave a talk on a tensor train optimiser for continuous functions. There's a library out there for it. This PR wraps this library to expose TTOpt as an optimiser.
Please verify that you have completed the following steps