-
Notifications
You must be signed in to change notification settings - Fork 43
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
add support for binary parameters #454
Conversation
This pull request was exported from Phabricator. Differential Revision: D65954134 |
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Differential Revision: D65954134
52ab256
to
bf2275a
Compare
This pull request was exported from Phabricator. Differential Revision: D65954134 |
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Differential Revision: D65954134
bf2275a
to
61f04a2
Compare
This pull request was exported from Phabricator. Differential Revision: D65954134 |
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Differential Revision: D65954134
61f04a2
to
6c0d957
Compare
This pull request was exported from Phabricator. Differential Revision: D65954134 |
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Differential Revision: D65954134
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Differential Revision: D65954134
6c0d957
to
8ed95b4
Compare
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Differential Revision: D65954134
This pull request was exported from Phabricator. Differential Revision: D65954134 |
8ed95b4
to
034b1b7
Compare
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Differential Revision: D65954134
This pull request was exported from Phabricator. Differential Revision: D65954134 |
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Differential Revision: D65954134
034b1b7
to
7127015
Compare
This pull request was exported from Phabricator. Differential Revision: D65954134 |
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Reviewed By: crasanders Differential Revision: D65954134
7127015
to
f69a76d
Compare
This pull request was exported from Phabricator. Differential Revision: D65954134 |
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Reviewed By: crasanders Differential Revision: D65954134
f69a76d
to
d0675de
Compare
This pull request was exported from Phabricator. Differential Revision: D65954134 |
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Reviewed By: crasanders Differential Revision: D65954134
d0675de
to
b82bfe5
Compare
This pull request was exported from Phabricator. Differential Revision: D65954134 |
Summary: Discrete parameter support added via rounding transform. Reviewed By: crasanders Differential Revision: D65699942
Summary: Transforming bounds requires additional logic that used to be part of ParameterTransforms, we move these to the parameters itself and have the ParameterTransforms look for these special methods when transforming bounds. We add a new ABC for our transforms as going forward it is likely that all of our transforms will have unique capabilities over the BoTorch base. This includes how we handle some finding options from configs. Reviewed By: crasanders Differential Revision: D65897908
) Summary: parameters.py was getting too big. Transforms themselves were moved to ops.py, leaving the base class and wrapper classes in parameters.py. Generic utility function that could be used elsewhere moved to base utils.py New parameter handling overwrites old ax support, so we remove ax related functions/tests Reviewed By: crasanders Differential Revision: D65898366
… tensors (facebookresearch#455) Summary: This removes the numpy dependency from the manual generators and changes the logic slightly so that we can support higher dimensional tensors of points, allowing us to support pairwise experiments Reviewed By: crasanders Differential Revision: D64607853
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Reviewed By: crasanders Differential Revision: D65954134
b82bfe5
to
22330c4
Compare
This pull request was exported from Phabricator. Differential Revision: D65954134 |
This pull request has been merged in 2ba8445. |
Summary: Pull Request resolved: #454 Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling. Reviewed By: crasanders Differential Revision: D65954134 fbshipit-source-id: 0afb0bd9639f26ccb965bbe7880ca5e6576a3b12
Summary: Binary parameters are just secretly discrete parameters bounded at [0, 1]. Config will accept binary as a par_type and do the necessary work to support it in modeling.
Differential Revision: D65954134