-
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 fixed parameters #457
Conversation
This pull request was exported from Phabricator. Differential Revision: D66012863 |
bfd9e3e
to
e0b6c77
Compare
Summary: Pull Request resolved: facebookresearch#457 Add support for fixed parameters. These parameters are just set in the config and will be removed before any model or generator is aware of them (then added back in whenever a model or generator is asked for an output. Differential Revision: D66012863
This pull request was exported from Phabricator. Differential Revision: D66012863 |
Summary: Pull Request resolved: facebookresearch#457 Add support for fixed parameters. These parameters are just set in the config and will be removed before any model or generator is aware of them (then added back in whenever a model or generator is asked for an output. Differential Revision: D66012863
e0b6c77
to
6237327
Compare
This pull request was exported from Phabricator. Differential Revision: D66012863 |
1 similar comment
This pull request was exported from Phabricator. Differential Revision: D66012863 |
Summary: Pull Request resolved: facebookresearch#457 Add support for fixed parameters. These parameters are just set in the config and will be removed before any model or generator is aware of them (then added back in whenever a model or generator is asked for an output. Differential Revision: D66012863
6237327
to
c3a1c49
Compare
Summary: Pull Request resolved: facebookresearch#445 Discrete parameter support added via rounding transform. Differential Revision: D65699942
…okresearch#452) Summary: Pull Request resolved: facebookresearch#452 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. Differential Revision: D65897908
c3a1c49
to
1c6acd6
Compare
Summary: Pull Request resolved: facebookresearch#457 Add support for fixed parameters. These parameters are just set in the config and will be removed before any model or generator is aware of them (then added back in whenever a model or generator is asked for an output. Reviewed By: crasanders Differential Revision: D66012863
This pull request was exported from Phabricator. Differential Revision: D66012863 |
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 Differential Revision: D65898366
… tensors 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 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. Differential Revision: D65954134
Summary: Add new parameter error that only is ever raised in config to allow meaningful error handling when parameter options aren't set correctly. ValueError for missing bounds/parameter names is still there to handle fallbacks but incorrectly set options will throw the new error. Differential Revision: D65957107
This pull request was exported from Phabricator. Differential Revision: D66012863 |
Summary: Pull Request resolved: facebookresearch#457 Add support for fixed parameters. These parameters are just set in the config and will be removed before any model or generator is aware of them (then added back in whenever a model or generator is asked for an output. Reviewed By: crasanders Differential Revision: D66012863
1c6acd6
to
0711018
Compare
Summary: Pull Request resolved: facebookresearch#457 Add support for fixed parameters. These parameters are just set in the config and will be removed before any model or generator is aware of them (then added back in whenever a model or generator is asked for an output. Reviewed By: crasanders Differential Revision: D66012863
0711018
to
589d95b
Compare
This pull request was exported from Phabricator. Differential Revision: D66012863 |
This pull request has been merged in deb8ce1. |
Summary: Pull Request resolved: #457 Add support for fixed parameters. These parameters are just set in the config and will be removed before any model or generator is aware of them (then added back in whenever a model or generator is asked for an output. Reviewed By: crasanders Differential Revision: D66012863 fbshipit-source-id: 9d5720b2f3acaf547c18269d762808dccc7a3383
Summary: Add support for fixed parameters. These parameters are just set in the config and will be removed before any model or generator is aware of them (then added back in whenever a model or generator is asked for an output.
Differential Revision: D66012863