Solver satisfying Dale's principle. #70
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
hunse
reviewed
Mar 28, 2018
| class DalesL2(Dales): | ||
| """Solves for weights subject to Dale's principle with regularisation.""" | ||
|
|
||
| def __call__(self, A, Y, rng=None, E=None, sigma=0.): |
There was a problem hiding this comment.
sigma shouldn't be a parameter that's passed in here. Rather, the instance should have a reg parameter, which is then used to compute sigma using the max value in A. Just like in LstsqL2.
hunse
reviewed
Mar 28, 2018
|
|
||
| # assert that weights themselves are close (this is true for L2 weights) | ||
|
|
||
| assert np.allclose(W1, W2) |
There was a problem hiding this comment.
You might want to check that the signs of the weights are as requested (i.e. the correct number positive/negative).
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Based off the code described in this Nengo issue. Equivalent in effect to the Parisien transform, but has better performance.
Still to do:
nnlssolver in SciPy has problems andcvxoptshould be used instead, as shown in this forum post.