Added docs for ROM, added docs for Cayley (and optimized routine), simplified calculation in attention layer. #76
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
…er in ROM context.
Codecov Report
@@ Coverage Diff @@
## main #76 +/- ##
==========================================
- Coverage 72.71% 72.02% -0.69%
==========================================
Files 93 94 +1
Lines 2294 2320 +26
==========================================
+ Hits 1668 1671 +3
- Misses 626 649 +23
... and 3 files with indirect coverage changes 📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
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Outdated suggestions cannot be applied.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.
This PR is a bit of a mess. I started working on different things (minor fixes) and thought it wouldn't be worth doing it on separate branches, but it ended up being bigger fixes and additions. Most of what's new here should affect the docs though, so I hope it's still possible to decipher what has changed. Short description:
kernels/inverses/inverse_kernels.jlwe made the computation of the new activation function slightly better by using@view.data_loader/data_loader.jlnow has a constructor that can be called with a single matrix (and relevant functions indata_loader/matrix_assign.jl). This is used for ROM autoencoders.symplectic_autoencoders) with a routine to generate data (generate_data.jl) and a routine for training (training.jl). The example is taken from 2112.10815.optimizers/manifold_related/retractions.jl) and wrote corresponding documentation.autoencoder.mdandsymplectic_autoencoder.md).To point 3.: this seems to be working really well. We easily get from a PSD error of 0.3 to 0.03 when using our symplectic autoencoders.