Skip to content
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

Added i) Inverse Dirichlet Adaptive Loss and ii) Neural Tangent Kernel Loss Implementations #673

Open
wants to merge 9 commits into
base: master
Choose a base branch
from

Conversation

@sphinx-tech
Copy link
Author

Instead of using the Jacobians for K (as implemented) by the authors; I have used dot products because we were only concerned with the main diagonal entries of the K matrix for our implementation - in Algorithm 1 (Pg. 10)

@xtalax
Copy link
Member

xtalax commented Apr 1, 2023

Put Plots and Revise in to your v1.8 environment, and add using Plots, Revise to your juliarc

@sphinx-tech sphinx-tech mentioned this pull request Apr 1, 2023
@ChrisRackauckas
Copy link
Member

Note that for this there is starter code in #504 and #506. It doesn't make sense to do both in the same PR though.

@sphinx-tech sphinx-tech marked this pull request as draft April 3, 2023 05:10
@sphinx-tech sphinx-tech marked this pull request as ready for review April 3, 2023 05:23
@sphinx-tech
Copy link
Author

sphinx-tech commented Apr 3, 2023

@xtalax @ChrisRackauckas I have made some changes according to your suggestions.

  • Two independent commits have been made that deal with the different loss functions
  • The Inverse Dirichlet is performing well, but the loss in Neural Tangent kernel is high

Pls let me know of any improvements that can be made in the implementations.
Also any direction on how to improve the testing part would be very helpful !

@ChrisRackauckas
Copy link
Member

Rebase onto master.

@ChrisRackauckas
Copy link
Member

The next step would be to setup tutorials and benchmarks with these.

@ChrisRackauckas
Copy link
Member

Looks like tests failed.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants