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

Refactoring moscot to include moscot.neural #763

Open
5 tasks
MUCDK opened this issue Nov 7, 2024 · 0 comments
Open
5 tasks

Refactoring moscot to include moscot.neural #763

MUCDK opened this issue Nov 7, 2024 · 0 comments
Assignees

Comments

@MUCDK
Copy link
Collaborator

MUCDK commented Nov 7, 2024

We want to have a moscot.neural module which implements the discrete problems as equivalent neural problems. Therefore, we consecutively do the following steps:

  • remove Protocol classes Removing Protocol Classes #765
  • moscot.neural.base should have NeuralOTProblem (currently CondOTProblem, instead of OTProblem in discrete case), and the NeuralOutput (instead of the OTTOutput in discrete case), this can be adapted from the current implementation. Therefore, restructure the output into a BaseOutput which is the parent class of DiscreteOTOutput and NeuralOTOutput
  • moscot.neural.solvers in future times should include GenotLin, GenotQuad, and OTCFM, start with GenotLin, because it's implemented already, and extend it to GenotQuad. OTCFM will be added later.
  • Port the dataloader (or roughly the idea of the dataloader) from CellFlow, and extend it to include quadratic terms.
  • Let's also distinguish between the data representation which we compute the coupling in, and the one which are input/output to the flow
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

No branches or pull requests

2 participants