Roadmap/Releases

[LimitingFactor]

Hello,

This is an exciting project with lots of possibilities.

Is there a development roadmap or release schedule for this project?

Thank you

[ngeneva]

Hi @LimitingFactor

Thanks for your interest in Modulus, we understand that in recent months the team has not been that clear regarding development but with good reason.

Perhaps you saw the GTC announcement, but Modulus is now moving to open-source development to give max transparency to our users + a lot of new features/improvements! Although I cannot give specifics about roadmaps/release schedules right now, please know that the team is working hard to make this significant transition with official releases coming in the not-so-distant future.

Thanks!

Physics-Informed Machine Learning Platform NVIDIA Modulus Is Now Open Source...

[LimitingFactor]

Thank you. I hadn’t seen that.

[occams_laser]

Hi @ngeneva , is there a timeline for when will documentation for v23 be released? I use Modulus for some of my projects and am not sure if I should wait for v23 or continue using v22.09.

I am super happy to see the new version on public GitHub!

[LimitingFactor]

@ngeneva Thank you this is an exciting project. Is there a plan to release a beta version of pysdf as well? At the moment it isn’t possible to run the beta code without it when trying to import stl files. For example trying to run the aneurysm example case.

Thanks

[tsltaywb]

Hi,

I thought we have always been able to download the full source code of Modulus, so how is this different from the project going open source now?

Will development still continue as before with regular releases?

Thanks!

[ngeneva]

Hi @tsltaywb

The major change is a switch to Github to make the code more accessible as well as a re-architecture of the underlying code base to make it easier for both PyTorch experts as well as non-DL engineers to work with Modulus. Existing APIs are still supported and releases will still be happening.

More information coming very soon. Thanks!

[michaltakac]

Is it possible to combine new modulus / modulus-launch with examples for modulus-sym? modulus-sym is using Domain, Constraints and Solver, modulus-launch is a more standard PyTorch way of writing training code. How can I use both modulus / modulu-launch features and modulu-sym PDEs and constraints?