Open rydeveraumn opened 1 year ago
Hi @rydeveraumn ,
Unfortunately the 3D solver happens to be under private wraps. Sorry about that
Hey @aadityacs! That makes sense to me! I was just hoping I wasn't overlooking something. Thank you for the info. Would you be able to link me to some resources about how to tackle the 3D FEA? If you have time of course if not no worries.
Shameless self promotion: TopOpt.jl (https://github.com/JuliaTopOpt/TopOpt.jl) allows you to implement the TOuNN approach quite easily using any NN architecture from https://github.com/FluxML/Flux.jl. See an example doing something similar in https://github.com/JuliaTopOpt/TopOpt.jl/blob/master/test/examples/neural.jl. TopOpt.jl supports 2D and 3D domains.
Nice! I will take shameless self promotion as long as I can learn some cool new things! Thanks for the resources! Julia too that is nice! I've heard there are some pretty nice speed ups in comparison with Python.
Feel free to join our #topopt channel on the Julia slack https://julialang.org/slack/ and say hi. We meet every 2 weeks to discuss anything topopt in Julia, open invitation.
Sorry @aadityacs for hijacking this issue.
No worries @mohamed82008 i might join the slack and say hello to! Glad to learn from the discussion there
And @rydeveraumn if you are considering starting out do consider Julia too! Imo it's definitely friendlier too and you don't have to bother vectorizing everything like in python
Please do!
Hey @aadityacs! Nice to chat again :) I had a question do you have any repositories/work on thermal topology optimization?
Late reply @rydeveraumn but you can find some implementation here https://github.com/UW-ERSL/AuTO
Hey @aadityacs! This is some really cool work you have done here. Is there a way to run the 3D version that was explored in your paper? Thank you for the help.