theory salon

« first day (2649 days earlier) ← previous day next day → last day (980 days later) »

6:06 AM

Hey man. Sorry I was busy with college apps over the weekend. I downloaded the dual.sphysics.org package, but I can't seem to figure out what exactly to import into blender. A guide said to go to File>User Preferences, but I am not seeing any user preferences menu on the mac version of blender 2.90

11 hours later…

vzn

4:48 PM

@AlanAbraham hi, college apps? you mean applications? have not used these pkgs myself...

vzn

4:59 PM

seems DualSPhysics is standalone. ie not blender related.

flipfluids is the blender plugin.

DualSPhysics tutorials off main pg dual.sphysics.org/index.php/tutorials

> If only accurate visuals are the main objective then from blender V2.82 and up has a new fluid, smoke and fire simulation engine called Mantaflow based on the FLIP solver.

acc to the blender post/ Q/A, mantaflow should be built into blender 2.9.

vzn

5:40 PM

VisualSPhysics is the DualSPhysics blender plugin github.com/EPhysLab-UVigo/VisualSPHysics

so there are quite a few options even without much looking. the simplest thing to do is go with straight Blender and see if the mantaflow engine is there & sufficient. suspect it is probably sufficient for lots of ideas...?

2 hours later…

Alan Abraham

7:51 PM

Ok, that makes more sense. Yeah, I was talking about college applications, since I am a senior...

1 hour later…

vzn

9:03 PM

@AlanAbraham so you like physics! :) that is significant esp given that the classes at your school are not ideal as you say. what brings/ brought you to it? what did you do in competition physics?

nbro

10:00 PM

Q: How can AI researchers avoid "overfitting" to commonly-used benchmarks as a community?

In fields such as Machine Learning, we typically (somewhat informally) say that we are overfitting if improve our performance on a training set at the cost of reduced performance on a test set / the true population from which data is sampled. More generally, in AI research, we often end up testi...