News for hackers. Stuff that matters.
Re:FluidNet – Accelerating Eulerian Fluid Simulation with Co... by londons_explore on Wednesday November 23, 2016 @05:58AM (#13021641)
Even this isn't true. The model is fully convolutional, so is independent of location within the studio (as physics should be).

I expect it would need retraining for different material properties (eg. different densities).

There are no comments on this comment yet.