Originally Posted by
KyouKiSu
I got question - will be any ways to simulate toribashNext physics faster than in-game? Like get some results of replay using some libs, for example to teach neural networks, without need to launch render of objects, interface and other things.
You can do this already with the normal Toribash.
Have a look at this:
https://forum.toribash.com/showthread.php?t=618230
I think if you train a model right now you'll be able to adapt it for TBN in some way later on, not sure if lua scripting will be supported any further, is it?