Re: Battle Field audiogame
hi,
talking about data science, (you are somehow write about computation on the gpu).
libraries like pytorch, mxnet, tensorflow etc support nvidia graphic cards, but for a simple linear model (things like decision trees, linear regression etc) really don't need that much gpu (scikit-learn uses numpy which doesn't support computation on the gpu at all).
also, for some kinds of neural networks (things like Dense layours, simple convolution networks etc), you don't need gpu computation (though I use gpu for it as well).
for loading the data and batchify it, you also don't need gpu (but for the computation you might).
don't think I'm against gpu, I'm saying that all of the ml algorithms don't need gpu, but for deep learning you probably need it.
and execution with a natively compiled language like C++ can't help you as well in the computation because it is computation expensive (consider that the computer want's to learn many kinds of sequence features, and do a backprop out of it and optimize the neural network's weights).
also, many of the deep learning algorithms require gpu, otherwise you should be alive maybe for a thousand years for it to complete the training process.
also for prediction, you don't need gpu because prediction is way way faster than training loop.
now for audio games, you are write. although python is more capable than development of audio games (even in it's direct form without using any natively compiled module like panda3d or pygame, but i advise you to use them)
thats write as well that python is slower to natively compiled languages, but I have to say that python is an interpreter not a compiler
-- Audiogames-reflector mailing list Audiogames-reflector@sabahattin-gucukoglu.com https://sabahattin-gucukoglu.com/cgi-bin/mailman/listinfo/audiogames-reflector
