i swear im not trying to shill.
honestly though, why havent you bought a Titan V so far if it effectively turns your PC into a supercomputer?
it costs $3000 but considering the speed at which it can do computations for (machine learning) is it not completely worth the money?
Because I don't do machine learning, and I'm not really interested in starting?
Because the average person does not need, say, a 1080ti much less a titan V. Unless you do machine learning, your an idiot for spending 2 pc's worth of cash on 1 way to overpowered gpu.
ill be honest; i havent started learning it either but realistically speaking it will be the driving force behind all industry moving forward.
no i agree completely; its not something i think that should be used for gaming.
but everyone should be doing some form of machine learning, wouldnt you agree? lest they end up completely fucked as automation consumes 80% of the job market
Cause I just bought a 1080ti for my new build...
it costs $3000.
wouldnt the benefit of machine learning outweigh that cost hand over fist?
>why havent you bought a Titan V so far if it effectively turns your PC into a supercomputer
>it costs $3000
Poorfag reporting in.
Is it the future, absolutely. People should put way more focus into Machine learning because its the inevitable next step, But at the very moment, technology such as the titan V is way too inaccessible for most, even those who want to learn ML, even in a university. I would say unless your in MIT, or an Ivy league, for the most part your shit out of luck and you'll be using older tech. I mean in the uni where I intern, one of the PC's in the lab still uses a fucking Core 2 Duo with some shit AMD gpu from the mid 2000's. And everything else is a standard windows 7 with an equally mediocre gpu/sandy bridge cpu's.
So would it be worth it, yes. But I feel this industry isn't going to take off until it is at a price where I could convince a university to spend cash on at least a few for labs.
>I feel this industry isn't going to take off
i mean the industry is primed to take off no matter what?
like this isnt even a case of "VR will become the new defacto standard and is primed to take off soon" when that is so uncertain.
but machine learning will be the driving force behind industry no matter what right?
>why havent you bought a Titan V
But I did?
Is machine learning a meme? I've only time for a couple more math courses I can take as an undergrad, so would that be better spent with advanced combinatorics or convex optimization (which is apparently the crux of machine learning)? Am I actually going to use it without going for a masters?
with the added graphics processing power? it could help if you already have something going that depends on such a thing, like video editing, 3D graphics, or mining shitcoins. I don't know much about machine learning's applications so I'm not sure how a single workstation with the card would be much use there.
good man. are there any models you are training atm that you are particularly proud of/hopeful for?
it is absolutely not a meme. it is literally without a doubt the future and there is no stopping it short of a giant natural catastrophe
Whan I said This industry wasn'nt going to take off, I meant to say it going to be slow for now because of the price which only select people/groups will be able to afford now.
But undoubtedly eventually it will reach a point of affordability and performance to really kick-start a period of nearly exponential growth. Just a matter of when. The titan V is a nice sneak peak into what the future will offer though.
Not really. Going to start learning a bit about it though. Mostly just bought it because my bonus was pretty decent this year, and it's pretty decent for gaming. I figured if somebody could figure out how to utilize tensor cores as additional shaders in graphics processing, this card will just be a future proofing purchase. But in the mean time, I figured it could be a fun side hobby to learn deep learning.
Anyone have good resources for getting started on that?
I kinda just browse Sup Forums and youtube.
And for the occasional round of smash and cuphead my 1050 does fine.
you mean cuckhead
>Titan V in a non-HEDT shitbox
Shame.
Why would I want to contribute to nvidia being the only GPU company for machine learning ?
>he cant afford the best just because
lmao
i really like how OP has no fucking idea what machine learning is, but is ready to spend 3 grand on a GPU which helps facilitate "high end machine learning"
sometimes i genuinely think you people are so fucking stupid that you have to be trolling and i must be the idiot for biting the bait
>why havent you bought a Titan V
I'm a poorfag.
>i really like how OP has no fucking idea what machine learning is
what do you mean
I could get like 135TB worth of SAS drives with that money
What an awful fucking thread. You should feel bad, OP.
... You would buy a Titan V built for machine learning without learning the first thing about ML? People should only upgrade their hardware AFTER they've already tested their model on CPU and determined the speedup of purchasing something like this. If you don't do serious GPGPU, why the fuck would you ever purchase this?
Unlikely. About 4yrs ago (I am CS changed to computational molecular biology for masters), we only touched on some topics like SVM design and loss function models for different classifiers. We got to implement our own SVM and what not, but the idea is that in industry these concepts aren't particularly that important, more of it becomes problem modeling. Some ways I do hate ML due to training time and the amount of energy you have to go through to get (both in terms of environmental and work) - ML is not necessarily the best way to solve every problem.