Octane - Deep Learning Edition

Generic forum to discuss Octane Render, post ideas and suggest improvements.
Forum rules
Please add your OS and Hardware Configuration in your signature, it makes it easier for us to help you analyze problems. Example: Win 7 64 | Geforce GTX680 | i7 3770 | 16GB
Post Reply
User avatar
Tutor
Licensed Customer
Posts: 531
Joined: Tue Nov 20, 2012 2:57 pm
Location: Suburb of Birmingham, AL - Home of the Birmingham Civil Rights Institute

The theme of the GTC 2015 GPU Technology Conference was deep learning [ https://www.youtube.com/watch?feature=p ... E7qCs1q6JQ ]. We were given a glimpse of wonders spring from the massive parallel nature of GPU processing, ranging from highly accurate photo identification to the development of self-driving cars. If 3d application providers and Otoy began embedding deep learning into their applications, thereby harnessing the compute capabilities of your GPUs (that are like parallel processors on steroids), what would you want first and foremost that you believe might be derived from embedding deep learning abilities into your applications?
Because I have 180+ GPU processers in 16 tweaked/multiOS systems - Character limit prevents detailed stats.
User avatar
Goldorak
OctaneRender Team
Posts: 2321
Joined: Sun Apr 22, 2012 8:09 pm
Contact:

It's an area of great interest to us in the long term. We are entering very exciting times with GPUS powered NN's advancing things so quickly in just the last two years.
User avatar
glimpse
Licensed Customer
Posts: 3740
Joined: Wed Jan 26, 2011 2:17 pm
Contact:

Goldorak wrote:It's an area of great interest to us in the long term. We are entering very exciting times with GPUS powered NN's advancing things so quickly in just the last two years.
I think it would be a dream that new Octane would have learning abilities..let's say program would be able to pic Your style & then..after year of learning You just press on that "Make_Nice_Image button" & it magically OctaneRender-it-out =DDD..ok.. time to sleep..'cos I'm dreaming with eyes wide open =p
User avatar
Tutor
Licensed Customer
Posts: 531
Joined: Tue Nov 20, 2012 2:57 pm
Location: Suburb of Birmingham, AL - Home of the Birmingham Civil Rights Institute

In the area of 3d and 2d, and as I said above, I foresee a need for both application providers and Otoy to collaborate more closely before deep learning is to be fully embedded into most of our workflows. Currently, 3d and 2d applications create form, space, appearance, animation and rendering. Octane doesn't create geometric forms, space or animation, except to provide lighting, texturing, rendering and related effects. While what Octane now does can significantly affect how we perceive an image's form (it's appearance) and animation, if we want to model or rig or create particles, etc., we have to use an underlying 3d or 2d application. Thus, I foresee a need for close collaboration between application providers and Otoy, unless Octane is given ability to learn to create and manipulate the forms and spacial relationships found in OBJ files in what it perceives to be artistically pleasing manners, before we just press on that "Make_Nice_Image button" & magically render a nice image.
Because I have 180+ GPU processers in 16 tweaked/multiOS systems - Character limit prevents detailed stats.
Post Reply

Return to “General Discussion”