Today's NVIDIA GTC 2020 Announcements

Parts 1 through 9 of the NVIDIA GTC 2020 KeyNote presentation are now available for your viewing pleasure.  Lots of fun new things to check out from NVIDIA.
Here's Part 1 of the 9 presentations. Note that the conference is happening remotely, as Jensen Huang gives his keynote talk from his home kitchen where he is for some reason wearing a black leather jacket.

So we recently discussed NVIDIA's Jetson Nano single board accelerators when we looked at embedded deployment solutions for deep learning neural nets. 

There is now a new Jetson Xavier NX module that is backwards compatible with Jetson Nano.
It includes the following features:
An integrated NVIDIA Volta 384-core Volta GPU with 48 Tensor Cores
Two NVIDIA Deep Learning Accelerator engines
Seven-way VLIW Vision Accelerator
Six-core NVIDIA Carmel 64-bit ARMv8.2 CPU
8-GB 128-bit LPDDR4x

Jetson Xavier now includes what NVIDIA calls 'cloud native technologies'.  So things like Docker based containerization and Kubernetes. The Jetson Xavier NX Developer kit is available now for $399. Only 5 per customer, so you will be constrained when you try to use those Kubernetes features to build your own custom GPU cluster.


New Application frameworks were also announced for 'conversational AI services' (fancy term for AI-Bot) and deep learning recommendation systems.  So AI systems that conversationally interact with people.


CUDA Toolkit 11 was also announced.  Here's a more detailed 'features revealed' post that talks about all of the new CUDA 11 features. Note conspicuous absence of Apple from the supported platforms table.


A new Expressive and Natural Speech Synthesis Model called Flowtron was also announced.  Flowtron does the voice narration for embedded commercial part of the keynote above (2.29 timestamp).  Here's a preprint paper that talks about the model.

 Flowtron learns an invertible mapping of data to a latent space that can be manipulated to control many aspects of speech synthesis (pitch, tone, speech rate, cadence, accent). You also have control over speech variation and style transfer.

Comments

Post a Comment

Popular posts from this blog

CycleGAN: a GAN architecture for learning unpaired image to image transformations

Image
 Like the Doublemint Twins touting the joys of Doublemint gum, 2 GANs are surely better than 1 GAN.  Especially if we package them together inside of one meta GAN module.  And this is exactly what the CycleGAN architecture does. Have you ever harbored dark secrets of turning a horse into a zebra?  The CycleGAN was developed to do just that. Learn how to turn a horse into a zebra.  And more. Now right away you can notice a difference between the image to image transformation GAN architectures we've been discussing over the last few posts.  Those last few posts described systems that learn from a database of matched input-output image pairs. And if your goal is to turn an edge representation into a nicely filled in continuous tone image, it's easy to build your database of matched input-output image pairs that your GAN system can then learn off of.  Take a continuous tone image (which will be the output of the database pair entry), then run it through an edge detector algorithm. 
Read more

Pix2Pix: a GAN architecture for image to image transformation

Image
 I thought following up yesterday's TraVelGAN post with a Pix2Pix GAN post would be useful to compare what is going on in the 2 architectures.  Two different approaches to the same problem. I stole this Pix2Pix Overview slide below from an excellent deeplearning.ai GAN course (note that they borrowed it from the original paper) because it gives you a good feel for what is going on inside of the Pix2Pix architecture.    Note how the Generator part is very much like an auto-encoder architecture, but rebuilt using the U-Net architecture features (based on skip-connections) that fastai has been discussing in their courses for several years before it became more widely known to the deep learning community at large  (and which originally came from an obscure medical image segmentation paper) . So the Generator in this Pix2Pix GAN is really pretty sophisticated, consisting of a whole image to image auto-encoder network with U-Net skip connections to generate better image quality at highe
Read more

Smart Fabrics

Image
"What makes smart fabrics revolutionary is that they have the ability to do many things that traditional fabrics cannot, including communicate, transform, conduct energy and even grow."  This is a quote from Rebeccah Pailes-Friedman, who is a professor at Pratt with a research focus on developing wearable technology and smart textiles. Enhancing textiles can be viewed through different lens.  Enhancements can be aesthetic, or performance boosting, or durability boosting. 
Read more