Facebook Trains ImageNet in 1 Hour

Research, Cluster / Supercomputing, Image Recognition, Machine Learning & Artificial Intelligence, Media & Entertainment, Tesla

Nadeem Mohammad, posted Jun 08 2017

Facebook published a paper today detailing how they are able to train nearly 1.3 million images in under an hour using 256 Tesla P100 GPUs that previously took days on a single system.      The team reduced the training time of a ResNet-50 deep learning model on ImageNet from 29 hours to one – which

Read more

AI Turns UI Designs Into Code

Research, CUDA, cuDNN, Image Recognition, Machine Learning & Artificial Intelligence, Media & Entertainment, Tesla

Nadeem Mohammad, posted Jun 05 2017

Copenhagen-based startup UIzard Technologies trained a neural network to automatically generate code from a graphical user interface screenshot. Currently, a UI designer mocks up the interface and then the front-end developer takes the design and translates it into code – but, what if an AI system can do this for you? The company’s slogan is

Read more

Developer Voices

Research, CUDA, cuDNN, Developer Voices, GeForce, Jetson

Nadeem Mohammad, posted Jun 02 2017

We love seeing all of the NVIDIA GPU-related tweets – here’s some that we came across this week: How many Jetsons are too many Jetsons? Correct ans = number of Jetsons you own + 1. @nvidia @nvidiadeveloper @GPUComputing #DeepLearning pic.twitter.com/7ALPHBWqq0 — Shreyas Skandan (@ShreyasSkandan) May 27, 2017 Our mission is to make #healthcare accessible and

Read more

GPUs and Deep Learning Could Help Control Robots in Space

News, Research, Aerospace, Astronomy & Astrophysics, cuDNN, Defense, Energy/Oil & Gas, GeForce, Higher Education / Academia, Image Recognition, Machine Learning & Artificial Intelligence

Nadeem Mohammad, posted Jun 01 2017

A six-person startup from Seattle developed augmented telerobotics software that gives humans better control of remotely operated robots which can be useful for exploring Mars or other planets. BluHaptics specializes in robotic control for underwater environments, but with a recently awarded grant funded by NASA, they are now applying their software to control robotic operations

Read more

Microsoft Releases New Version of High-Performance, Open-Source, Deep Learning Toolkit

News, Research, Higher Education / Academia, Image Recognition, Machine Learning & Artificial Intelligence, Volta

Nadeem Mohammad, posted Jun 01 2017

Previously known as CNTK, the Microsoft Cognitive Toolkit version 2.0 allows developers to create, train, and evaluate their own neural networks that can scale across multiple GPUs and multiple machines on massive data sets. The open-source toolkit, available on GitHub, offers hundreds of new features, performance improvements and fixes that have been added since the

Read more