Blog

Jul 25, 2016   |   Scott Gray

Still not slowing down: Benchmarking optimized Winograd implementations

By: Scott Gray and Urs Köster This is part 3 of a series of posts on using the Winograd algorithm to make convolutional networks faster than ever before. In the second part we provided a  technical overview of how the algorithm works. Since the first set of Winograd kernels in neon, which we described in…

Read more

#neon

Jul 16, 2016   |   Hanlin Tang

Learn about neon™ with the Nervana Deep Learning Course

Intel Nervana is excited to share a series of short Nervana videos and accompanying exercises to learn how to build deep learning models with neon, our deep learning framework. We start with a basic introduction into deep learning concepts, provide an overview of the neon framework, and discuss key neon concepts such as loading data…

Read more

#neon

Jul 01, 2016   |   Jennifer Myers

neon v1.5 released!

We’re excited to release neon v1.5 with Python 2 and Python 3 support, support for Pascal GPUs (GTX 1080) and performance enhancements such as persistent RNN kernels (based on the paper by Greg Diamos at Baidu), bringing a 12x performance gain compared to v1.4.0. Highlights from this release include: Python2/Python3 compatibility [#191] Support for Pascal…

Read more

#Release Notes

Jun 29, 2016   |   Urs Köster

Going beyond full utilization: The inside scoop on Nervana's Winograd kernels

By: Urs Köster and Scott Gray This is part 2 of a series of posts on how Nervana uses the Winograd algorithm to make convolutional networks faster than ever before. In the first part we focused on benchmarks demonstrating a 2-3x algorithmic speedup. This part will get a bit more technical and dive into the guts of…

Read more

#neon

May 25, 2016   |   Aravind Kalaiah

Transfer learning using neon

Introduction In the last few years plenty of deep neural net (DNN) models have been made available for a variety of applications such as classification, image recognition and speech translation. Typically, each of these models are designed for a very specific purpose, but can be extended to novel use cases. For example, one can train…

Read more

#Model Zoo #Transfer Learning

Apr 30, 2016   |   Jennifer Myers

neon v1.4.0 released!

Highlights from this release include:  * VGG16 based Fast R-CNN model using winograd kernels * new, backward compatible, generic data loader * C3D video loader model trained on UCF101 dataset * Deep Dream example * make conv layer printout more informative [#222] * fix some examples to use new arg override capability * improve performance…

Read more

#Release Notes

Apr 27, 2016   |   JD Co-Reyes

Blog Post (Part III): Deep Reinforcement Learning with OpenAI Gym

This is part 3 of a blog series on deep reinforcement learning. See “Part 1: Demystifying Deep Reinforcement Learning” for an introduction to the topic and “Part 2: Deep Reinforcement Learning with Neon” for the original implementation in Simple-DQN. In this blog post we will extend a Simple-DQN to work with OpenAI Gym, a new…

Read more

#neon #Reinforcement Learning

Mar 04, 2016   |   Scott Gray

"Not so fast, FFT": Winograd

Deep learning thrives on speed. Faster training enables the construction of larger and more complex networks to tackle new domains such as speech or decision making. Recently, small convolutional filter sizes have become an important component in convolutional neural networks such as Google’s AlphaGo network or Microsoft’s deep residual networks. While most convolutions are computed…

Read more

#neon

Feb 09, 2016   |   Hunter Lang

Using neon for Scene Recognition: Mini-Places2

Introduction Much of the latest research in computer vision has focused on deep learning techniques. It has been applied to object recognition, where the goal is to predict what type of object is pictured in an image, and object localization, where the goal is to predict an object’s location in an image. Scene recognition is…

Read more

#Model Zoo #Scene Recognition

Feb 02, 2016   |   Arjun Bansal

neon v1.2 release: Kepler & AWS support are back, Deep ResNets, and more

We are excited to share neon’s v1.2 release with the community, which has several major features (Kepler support, new macrobatch and serialization enhancements) and examples, along with an expanded Model Zoo to help users get started with their use cases. New storage format (docs) and data loader for loading datasets that do not fit in…

Read more

#Release Notes #Scene Recognition

Stay Connected

Keep tabs on all the latest news with our monthly newsletter.