 |
||
|
||
Accueil > Forum TERATEC > Atelier 2
Big Data : Optimiser la prise de décision grâce aux Data Analytics
Les GPUs pour la classification d’images basée sur des techniques de réseaux de neurone
Julien DEMOUTH, DevTech, NVIDIA
In this session, we will introduce Deep Neural Networks (DNN) and present how GPU computing has revolutionized the utilization of those DNN by enabling the training of much larger networks at a much smaller cost. Deep Neural Networks are at the core of modern machine learning and can be used to classify images or do speech recognition; among other things. NVIDIA develops GPUs and software components, like cuDNN and Digits, which help improve the performance and usability of the main DNN frameworks (Theano, Torch, Caffe). In this technical presentation, we will present how those frameworks take advantage of NVIDIA's technology..
