CNTK: Microsoft's Open-Source Deep-Learning Toolkit
13 108
38.7
Microsoft Research321 тыс
Следующее
03.06.16 – 1 4155:29Популярные
233 дня – 1 7001:18Опубликовано 27 мая 2016, 18:48
Speaker: Frank Seide
This talk will introduce CNTK, Microsoft’s cutting-edge open-source deep-learning toolkit for Windows and Linux. CNTK is a computation-graph based deep-learning toolkit for training and evaluating deep neural networks. Microsoft product groups use CNTK, for example to create the Cortana speech models and web ranking. CNTK supports feed-forward, convolutional, and recurrent networks for speech, image, and text workloads, also in combination. Popular network types are supported either natively (convolution) or can be described as a CNTK configuration (LSTM, sequence-to-sequence). CNTK scales to multiple GPU servers and is designed around efficiency. We will give an overview of CNTK's general architecture and describe the specific methods and algorithms used for automatic differentiation, recurrent-loop inference and execution, memory sharing, on-the-fly randomization of large corpora, and multi-server parallelization. We will then discuss how typical uses looks like for relevant tasks like image recognition, sequence-to-sequence modeling, and speech recognition.
research.microsoft.com/latamfa...
This talk will introduce CNTK, Microsoft’s cutting-edge open-source deep-learning toolkit for Windows and Linux. CNTK is a computation-graph based deep-learning toolkit for training and evaluating deep neural networks. Microsoft product groups use CNTK, for example to create the Cortana speech models and web ranking. CNTK supports feed-forward, convolutional, and recurrent networks for speech, image, and text workloads, also in combination. Popular network types are supported either natively (convolution) or can be described as a CNTK configuration (LSTM, sequence-to-sequence). CNTK scales to multiple GPU servers and is designed around efficiency. We will give an overview of CNTK's general architecture and describe the specific methods and algorithms used for automatic differentiation, recurrent-loop inference and execution, memory sharing, on-the-fly randomization of large corpora, and multi-server parallelization. We will then discuss how typical uses looks like for relevant tasks like image recognition, sequence-to-sequence modeling, and speech recognition.
research.microsoft.com/latamfa...
Свежие видео
2 дня – 36 70210:54
6 дней – 9 2883:26
7 дней – 1003:32
7 дней – 550:36
7 дней – 48 93525:26
16 дней – 12 1261:35Случайные видео
171 день – 6 72416:57
195 дней – 5 097 2220:43
306 дней – 584 4317:59
317 дней – 11 4080:34
18.01.22 – 1 3800:15
10.11.21 – 1 950 9061:29
4 часа – 1 1658:01
14 часов – 49238:33
14 часов – 14638:24
2 дня – 8 3591:03
3 дня – 2470:31
6 дней – 2 5170:42
10 часов – 3 89517:08
14 часов – 41943:11