Tag：Deep learning
-
Time：2020-1-18
abstractMxnet is one of the main frameworks in the field of deep learning. This paper introduces mxnet from the aspects of characteristics, architecture and programming mode. Answers how to quickly deploy and run mxnet on alicloud, and introduces some performance practices of mxnet on alicloud. Introduction to the speaker: Xie Feng, Alibaba cloud heterogeneous computing […]
-
Time：2020-1-17
Different from the “handgrip” of other introductory tutorials, this article emphasizes “cause” rather than “effect”. The reason why I add “general” is that after you understand the development environment, you will not make those very low mistakes.We all know that deep learning involves a large number of models and algorithms. Looking at those messy formula […]
-
Time：2020-1-15
Natural language processing, voice processing, text processing. Speech recognition enables the computer to “understand” human speech and “extract” the text information of speech. Japan’s Wells Fargo life insurance company spent $1.7 million to install artificial intelligence systems, convert text to customer language, and analyze positive or negative words. Intelligent customer service is the research focus […]
-
Time：2020-1-14
Abstract: machine learning practice part: build your own neural network model with simple code ~ Keras is one of the most popular deep learning libraries, which has made great contribution to the commercialization of artificial intelligence. It’s very simple to use, and it allows you to build powerful neural networks in a few lines of […]
-
Time：2020-1-14
Professor Li Feifei, Artificial Intelligence Laboratory of Stanford University, realizes three elements of artificial intelligence: syntax, semantics and inference. Language, vision. By using Grammar (language grammar analysis, visual three-dimensional structure analysis) and semantics (language semantics, visual special action meaning) as model input training data, reasoning ability is realized, training learning ability is applied to work, […]
-
Time：2020-1-13
Absrtact: This paper introduces the spatial transformation network and its working principle in detail. At last, the whole mechanism is reduced to two familiar concepts: affine transformation and bilinear interpolation. Pure dry goods: space transformation network implemented by deep learning-part1 In the first part, we mainly introduce two very important concepts: affine transformation and bilinear […]
-
Time：2020-1-13
Distributed tensorflow is supported by the underlying technology of high performance grpc library. Martin Abadi, Ashish Agarwal, Paul Barham, tensorflow: large scale machine learning on heterogeneous distributed systems. Distributed principle. The distributed cluster consists of multiple server processes and client processes. Deployment mode: single machine multi card, distributed (multi machine multi card). The tensorflow is […]
-
Time：2020-1-12
Abstract: This article describes how tensorrt integration can speed up tensorflow reasoning and help you get started with case demonstration NVIDIA announced the completion of the reasoning optimization tool tensorrt and tensorflow to work together. Tensorrt integration will be available for tensorflow version 1.7. Tensorflow is still the most popular deep learning framework, and NVIDIA […]
-
Time：2020-1-12
XLA (accelerated linear algebra), the main specific compiler in the field of linear algebra, optimizes tensorflow calculation. Just in time (JIT) compilation or ahead of time (AOT) compilation to implement XLA is helpful for hardware acceleration. XLA is still in the experimental stage. https://www.tensorflow.org/ve… 。 XLA advantage. Special compiler in the field of linear algebra, […]
-
Time：2020-1-11
Tensorflow debugger (tfdbg), a special debugger for tensorflow. The real-time data flow is displayed by breakpoints and computer graphics, and the internal structure and state of tensorflow graphics are visualized. It helps to train reasoning and debug model errors. https://www.tensorflow.org/pr… 。 Common error types, non numeric (Nan), infinite (INF). Tfdbg command line interface (CLI). Debugger […]
-
Time：2020-1-10
Alphago, 1000 nodes for each experiment, 4 GPUs and 4000 GPUs for each node. Siri, 2 nodes and 8 GPUs for each experiment. AI research, relying on massive data computing, is far away from performance computing resources. A larger cluster operation model reduces the weekly training time to the day level and hour level. Kubernetes, […]
-
Time：2020-1-9
Hadoop ecological big data system is divided into yam, HDFS and MapReduce computing frameworks. Tensorflow distributed is equivalent to MapReduce computing framework and kubernetes is equivalent to yam scheduling system. Tensorflowonspark uses remote direct memory access (RDMA) to solve storage functions and scheduling, and realizes deep learning and big data fusion. Tensorflowonspark (TFOs), an open […]
