0%

Book Description

Deep learning is enabling the next generation of successful companies. The question is no longer whether enterprises will use deep learning (they will), but how involved each organization becomes with the technology.

Sean Murphy and Allen Leis introduce deep learning from an enterprise perspective and offer an overview of the TensorFlow library and ecosystem. If your company is adopting deep learning, this report will help you navigate the initial decisions you must make—from choosing a deep learning framework to integrating deep learning with the other data analysis systems already in place—to ensure you're building a system capable of handling your specific business needs.

  • Explore fundamental concepts and core questions about deep learning in the enterprise
  • Familiarize yourself with available framework options, including TensorFlow, MXNet, Microsoft Cognitive Toolkit, and Deeplearning4J
  • Dive into TensorFlow's library and ecosystem, from tools such as estimators, prebuilt neural networks, Keras, ML Toolkit for TensorFlow, Tensor2Tensor (T2T), TensorBoard, and TensorFlow Debugger, to model deployment and management with TensorFlow Serving
  • See how companies such as Jet.com and PingThings have implemented deep learning to improve the accuracy and enhance the performance of a number of tasks

Table of Contents

  1. Introduction
  2. 1. Choosing to Use Deep Learning
    1. General Rationale
    2. Specific Incentives
      1. Using Sequence Data
      2. Using Images and Video
    3. Specific Enterprise Examples
    4. Potential Downsides
    5. Summary
  3. 2. Selecting a Deep Learning Framework
    1. Enterprise-Ready Deep Learning
      1. TensorFlow
      2. MXNet
      3. Microsoft Cognitive Toolkit (CNTK)
      4. Deeplearning4J
    2. Industry Perspectives
    3. Summary
  4. 3. Exploring the Library and the Ecosystem
    1. Improving Network Design and Training
      1. Estimators
      2. Prebuilt Neural Networks
      3. Keras
      4. Machine Learning Toolkit for TensorFlow
      5. Tensor2Tensor (T2T)
      6. TensorBoard
      7. TensorFlow Debugger
    2. Deploying Networks for Inference
      1. TensorFlow Serving
      2. In-Process Serving
    3. Integrating with Other Systems
      1. Data Ingestion Options
      2. TensorFlowOnSpark
      3. “Ecosystem” Repo
    4. Accelerating Training and Inference
      1. GPUs and CUDA
      2. Tensor Processing Units
      3. Google Cloud TPU and CloudML
    5. Summary
  5. Conclusion