Deep Learning Cookbook: Practical Recipes to Get Started Quickly

Douwe Osinga

  • 出版商: O'Reilly
  • 出版日期: 2018-07-17
  • 定價: $2,100
  • 售價: 8.0$1,680
  • 語言: 英文
  • 頁數: 252
  • 裝訂: Paperback
  • ISBN: 149199584X
  • ISBN-13: 9781491995846
  • 相關分類: DeepLearning
  • 相關翻譯: 深度學習實戰 (簡中版)
  • 立即出貨

商品描述

Recent developments in deep learning have put the field center stage for innovation in software engineering. New algorithms and techniques in academia hold promise for many real world problems, and new machine learning platforms are powerful, but aren’t necessarily easy to get started with.

With this hands-on cookbook, you'll discover that deep learning doesn't need to be intimidating. Aimed at readers who are new to deep learning, this cookbook enables you to solve problems quickly, using the most appropriate platform for each application. You'll learn how to leverage the work of Google by reusing pre-trained networks, use non-final layers to map data, and build recommender systems out of any correlation data.

  • Work with step-by-step recipes that address familiar problems in areas such as text embeddings, text labeling and generation, and image classification and generation
  • Walk through a practical solution for each recipe, using modern machine learning frameworks
  • Learn how your newly-trained models can be easily ported for use in production settings
  • Build applications that go from interesting results to serving real users
  • Use deep learning in production, including how to query embeddings with the Postgres database, and how export and serve models using TensorFlow
  • Set up a microservice using Python, and run models on mobile devices

商品描述(中文翻譯)

近年來,深度學習在軟體工程創新方面成為熱門話題。學術界的新演算法和技術為許多現實世界的問題帶來了希望,而新的機器學習平台功能強大,但並不一定容易上手。

這本實用手冊專為初學者設計,讓你發現深度學習並不可怕。通過這本手冊,你可以快速解決問題,並選擇最適合每個應用的平台。你將學習如何通過重用預訓練網絡、使用非最終層來映射數據,以及利用任何相關數據構建推薦系統。

本書特點如下:
- 提供一系列逐步解決常見問題的實用配方,包括文本嵌入、文本標記和生成,以及圖像分類和生成。
- 使用現代機器學習框架,逐步實現每個配方的解決方案。
- 學習如何將新訓練的模型輕鬆移植到生產環境中使用。
- 構建從有趣結果到為真實用戶提供服務的應用程式。
- 在生產環境中使用深度學習,包括如何使用Postgres數據庫查詢嵌入向量,以及如何使用TensorFlow導出和提供模型。
- 使用Python設置微服務,並在移動設備上運行模型。