生成式深度学习 第2版(影印版)
David Foster
出版时间:2024年03月
页数:453
“本书通俗易懂地介绍了用于生成式建模的深度学习工具包。如果你是一个喜欢捣鼓代码的创意工作者,希望在工作中应用深度学习,那么这本书就是你的最佳选择。”
——David Ha
Stability A战略主管
“这本杰作深入研究了最先进的生成式深度学习背后的所有主要技术。这是对人工智能最迷人领域之一的一次激动人
心的探索!”
——Francois Chollet
Keras创作

生成式AI是科技领域最炙手可热的话题。这本实践用书教机器学习工程师和数据科学人员如何利用TensorFlow和Keras从零开始创建令人印象深刻的生成式深度学习模型,包括变分自编码器(VAE)、生成对抗网络(GAN)、Transformers、归一化流、基于能量的模型和去噪扩散模型。
本书从深度学习的基础知识开始,逐步讲解最前沿的架构。通过各种技巧和窍门,你将理解如何使你的模型更高效地学习,变得更具创造力。
● 可了解VAE如何改变照片中的面部表情
● 训练GAN基于你自己的数据集生成图像
● 构建扩散模型,产生新品种的花卉
● 训练自己的GPT进行文本生成
● 学习ChatGPT等大语言模型的训练方式
● 探索StyleGAN2和ViTVQGAN等最先进的架构
● 使用Transformers和MuseGAN创作多声部音乐
● 理解生成世界模型如何解决强化学习任务
● 深入研究DALL.E 2、Imagen、Stable Diffusion等多模态模型
书中还探讨了生成式AI的未来,以及个人和公司如何积极利用这一引人注目的新技术创造竞争优势。
  1. Foreword
  2. Preface
  3. Part I. Introduction to Generative Deep Learning
  4. 1. Generative Modeling
  5. What Is Generative Modeling?
  6. Our First Generative Model
  7. Core Probability Theory
  8. Generative Model Taxonomy
  9. The Generative Deep Learning Codebase
  10. Summary
  11. 2. Deep Learning
  12. Data for Deep Learning
  13. Deep Neural Networks
  14. Multilayer Perceptron (MLP)
  15. Convolutional Neural Network (CNN)
  16. Summary
  17. Part II. Methods
  18. 3. Variational Autoencoders
  19. Introduction
  20. Autoencoders
  21. Variational Autoencoders
  22. Exploring the Latent Space
  23. Summary
  24. 4. Generative Adversarial Networks
  25. Introduction
  26. Deep Convolutional GAN (DCGAN)
  27. Wasserstein GAN with Gradient Penalty (WGAN-GP)
  28. Conditional GAN (CGAN)
  29. Summary
  30. 5. Autoregressive Models
  31. Introduction
  32. Long Short-Term Memory Network (LSTM)
  33. Recurrent Neural Network (RNN) Extensions
  34. PixelCNN
  35. Summary
  36. 6. Normalizing Flow Models
  37. Introduction
  38. Normalizing Flows
  39. RealNVP
  40. Other Normalizing Flow Models
  41. Summary
  42. 7. Energy-Based Models
  43. Introduction
  44. Energy-Based Models
  45. Summary
  46. 8. Diffusion Models
  47. Introduction
  48. Denoising Diffusion Models (DDM)
  49. Summary
  50. Part III. Applications
  51. 9. Transformers
  52. Introduction
  53. GPT
  54. Other Transformers
  55. Summary
  56. 10. Advanced GANs
  57. Introduction
  58. ProGAN
  59. StyleGAN
  60. StyleGAN2
  61. Other Important GANs
  62. Summary
  63. 11. Music Generation
  64. Introduction
  65. Transformers for Music Generation
  66. MuseGAN
  67. Summary
  68. 12. World Models
  69. Introduction
  70. Reinforcement Learning
  71. World Model Overview
  72. Collecting Random Rollout Data
  73. Training the VAE
  74. Collecting Data to Train the MDN-RNN
  75. Training the MDN-RNN
  76. Training the Controller
  77. In-Dream Training
  78. Summary
  79. 13. Multimodal Models
  80. Introduction
  81. DALL.E 2
  82. Imagen
  83. Stable Diffusion
  84. Flamingo
  85. Summary
  86. 14. Conclusion
  87. Timeline of Generative AI
  88. The Current State of Generative AI
  89. The Future of Generative AI
  90. Final Thoughts
  91. Index
书名:生成式深度学习 第2版(影印版)
作者:David Foster
国内出版社:东南大学出版社
出版时间:2024年03月
页数:453
书号:978-7-5766-1200-4
原版书书名:Generative Deep Learning, 2nd Edition
原版书出版商:O'Reilly Media
David Foster
 
David Foster是Applied Data Science的联合创始人,这是一家数据科学咨询公司,为客户提供创新的解决方案。他拥有英国剑桥三一学院的数学硕士学位,以及华威大学运筹学硕士学位。
David曾多次赢得国际机器学习竞赛,包括 InnoCentive Predicting Product Purchase 大奖赛,并获得了可视化的第一名,这项技术可以帮助美国的制药公司优化临床试验的选址。
David活跃在在线数据科学社区,并撰写了一系列有关深度强化学习的博客文章,包括《How To Build Your Own AlphaZero AI Using Python and Keras》(地址:http://bit. ly/2J6fGhU)。
 
 
The Pyrrhura genus falls under the family Psittacidae, one of three families of parrots. Within its subfamily Arinae are several macaw and parakeet species of the Western Hemisphere. The painted parakeet inhabits the coastal forests and mountains of northeastern South America.

Bright green feathers cover most of a painted parakeet, but they are blue above the beak, brown in the face, and reddish in the breast and tail. Most strikingly, the feathers on the painted parakeet’s neck look like scales; the brown center is outlined in offwhite. This combination of colors camouflages the birds in the rainforest.

Painted parakeets tend to feed in the forest canopy, where their green plumage masks them best. They forage in flocks of 5 to 12 birds for a wide variety of fruits, seeds, and flowers. Occasionally, when feeding below the canopy, painted parakeets will eat algae from forest pools. They grow to about 9 inches in length and live for 13 to 15 years. A clutch of painted parakeet chicks—each of which are less than an inch wide at hatching—is usually around five eggs.
购买选项
定价:132.00元
书号:978-7-5766-1200-4
出版社:东南大学出版社