2024-02-08 19:00:18

totrans/gen-dl_05.yaml

@@ -1,52 +1,70 @@
 - en: Part II. Methods
+  id: totrans-0
   prefs:
   - PREF_H1
   type: TYPE_NORMAL
+  zh: 第二部分. 方法
 - en: In [Part II](#part_methods) we will dive into the six families of generative
     models, including the theory behind how they work and practical examples of how
     to build each type of model.
+  id: totrans-1
   prefs: []
   type: TYPE_NORMAL
+  zh: 在[第二部分](#part_methods)中，我们将深入探讨六种生成模型系列，包括它们工作原理的理论以及如何构建每种类型模型的实际示例。
 - en: In [Chapter 3](ch03.xhtml#chapter_vae) we shall take a look at our first generative
     deep learning model, the *variational autoencoder*. This technique will allow
     us to not only generate realistic faces, but also alter existing images—for example,
     by adding a smile or changing the color of someone’s hair.
+  id: totrans-2
   prefs: []
   type: TYPE_NORMAL
+  zh: 在[第3章](ch03.xhtml#chapter_vae)中，我们将看一下我们的第一个生成式深度学习模型，*变分自动编码器*。这种技术不仅可以让我们生成逼真的人脸，还可以改变现有的图像，例如添加微笑或改变某人头发的颜色。
 - en: '[Chapter 4](ch04.xhtml#chapter_gan) explores one of the most successful generative
     modeling techniques of recent years, the *generative adversarial network*. We
     shall see the ways that GAN training has been fine-tuned and adapted to continually
     push the boundaries of what generative modeling is able to achieve.'
+  id: totrans-3
   prefs: []
   type: TYPE_NORMAL
+  zh: '[第4章](ch04.xhtml#chapter_gan)探讨了近年来最成功的生成建模技术之一，*生成对抗网络*。我们将看到GAN训练如何被微调和调整，以不断推动生成建模能够实现的边界。'
 - en: In [Chapter 5](ch05.xhtml#chapter_autoregressive) we will delve into several
     examples of *autoregressive models*, including LSTMs and PixelCNN. This family
     of models treats the generation process as a sequence prediction problem—it underpins
     today’s state-of-the-art text generation models and can also be used for image
     generation.
+  id: totrans-4
   prefs: []
   type: TYPE_NORMAL
+  zh: 在[第5章](ch05.xhtml#chapter_autoregressive)中，我们将深入探讨几个*自回归模型*的例子，包括LSTMs和PixelCNN。这个模型系列将生成过程视为一个序列预测问题，它支撑着今天最先进的文本生成模型，并且也可以用于图像生成。
 - en: In [Chapter 6](ch06.xhtml#chapter_flow) we will cover the family of *normalizing
     flow models*, including RealNVP. This model is based on a change of variables
     formula, which allows the transformation of a simple distribution, such as a Gaussian
     distribution, into a more complex distribution in way that preserves tractability.
+  id: totrans-5
   prefs: []
   type: TYPE_NORMAL
+  zh: 在[第6章](ch06.xhtml#chapter_flow)中，我们将涵盖*归一化流模型*系列，包括RealNVP。这个模型基于变量变换公式，允许将简单分布（如高斯分布）转换为更复杂的分布，同时保持可处理性。
 - en: '[Chapter 7](ch07.xhtml#chapter_energy_based_models) introduces the family of
     *energy-based models*. These models train a scalar energy function to score the
     validity of a given input. We will explore a technique for training energy-based
     models called contrastive divergence and a technique for sampling new observations
     called Langevin dynamics.'
+  id: totrans-6
   prefs: []
   type: TYPE_NORMAL
+  zh: '[第7章](ch07.xhtml#chapter_energy_based_models)介绍了*基于能量的模型*系列。这些模型训练一个标量能量函数来评分给定输入的有效性。我们将探讨一种用于训练基于能量模型的技术，称为对比散度，以及一种用于采样新观测的技术，称为Langevin动力学。'
 - en: Finally, in [Chapter 8](ch08.xhtml#chapter_diffusion) we shall explore the family
     of *diffusion models*. This technique is based on the idea of iteratively adding
     noise to an image and then training a model to remove the noise, giving us the
     ability to transform pure noise into realistic samples.
+  id: totrans-7
   prefs: []
   type: TYPE_NORMAL
+  zh: 最后，在[第8章](ch08.xhtml#chapter_diffusion)中，我们将探索*扩散模型*系列。这种技术基于迭代地向图像添加噪声，然后训练模型去除噪声，使我们能够将纯噪声转换为逼真的样本。
 - en: By the end of [Part II](#part_methods) you will have built practical examples
     of generative models from each of the six generative modeling families and be
     able to explain how each works from a theoretical perspective.
+  id: totrans-8
   prefs: []
   type: TYPE_NORMAL
+  zh: 到[第二部分](#part_methods)结束时，您将从每个生成建模系列中构建实际示例，并能够从理论角度解释每种模型的工作原理。