|
所在平台: Coursera |
课程主页: https://www.coursera.org/learn/building-deep-learning-models-with-tensorflow
课程评论：没有评论
课程名称：使用 TensorFlow 构建深度学习模型 课程概述： 世界上大多数数据都是未标记和非结构化的。浅层神经网络难以捕捉到例如图像、声音和文本数据中的相关结构。深度网络能够发现此类数据中的隐藏结构。在本课程中，您将使用 TensorFlow 库将深度学习应用于不同的数据类型，以解决实际问题。 学习目标： 完成本课程后，学习者将能够： - 解释 TensorFlow 的基础概念，包括主要功能、操作和执行管道。 - 描述如何在曲线拟合、回归、分类和误差函数最小化中使用 TensorFlow。 - 理解不同类型的深度架构，如卷积网络、递归网络和自编码器。 - 应用 TensorFlow 进行反向传播，以在训练神经网络时调整权重和偏差。 课程大纲： 1. 高级 Keras 功能 - 概述 Keras 的高级特性，涵盖复杂模型创建的 Keras 函数 API，创建自定义层和模型的过程，及 Keras 与 TensorFlow 2.x 的集成以增强功能，含实验和测验。 2. Keras 中的高级 CNN - 开发高级卷积神经网络（CNN）并学习数据增强技术，实施迁移学习和利用预训练模型，使用 TensorFlow 增强图像处理能力。 3. Keras 中的 Transformer - 构建和训练高级 Transformer，开发用于序列数据和时间序列的 Transformer 模型，实施高级文本生成技术。 4. Keras 中的无监督学习和生成模型 - 学习无监督学习的原则，构建和训练自编码器及扩散模型，开发生成对抗网络（GAN）并整合 TensorFlow 以应对高级无监督学习任务。 5. Keras 的高级技术 - 掌握 Keras 的高级模型开发技术，创建自定义培训循环，优化模型及进行超参数调整。 6. Keras 的强化学习简介 - 学习强化学习的基本概念及其应用，涵盖使用 Keras 的 Q-Learning 算法，开发和训练深度 Q 网络（DQN）。 7. 期末项目和作业 - 实施最终项目并进行最终评估。 通过本课程的学习，您将获得深度学习模型开发的实用技能，能够有效应用于各种数据分析和处理任务。
Name:Advanced Keras Functionalities
Description:This module provides an overview of Keras advanced features. It will cover Keras functional API for complex model creation. It also includes the creation of custom layers and models in Keras. Then the module describes the integration of Keras with TensorFlow 2.x for enhanced functionality. You will apply your learnings in labs and test your concepts in quizzes.
Name:Advanced CNNs in Keras
Description:In this module, you will learn to develop advanced convolutional neural networks (CNNs) using Keras. You will learn data augmentation techniques with Keras. In addition, you will implement transfer learning with Keras and leverage pre-trained models. Finally, you will learn how to use TensorFlow for enhancing image processing capabilities. You will apply your learnings in labs and test your concepts in quizzes.
Name:Transformers in Keras
Description:This module covers building and training advanced Transformers using Keras. You will further develop Transformer models for sequential data and time series using TensorFlow with Keras. In addition, you will learn to implement advanced Transformer techniques for text generation. You will apply your learnings in labs and test your concepts in quizzes.
Name:Unsupervised Learning and Generative Models in Keras
Description:In this module, you will learn the principles of unsupervised learning in Keras. You will learn to build and train autoencoders and diffusion models. In addition, you will develop generative adversarial networks (GANs) using Keras and integrate TensorFlow for advanced unsupervised learning tasks. You will apply your learnings in labs and test your concepts in quizzes.
Name:Advanced Keras Techniques
Description:In this module, you will learn advanced techniques in Keras for model development. You will create custom training loops and optimize models using Keras and perform hyperparameter tuning with Keras Tuner. Finally, you will learn to use TensorFlow for model optimization and custom training loops. You will apply your learnings in labs and test your concepts in quizzes.
Name:Introduction to Reinforcement Learning with Keras
Description:In this module, you will learn the fundamentals of reinforcement learning and its applications in Keras. The module also covers the Q-Learning algorithms using Keras. You will develop and train deep Q-networks (DQNs) with Keras for advanced reinforcement learning tasks. You will apply your learnings in labs and test your concepts in quizzes.
Name:Final Project and Assignment
Description:In this module, you will implement the final project and attempt the final assessment.
The majority of data in the world is unlabeled and unstructured. Shallow neural networks cannot easily capture relevant structure in, for instance, images, sound, and textual data. Deep networks are capable of discovering hidden structures within this type of data. In this course you’ll use TensorFlow library to apply deep learning to different data types in order to solve real world problems. Learning Outcomes: After completing this course, learners will be able to: • explain foundational TensorFlow concepts such as the main functions, operations and the execution pipelines. • describe how TensorFlow can be used in curve fitting, regression, classification and minimization of error functions. • understand different types of Deep Architectures, such as Convolutional Networks, Recurrent Networks and Autoencoders. • apply TensorFlow for backpropagation to tune the weights and biases while the Neural Networks are being trained.
