书籍 Deep Learning with Python, Second Edition的封面

Deep Learning with Python, Second Edition

François Chollet

出版时间

2020-06-03

ISBN

9781617296864

评分

★★★★★
书籍介绍

Deep Learning with Python introduces the field of deep learning using the Python language and the powerful Keras library. You’ll learn directly from the creator of Keras, François Chollet, building your understanding through intuitive explanations and practical examples. Updated from the original bestseller with over 50% new content, this second edition includes new chapters, cutting-edge innovations, and coverage of the very latest deep learning tools. You'll explore challenging concepts and practice with applications in computer vision, natural-language processing, and generative models. By the time you finish, you'll have the knowledge and hands-on skills to apply deep learning in your own projects.

what's inside

Deep learning from first principles

Image-classification, imagine segmentation, and object detection

Deep learning for natural language processing

Timeseries forecasting

Neural style transfer, text generation, and image generation

François Chollet works on deep learning at Google in Mountain View, CA. He is the creator of the Keras deep-learning library, as well as a contributor to the TensorFlow machine-learning framework. He also does AI research, with a focus on abstraction and reasoning. His papers have been published at major conferences in the field, including the Conference on Computer Vision and ...

(展开全部)

用户评论
2nd edition claims to be fairly different from 1st edition, with 1/3 longer and 75% novel content. However, after reading the whole book (7-8h), the main contents of the 2nd edition is quite similar to the 1st edition, only with more noticeable new contents in ch01-02, ch09, and ch14. Overall, it is a fantastic book to deepen understanding of DL.
最好的深层学习入门书,没有之一。相较第一版至少有一半以上内容增加,囊括了最新的transformer。阅读的时候就有一种:新书到手,天下我有!的感觉。不愧是keras的作者,对deep learning的讲解深入浅出,尤其是很多概念的由来,模型的训练tricks。除了最后一章对于未来machine learning的构想只适合了解以外,私以为每一章的每个细节都适合多次阅读,并且在实践中反复测试。读完以后,一个感觉就是有些相对厉害的中牛(姑且称之为不愿意分享核心技术的炼丹师们),不会告诉你他们读过这本书,因为这本书不仅仅是理论+实践那么简单,里面还有很多心法。很多机器学习和深层学习的书会给你讲原理,但很少告诉你如何提高训练精度。这本书会让你少走很多弯路!
一年前看过第一版,但忘的差不多了,这次就第一版和第二版一起看的,差别还是挺大的,前面那个说“quite similar to the 1st edition”的人压根就没仔细看。我所有代码都跑了,同时和第一版的做了对比,差别在:很多数据预处理用了最新的集成在keras里的函数,而第一版基本是用base python做的,其他的部分如果keras有更新,书里也跟进了;介绍了第一版以来到2020年书出版的时候DL的新进展,比如介绍了attention;对于用RNN来做时序分析单独写了一章;第7章关于深入实用keras的介绍比第一版深入多了,不过这部分内容更推荐那本《机器学习实战》 hands-on那本,虽然本书是keras之父写的,那不如人家介绍的好。最后,能看第二版的尽量第二版
略读了一遍
是最好的deep learning教材了:理论部分intuitive的恰到好处,不浅薄也不过于硬核,尤其是CNN,RNN,Transformer这三章给我留下了非常深刻的印象,穿插的例子都相当make sense;代码部分详略得当,text 这一章比较复杂,顺便刷新了对class的认识,Keras真是好用,谁用谁知道! 最后一章深入讨论了AI,我很认同artificial intelligence 应该称为 artificial cognition这个论点,眼下的AI 和底特律里的仿生人完全不是一回事,它们只是在gradient decent 罢了【不是】
非常优秀的入门书! 这本书对时间序列、NLP、计算机视觉、图像生成都做了充分而细致的介绍,紧跟前沿。同时揭示了深度网络的本质——找到高维数据空间的低维流体,并解释了基于几何变换和差分的深度网络的局限性。 本书还对人工智能的未来进行了展望,例如添加程序处理原语,让神经网络不仅能捕捉特征,还能操控结构。 同时,作为 Keras 开发者,作者对架构、抽象都有深刻理解,在介绍概念和细节的同时,会解释它在系统中的作用、设计思路和数学原理。从使用者、研究者和框架开发者三个角度进行阐述,完整的回答了“怎么用”、“为什么可以用”和“这么用的意义”的问题。
写的通俗易懂,入门最佳读物。