Aurélien Geron comments

Results 284 comments of


                                            Aurélien Geron

3D image rotation Ref. Figure 5-12, p. 165 (add %matplotlib notebook)

Thanks for the tip, I'll check it out! 👍

Chapter 12 Losses and Metrics Based on Model Internals error trying to fit the model

Thanks for the useful discussion, everyone! I updated the notebook to simply comment out `super().build()` inside the `build()` method (instead of removing the `build()` method entirely), and everything works fine...

[Chapter 17] Question : Does an autoencoder based on advanced CNN such as SE-ResNET make any sense ?

Thanks for your interesting question. The optimal model always depends on the dataset (that's basically the conclusion of the "no free lunch" theorem). For example, there will be cases where...

[Chapter 17] Question : Does an autoencoder based on advanced CNN such as SE-ResNET make any sense ?

Thanks for your feedback and the interesting ideas and links. Sorry, I'm not aware of additional work on asymmetric autoencoders: it seems like an interesting thing to dig into!

Ch7: clarify comments, fix typo, update score outputs

Excellent, thanks @vi3itor! 👍

Chapter 1 Training and running a linear model using Scikit-Learn

Hi @RohitKumar23-11 , As @Praful932 said, this error means that the `oecd_bli` DataFrame does not have a column called `INEQUALITY`. Looking at your code, it looks like your code is...

[QUESTION]what is changed in this book since i have started reading second edition. is it necessary to move to third edition

Hi @kartikcpp , The second edition is fine, it's still mostly up to date, but the 3rd edition covers more topics and updates the code to the latest libraries. Here's...

[QUESTION]what is changed in this book since i have started reading second edition. is it necessary to move to third edition

Yes, the chapter on NLP includes a presentation of Transformer models, how to build them, how to use the pretrained models from TF Hub, and a quick intro to HuggingFace's...

[QUESTION]what is changed in this book since i have started reading second edition. is it necessary to move to third edition

The early release is already available on O'Reilly's Safari platform at https://homl.info/er3 It currently includes chapters 1 to 12, and more chapters will be added every few weeks. I expect...

[Chapter 18] Reinforcement Learning Breakout Frames

Hi @Praful932 , Thanks for your questions. If I remember correctly, the 108k number was just one of the hyperparameters listed in the original DeepMind paper. Let's suppose the ALE...