The following sentence from "INSATALL.md" is somewhat outdated:

but the good news is that they will be installed automatically when you install the tensorflow-gpu package from Anaconda.

It is because the official command below uses tf version 2.4 which is not compatible with the jupyter notebooks for homl3.

conda install -c anaconda tensorflow-gpu

Cf. https://anaconda.org/anaconda/tensorflow-gpu

Is there any other easy way to use tensorflow-gpu?

Cloned repo and created new env from the environment.yml file today. When I ran through notebook, "10_neural_nets_with_keras", the, "tf.keras.utils.plot_model(...)" cell failed with missing pydot requirement. Cell is marked as "extra code", and overall this doesn't really impact running through the notebook.

Thanks for helping us improve this project!

Before you create this issue Please make sure you are using the latest updated code and libraries: see https://github.com/ageron/handson-ml3/blob/main/INSTALL.md#update-this-project-and-its-libraries

Also please make sure to read the FAQ (https://github.com/ageron/handson-ml3#faq) and search for existing issues (both open and closed), as your question may already have been answered: https://github.com/ageron/handson-ml3/issues

Describe the bug Part 1 - Chapter 2 - (EndtoEndML), in creating ColumnTransformer supplying feature_names_out as argument rises TypeError exception.

To Reproduce Please copy the code that fails here, using code blocks like this:

def ratio_pipeline(name=None):
    return make_pipeline(
        SimpleImputer(strategy="median"),
        FunctionTransformer(column_ratio, feature_names_out= lambda input_features: [name]),
        StandardScaler())

And if you got an exception, please copy the full stacktrace here:

Traceback (most recent call last):
  File "F:\*****\*****\*****\CHAPTER2-END_TO_END_PROJECT\final-example.py", line 49, in <module>
    ("bedrooms_ratio", ratio_pipeline("bedrooms_ratio"), ["total_bedrooms", "total_rooms"]),
  File "F:\*****\*****\*****\CHAPTER2-END_TO_END_PROJECT\final-example.py", line 38, in ratio_pipeline
    FunctionTransformer(column_ratio, feature_names_out= lambda input_features: [name]),
TypeError: FunctionTransformer.__init__() got an unexpected keyword argument 'feature_names_out'

Expected behavior The preprocessing fit and transform the dataset

Versions (please complete the following information):

• OS: Windows 10
• Python: 3.10
• TensorFlow: ****
• Scikit-Learn: 1.0.2
• Other libraries that may be connected with the issue:*****

Additional context

This is my full code

import numpy as np
from sklearn.base import BaseEstimator, TransformerMixin
from sklearn.cluster import KMeans
from sklearn.compose import ColumnTransformer, make_column_selector
from sklearn.impute import SimpleImputer
from sklearn.metrics.pairwise import rbf_kernel
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import FunctionTransformer, StandardScaler, OneHotEncoder
from data import housing


class ClusterSimilarity(BaseEstimator, TransformerMixin):
    def __init__(self, n_clusters=10, gamma=1.0, random_state=None):
        self.n_clusters = n_clusters
        self.gamma = gamma
        self.random_state = random_state

    def fit(self, X, y=None, sample_weight=None):
        self.kmeans_ = KMeans(self.n_clusters, random_state=self.random_state)
        self.kmeans_.fit(X, sample_weight=sample_weight)
        return self  # always return self!

    def transform(self, X):
        return rbf_kernel(X, self.kmeans_.cluster_centers_, gamma=self.gamma)

    def get_feature_names_out(self, names=None):
        return [f"Cluster {i} similarity" for i in range(self.n_clusters)]


def column_ratio(X):
    return X[:, [0]] / X[:, [1]]


def ratio_pipeline(name=None):
    return make_pipeline(
        SimpleImputer(strategy="median"),
        FunctionTransformer(column_ratio, feature_names_out= lambda input_features: [name]),
        StandardScaler())


log_pipeline = make_pipeline(SimpleImputer(strategy="median"), FunctionTransformer(np.log), StandardScaler())
cluster_simil = ClusterSimilarity(n_clusters=10, gamma=1., random_state=42)
default_num_pipeline = make_pipeline(SimpleImputer(strategy="median"),StandardScaler())

cat_pipeline = make_pipeline(SimpleImputer(strategy="most_frequent"), OneHotEncoder(handle_unknown="ignore"))

preprocessing = ColumnTransformer([
    ("bedrooms_ratio", ratio_pipeline("bedrooms_ratio"), ["total_bedrooms", "total_rooms"]),
    ("rooms_per_house", ratio_pipeline("rooms_per_house"), ["total_rooms", "households"]),
    ("people_per_house", ratio_pipeline("people_per_house"), ["population", "households"]),
    ("log", log_pipeline, ["total_bedrooms", "total_rooms", "population", "households", "median_income"]),
    ("geo", cluster_simil, ["latitude", "longitude"]),
    ("cat", cat_pipeline, make_column_selector(dtype_include=np.object)),
], remainder=default_num_pipeline)  # one column remaining: housing_median_age


if __name__ == '__main__':
    housing_prepared = preprocessing.fit_transform(housing)
    print(housing_prepared.shape)
    print(preprocessing.get_feature_names_out())

What's the right definition of the cost function for Ridge regression?

1. The doc version:

2. The book version:

In the 2nd edition, 2 is used instead of m. Anyway, it's not obvious why with m.

Hi @ageron,

This PR clarifies some comments, fixes several typos and deprecated np.object, and updates output scores in the solutions for exercises 8 and 9, because it makes some difference for the comments.

I also updated the answers to Solutions 6 and 7 about decreasing/increasing the learning rate. Please double-check: it makes sense to me, but I may be wrong here.

Hello, I'm using Nvidia Geforce 1050 ti 4gb, after running cell [49], I'm getting this error. Kindly share the requirements of GPU which will run smoothly with this book.

W tensorflow/core/framework/op_kernel.cc:1733] RESOURCE_EXHAUSTED: failed to allocate memory

Hi @ageron,

This PR updates some figure labels, corrects several typos (mostly in the solution section) and updates the link to NumPy documentation.

Also, you finished Chapter 4 with the following sentences: We’ve opened up the first Machine Learning black boxes! In the next chapters, we will open many more, starting with Decision Trees. Did you mean "Support Vector Machines"?

Hi @ageron,

Yet another PR :)

• updated documentation links,
• removed unnecessary import for 3D plotting: it's not been required since Matplotlib 3.2.0 (March 2020),
• added the link to matplotlib's tutorial for 3d plotting,
• corrected a few typos,
• replaced "attributes" with "parameters" and "arguments" depending on the context.

According to sklearn's document, it seems that the hyperparmeter decision_function_shape for SVC model is "ovr", not "ovo". But in the book, "ovo" is mentioned and explained in the section about Multiclass classification, Chapter 3.

On the other hand, it is written in the jupyter notebook as follows:

If you want decision_function() to return all 45 scores, you can set the decision_function_shape hyperparameter to "ovo". The default value is "ovr", but don't let this confuse you: SVC always uses OvO for training. This hyperparameter only affects whether or not the 45 scores get aggregated or not:

This should be also explained in the book.

The definition of ratio_pipeline() function in the jupyter notebook of Chapter 2 contains the following transformer:

FunctionTransformer(column_ratio,
                    feature_names_out=[name]),

But the keyword feature_names_out is not listed as an argument in the scikit-learn document of FunctionTransformer class or its parent classes, and it's not clear how to understand it.

Is that just for naming the transformer to be built or for any other pourpose? And why isn't it listed as a keyword argment of the FunctionTransformer's constructor or of its parent classes'?

load_housing_data() function had an indentation error because of which if the 'housing.tgz' already existed in the drive, the code won't extract the 'housing.csv' and 'datasets/housing/housing.csv' won't exist.

Thanks for helping us improve this project!

Before you create this issue Please make sure you are using the latest updated code and libraries: see https://github.com/ageron/handson-ml3/blob/main/INSTALL.md#update-this-project-and-its-libraries

Also please make sure to read the FAQ (https://github.com/ageron/handson-ml3#faq) and search for existing issues (both open and closed), as your question may already have been answered: https://github.com/ageron/handson-ml3/issues

Describe the bug Running cell 72 in the Chapter 2 returns a 2x5 sparse matrix and not a numpy array. I run this in Jetbrains DataSpell.

To Reproduce Please copy the code that fails here, using code blocks like this: cell72:

cat_encoder.transform(df_test)

And if you got an exception, please copy the full stacktrace here: not an exception and the output is correct - it just misses the step of converting the output to a numpy array. The np array output is displayed both in the book (page 74 at the top as well as in the notebook).

<2x5 sparse matrix of type '<class 'numpy.float64'>'
	with 2 stored elements in Compressed Sparse Row format>

Expected behavior The expected out put is the output in the book (p. 74) or in the notebook (out:72) I achieve this result by adding .toarray() to the code

Screenshots If applicable, add screenshots to help explain your problem.

Versions (please complete the following information):

• OS: MacOS Ventura 13.1
• Python: 3.10
• TensorFlow: as per requirements.txt
• Scikit-Learn: as per requirements.txt
• Other libraries that may be connected with the issue: Using Jetbrains DataSpell

Additional context Add any other context about the problem here.

Having trouble creating HOML3 environment on Windows 10. I thought this might be due to having standalone Python 3.10 installed, so I uninstalled that, then uninstalled and reinstalled anaconda. Each time I retry creating the HOML3 environment, I first clean up by deleting env3\homl3 in the anaconda folder.

Here's the error. Before trying to create the HOML3 environment this time, I did "python -m pip install libtorrent" and got "Requirement already satisfied: libtorrent in c:\users\x475a\anaconda3\lib\site-packages (2.0.7)"

Thank you.

Building wheel for AutoROM.accept-rom-license (pyproject.toml): finished with status 'error' Failed to build AutoROM.accept-rom-license

Pip subprocess error: error: subprocess-exited-with-error

× Building wheel for AutoROM.accept-rom-license (pyproject.toml) did not run successfully. │ exit code: 1 ╰─> [66 lines of output] C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools\config\setupcfg.py:508: SetuptoolsDeprecationWarning: The license_file parameter is deprecated, use license_files instead. warnings.warn(msg, warning_class) running bdist_wheel running build running build_py creating build creating build\lib copying AutoROM.py -> build\lib installing to build\bdist.win-amd64\wheel running install running install_lib creating build\bdist.win-amd64 creating build\bdist.win-amd64\wheel copying build\lib\AutoROM.py -> build\bdist.win-amd64\wheel. running install_egg_info running egg_info writing AutoROM.accept_rom_license.egg-info\PKG-INFO writing dependency_links to AutoROM.accept_rom_license.egg-info\dependency_links.txt writing requirements to AutoROM.accept_rom_license.egg-info\requires.txt writing top-level names to AutoROM.accept_rom_license.egg-info\top_level.txt reading manifest file 'AutoROM.accept_rom_license.egg-info\SOURCES.txt' reading manifest template 'MANIFEST.in' adding license file 'LICENSE.txt' writing manifest file 'AutoROM.accept_rom_license.egg-info\SOURCES.txt' Copying AutoROM.accept_rom_license.egg-info to build\bdist.win-amd64\wheel.\AutoROM.accept_rom_license-0.5.0-py3.10.egg-info running install_scripts Traceback (most recent call last): File "C:\Users\x475a\anaconda3\envs\homl3\lib\site-packages\pip_vendor\pep517\in_process_in_process.py", line 351, in main() File "C:\Users\x475a\anaconda3\envs\homl3\lib\site-packages\pip_vendor\pep517\in_process_in_process.py", line 333, in main json_out['return_val'] = hook(**hook_input['kwargs']) File "C:\Users\x475a\anaconda3\envs\homl3\lib\site-packages\pip_vendor\pep517\in_process_in_process.py", line 249, in build_wheel return _build_backend().build_wheel(wheel_directory, config_settings, File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools\build_meta.py", line 413, in build_wheel return self._build_with_temp_dir(['bdist_wheel'], '.whl', File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools\build_meta.py", line 398, in _build_with_temp_dir self.run_setup() File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools\build_meta.py", line 484, in run_setup super(BuildMetaLegacyBackend, File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools\build_meta.py", line 335, in run_setup exec(code, locals()) File "", line 18, in File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools_init.py", line 87, in setup return distutils.core.setup(**attrs) File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools_distutils\core.py", line 185, in setup return run_commands(dist) File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools_distutils\core.py", line 201, in run_commands dist.run_commands() File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools_distutils\dist.py", line 969, in run_commands self.run_command(cmd) File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools\dist.py", line 1208, in run_command super().run_command(command) File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools_distutils\dist.py", line 988, in run_command cmd_obj.run() File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\wheel\bdist_wheel.py", line 360, in run self.run_command("install") File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools_distutils\cmd.py", line 318, in run_command self.distribution.run_command(command) File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools\dist.py", line 1208, in run_command super().run_command(command) File "C:\Users\x475a\AppData\Local\Temp\pip-build-env-vk00i65p\overlay\Lib\site-packages\setuptools_distutils\dist.py", line 988, in run_command cmd_obj.run() File "", line 11, in run File "C:\Users\x475a\AppData\Local\Temp\pip-install-n3yw4sxy\autorom-accept-rom-license_e037a052a27841a5a54970ef670daf18\AutoROM.py", line 13, in import libtorrent as lt ImportError: DLL load failed while importing libtorrent: The specified module could not be found. [end of output]

note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for AutoROM.accept-rom-license ERROR: Could not build wheels for AutoROM.accept-rom-license, which is required to install pyproject.toml-based projects

failed

CondaEnvException: Pip failed

The field of graph representation learning has grown at an incredible and sometimes unwieldy pace over the past few years, and a lot of new algorithms and innovations were made in the field. I read the second version completely it's a power-packed guide for beginners to gain knowledge on both machine learning and deep learning. But I find this important puzzle was missing there, I request the author to add a chapter on graph neural networks in upcoming versions. The hands-on experience with graph neural networks will be useful for all the readers. Thanks.

I'm currently reading the book from O'reilly website and in chapter 11 there is one link that is not working. It is under Unsupervised Pretraining header and when i'm trying to open the link https://homl.info/extra-anns which redirects to https://colab.research.google.com/github/ageron/handson-ml3/blob/main/extra_ann_architectures.ipynb but this is not working and when i cheked the repo, I dont find it as well.

@ageron can you please update it.

After training mulvar_model with train_mulvar_ds and valid_mulvar_ds, I am trying to predict the model on mulvar_test using model.predict(mulvar_test). and it's returning the array of floats containing 914 values(914 values maybe because of the shape of mulvar_test is (914, 5 ) ) and i guess thats the array of predictions on rail. all i want to know if there's any way that model.predict will return only one value instead of array 914 values?

=> ERROR [ 4/12] RUN echo ' - pyvirtualdisplay' >> /tmp/environment. 14.1s

[ 4/12] RUN echo ' - pyvirtualdisplay' >> /tmp/environment.yml && conda env create -f /tmp/environment.yml && conda clean -afy && find /opt/conda/ -follow -type f -name '.a' -delete && find /opt/conda/ -follow -type f -name '.pyc' -delete && find /opt/conda/ -follow -type f -name '*.js.map' -delete && rm /tmp/environment.yml: #8 0.444 Collecting package metadata (repodata.json): ...working... done #8 12.95 Solving environment: ...working... failed #8 12.96 #8 12.96 ResolvePackageNotFound: #8 12.96 - pyglet=1.5 #8 12.96 - box2d-py #8 12.96

executor failed running [/bin/sh -c echo ' - pyvirtualdisplay' >> /tmp/environment.yml && conda env create -f /tmp/environment.yml && conda clean -afy && find /opt/conda/ -follow -type f -name '.a' -delete && find /opt/conda/ -follow -type f -name '.pyc' -delete && find /opt/conda/ -follow -type f -name '*.js.map' -delete && rm /tmp/environment.yml]: exit code: 1 ERROR: Service 'handson-ml3' failed to build : Build failed