Keras and TensorFlow
Keras and TensorFlow are open source Python libraries for working with neural networks, creating machine learning models, performing deep learning and
In general, there are two ways to install Keras and TensorFlow:
Install a Python distribution that includes hundreds of popular packages (including Keras and TensorFlow) such as ActivePython.
Use pip to install TensorFlow, which will also install Keras at the same time.Pip Install TensorFlow
Instead of pip installing each package separately, the recommended approach is to install Keras as part of the TensorFlow installation. When you install TensorFlow 2.0+, Keras will be automatically installed, as well.
The simplest way to install TensorFlow is to install the binary version using one of the official releases on the Python Package Index (PyPI).
TensorFlow can be run on three different processor platforms, with the main difference being the speed at which your neural network will be trained. Each platform has different hardware requirements and offers different performance:
CPU – any modern computer can run this version, but it offers the slowest training speeds.
TPU – only available currently on Google’s Colaboratory (Colab) platform, Tensor Processing Units (TPUs) offer the highest training speeds.
GPU – most high end computers feature a separate Graphics Processing Unit (GPU) from Nvidia or AMD that offer training speeds much faster than CPUs, but not as fast as TPUs.
TensorFlow Requirements
TensorFlow and Keras require Python 3.6+ (Python 3.8 requires TensorFlow 2.2+) , and the latest version of pip. You can determine the version of Python installed on your computer by running the following command:
python3 --version
Output should be similar to:
Python 3.8.2
Run the following command to ensure that the latest version of pip is installed:
pip3 install --upgrade pip
To install TensorFlow for CPU and GPU processors, run the following command:
pip3 install tensorflow
pip3 install kerasIf you’re fine with using the CPU to train your neural network, your installation is done. If you want to use your GPU to the training, you’ll need to do the following:
For AMD GPUs, refer to the article Install Tensorflow 2 for AMD GPUs
For Nvidia GPUs:
Ensure you’re running a CUDA®-enabled card
Install v11 or later of the CUDA® Toolkit
If you’re working with Deep Neural Networks, you’ll should also install the latest version of the cuDNN library
The installation installs a slew of TensorFlow and Keras dependencies:
tensorflow
├── absl-py~=0.10
│ └── six
├── astunparse~=1.6.3
│ ├── six<2.0,>=1.6.1
│ └── wheel<1.0,>=0.23.0
├── flatbuffers~=1.12.0
├── gast==0.3.3
├── google-pasta~=0.2
│ └── six
├── grpcio~=1.32.0
│ └── six>=1.5.2
├── h5py~=2.10.0
│ ├── numpy>=1.7
│ └── six
├── keras-preprocessing~=1.1.2
│ ├── numpy>=1.9.1
│ └── six>=1.9.0
├── numpy~=1.19.2
├── opt-einsum~=3.3.0
│ └── numpy>=1.7
├── protobuf>=3.9.2
│ └── six>=1.9
├── six~=1.15.0
├── tensorboard~=2.4
│ ├── absl-py>=0.4
│ │ └── six
│ ├── google-auth-oauthlib<0.5,>=0.4.1
│ │ ├── google-auth>=1.0.0
│ │ │ ├── cachetools<5.0,>=2.0.0
│ │ │ ├── pyasn1-modules>=0.2.1
│ │ │ │ └── pyasn1<0.5.0,>=0.4.6
│ │ │ ├── rsa<5,>=3.1.4
│ │ │ │ └── pyasn1>=0.1.3
│ │ │ ├── setuptools>=40.3.0
│ │ │ └── six>=1.9.0
│ │ └── requests-oauthlib>=0.7.0
│ │ ├── oauthlib>=3.0.0
│ │ └── requests>=2.0.0
│ │ ├── certifi>=2017.4.17
│ │ ├── chardet<5,>=3.0.2
│ │ ├── idna<3,>=2.5
│ │ └── urllib3<1.27,>=1.21.1
│ ├── google-auth<2,>=1.6.3
│ │ ├── cachetools<5.0,>=2.0.0
│ │ ├── pyasn1-modules>=0.2.1
│ │ │ └── pyasn1<0.5.0,>=0.4.6
│ │ ├── rsa<5,>=3.1.4
│ │ │ └── pyasn1>=0.1.3
│ │ ├── setuptools>=40.3.0
│ │ └── six>=1.9.0
│ ├── grpcio>=1.24.3
│ │ └── six>=1.5.2
│ ├── markdown>=2.6.8
│ ├── numpy>=1.12.0
│ ├── protobuf>=3.6.0
│ │ └── six>=1.9
│ ├── requests<3,>=2.21.0
│ │ ├── certifi>=2017.4.17
│ │ ├── chardet<5,>=3.0.2
│ │ ├── idna<3,>=2.5
│ │ └── urllib3<1.27,>=1.21.1
│ ├── setuptools>=41.0.0
│ ├── six>=1.10.0
│ ├── tensorboard-plugin-wit>=1.6.0
│ ├── werkzeug>=0.11.15
│ └── wheel>=0.26
├── tensorflow-estimator<2.5.0,>=2.4.0
├── termcolor~=1.1.0
├── typing-extensions~=3.7.4
├── wheel~=0.35
└── wrapt~=1.12.1
Update Tensorflow and Keras Using Pip
If you already have TensorFlow and Keras installed, they can be updated by running the following command:
pip3 install -U tensorflow
You can verify the TensorFlow installation with the following command:
python -m pip show tensorflow
Output should be similar to:
Name: tensorflow
Version: 2.2.0
Summary: TensorFlow is an open source machine learning framework for everyone.
Home-page: https://www.tensorflow.org/
Author: Google Inc.
Author-email: packages@tensorflow.org
License: Apache 2.0
Location: c:\python38\lib\site-packages
Requires: google-pasta, gast, six, protobuf, tensorboard, h5py, termcolor, absl-py, opt-einsum, wrapt, grpcio, keras-preprocessing, tensorflow-estimator, numpy, astunparse, wheel, scipy
Required-by:
If you intend to create plots based on TensorFlow and Keras data, then consider installing Matplotlib. For information about Matplotlib and how to install it, refer to What is Matplotlib in Python?
How to Import Keras and TensorFlow
Once TensorFlow and Keras are installed, you can start working with them.
# Begin a Keras script by importing the Keras library:
import keras
or
from tensorflow import keras
# Import TensorFlow:
import tensorflow as tf
It’s not necessary to import all of the Keras and Tensorflow library functions. Instead, import just the function(s) you need for your project.
# Import the Sequential model class from Keras
# to form the framework for a Sequential neural network:
from keras.models import Sequential
For more information on working with Keras, refer to What is a Keras Model. <– link needed
The following tutorials will provide you with step-by-step instructions on how to work with machine learning Python packages:
Get a version of Python, pre-compiled with Keras and other popular ML Packages
ActiveState Python is the trusted Python distribution for Windows, Linux and Mac, pre-bundled with top Python packages for machine learning – free for development use.
Some Popular ML Packages You Get Pre-compiled – With ActiveState Python
Machine Learning:
TensorFlow (deep learning with neural networks)*
scikit-learn (machine learning algorithms)
keras (high-level neural networks API)
Data Science:
pandas (data analysis)
NumPy (multidimensional arrays)
SciPy (algorithms to use with numpy)
HDF5 (store & manipulate data)
matplotlib (data visualization)