I personally summarized the installation of Anaconda 3 on Windows 10.
Python has many useful libraries that make it easy for you to do advanced numerical calculations, data analysis, machine learning, and more. However, many of these libraries are not included as standard with Python, and you have to install them yourself to prepare the environment.
It's easy for people who are used to Python or who are already using it, but for those who are new to Python, building an environment is surprisingly troublesome. On the other hand, Anaconda is a Python package that includes Python itself and libraries that are often used in Python. Just installing Anaconda will install Python itself and the library, so it will be much easier to build the environment.
On the other hand, for engineers who are accustomed to python, using Anaconda has the disadvantage of not being able to build their own Python environment. For beginners, the benefits of using Anaconda may be greater.
・ Windows10 Home (64bit) ・ Anaconda3 2019.10 (Python 3.7)
Go to the Anaconda3 Download Site.
The person who installs this time is Python3, so click [Download] in [Python 3.7 Version]. ** Python 2 series will end support on January 1, 2020 **
Run when downloaded
When the setup wizard starts, click [Next>]
Look at the license terms and click [I Agree].
[Just Me (recommended)] if you want to install only for individual users [All Users (requires admin privileges)] to install for all users
This time, I selected [All Users (requires admin privileges)].
If you are not particular about it, just click [Next>]. If you want to change it, click [Browse] and specify the folder you want to install.
If you check the first [Add Anaconda to the system PATH environment variable], the directory containing the Python executable file installed by Anaconda will be added to the PATH environment variable. This makes it possible to execute Anaconda's Python as it is with cmd etc.
For the second [Register Anaconda as the system Python 3.7], if you check it, other programs that use Python will use Python installed by Anaconda.
Neither setting is a problem if you only use Python installed with Anaconda. ** Please note that if you have already installed Python in another way, or if you try to install Python separately later, you may have trouble. ** **
This time I checked only the second one. Click [Install] to start the installation.
Wait patiently, and if you can install it, you're done.
Press the Start Menu at the bottom left to launch Anacondqa Powershell Prompt. Anaconda3 (64-bit) -> Anaconda Powershell Prompt
Python version
# --V is uppercase.
python --V
(base) PS C:\Users\USERNAME> python --V Python 3.7.4
If you can confirm the version, it's OK.
conda
conda info
If no error message appears, it's OK.
conda
python -c "import numpy; print( numpy.__version__ )"
(base) PS C:\Users\USERNAME> python -c "import numpy; print( numpy.version )" 1.16.5
If the version is displayed, it's OK.
Press the Start Menu at the bottom left, right-click Anacondqa Powershell Prompt-> Other-> Run as administrator
In the newly opened Anacondqa Powershell Prompt, update the Anaconda3 conda package and delete the old conda package file.
[conda-forge]If there is, delete it
#The following command will be sent to the conda channel[conda-forge]It is a command to delete if is included.
#[conda-forge]Should not be included, but it is running just in case.
conda config --remove channels conda-forge
Package update
conda upgrade --all
#When I get a question asking y or n, I want to continue[y]
Remove old packages
conda clean --packages
#When I get a question asking y or n, I want to continue[y]
pip update
pip install --upgrade pip
(base) PS C:\Windows\system32> pip install --upgrade pip Requirement already up-to-date: pip in c:\programdata\anaconda3\lib\site-packages (19.3.1)
conda install -y -c spyder-ide spyder
conda install -y numpy scipy h5py scikit-learn scikit-image matplotlib seaborn pandas pillow
conda install -y jupyter pytest docopt pyyaml cython
pip install opencv-python
pip install opencv-contrib-python
pip install chainer
pip install chainerrl
pip install tensorflow==2.0.0
pip install keras
pip install natsort
Recommended Posts