[PYTHON] Things to keep in mind when deploying Keras on your Mac

Introduction

When I introduced Keras (tensorflow) to my Mac, I ran into various things, so I summarized it as a memorandum.

environment

Mac OSX Mojave(10.14.6) Python3.7.3

Install tensorflow and keras

Execute the following command in the terminal. If the version of pip is old, it will not work well, so you need to upgrade the version of pip first **

pip install --upgrade pip pip3 install --upgrade tensorflow pip3 install keras

After the above, actually launch python3 and check the operation.

Python 3.7.3 (v3.7.3:ef4ec6ed12, Mar 25 2019, 16:52:21) 
[Clang 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow
>>> import keras
Using TensorFlow backend.
>>> 

I was able to install it safely.

mnist.load_data () doesn't work

import keras
from keras.datasets import mnist
(train_images, train_labels), (test_images, test_labels) = 
mnist.load_data()

It is a familiar mnist with the notation such as, but if you do mnist.load_data () with python3, the following (although it is abbreviated notation)

Exception: URL fetch failure on https://s3.amazonaws.com/img-
datasets/mnist.npz : None -- [SSL: CERTIFICATE_VERIFY_FAILED] 
certificate verify failed: unable to get local issuer certificate 
(_ssl.c:1056)

I get an error. This is the same for cifar10.load_data (). Solved by referring to here. The cause is

OpenSSL installed by default on macOS is too old, so since Python 3.6, the installer for macOS includes OpenSSL and the system's OpenSSL is no longer referenced.

And that. You can download the certifi module by executing the following code in the terminal and refer to the root certificate included in it.

cd /Applications/Python\ 3.7/Install\ Certificates.command

from now on

For deep learning in Python, I referred to the summary of here. In addition, we will update it as needed when problems occur.