keras setup

README.md

Getting started with Machine Learning in Python

Configuring Local Dev Environment

Install anaconda from source: https://www.anaconda.com/download/#macos then restart terminal and check to see if it is installed:

conda --version #> conda 4.5.11
which conda #> /anaconda3/bin/conda

Create a new virtual environment, named something like ml-env-1, with the necessary python packages:

conda create -n ml-env-1 python=3.6 # installing packages here seems to not work, also version 3.7 not yet supported by tensorflow, so use python 3.6
conda env list #> you should see your new environment included

Enter the virtual environemnt:

conda activate ml-env-2  # ... to deactivate: conda deactivate

which python #> /anaconda3/envs/ml-env-2/bin/python
python --version #> Python 3.6.7 :: Anaconda, Inc.

which pip #> /anaconda3/envs/ml-env-2/bin/pip
pip --version #> pip 18.1 from /anaconda3/envs/ml-env-2/lib/python3.6/site-packages/pip (python 3.6)

Install package dependencies inside the virtual environment:

pip install keras tensorflow

pip list #> should see both, with supporting packages

Creating a Model

python my_model.py

CREDITS.md

Credits, notes and reference

Python / General

• https://www.dataweekends.com/blog/2017/03/09/set-up-your-mac-for-deep-learning-with-python-keras-and-tensorflow
• https://github.com/prof-rossetti/nyu-info-2335-201805/blob/master/notes/programming-languages/python/modules/pdb.md
• https://github.com/prof-rossetti/nyu-info-2335-201805/blob/master/notes/programming-languages/python/datatypes/strings.md

Anaconda

• https://www.anaconda.com/download/#macos
• https://docs.anaconda.com/anaconda/
• https://docs.anaconda.com/_downloads/Anaconda-Starter-Guide-Cheat-Sheet.pdf

Keras

• https://github.com/keras-team/keras
• https://keras.io/backend/
• https://keras.io/#configuring-your-keras-backend
• keras-team/keras#4889
• https://transcranial.github.io/keras-js/#/

Tensorflow

• https://www.tensorflow.org/install/source -- can run on local mac laptop without crashing? let's see how resource intensive this is...
• https://stackoverflow.com/questions/38896424/tensorflow-not-found-using-pip not yet on Python 3.7... use 3.6...

Datasets

• https://keras.io/datasets/

MNIST Handwritten Digits

• https://keras.io/datasets/#mnist-database-of-handwritten-digits
• https://nextjournal.com/gkoehler/digit-recognition-with-keras
• https://topicfly.io/classify-mnist-digits-keras/
• https://medium.com/@cafielo/build-a-handwritten-digit-recognition-model-with-keras-b8733274574c
• https://medium.com/coinmonks/handwritten-digit-prediction-using-convolutional-neural-networks-in-tensorflow-with-keras-and-live-5ebddf46dc8

Numpy / Scipy

N-dimensional Data arrays:

• https://docs.scipy.org/doc/numpy-1.13.0/reference/arrays.ndarray.html
• https://docs.scipy.org/doc/numpy/reference/generated/numpy.ndarray.shape.html

Matplotlib

• https://stackoverflow.com/questions/21784641/installation-issue-with-matplotlib-python#comment56913201_21789908
• https://stackoverflow.com/a/35085705/670433

my_model.py

import pdb
from keras.models import Sequential
from keras.datasets import mnist

# from matplotlib import pyplot as plt #> ImportError: Python is not installed as a framework....
# h/t: https://stackoverflow.com/questions/21784641/installation-issue-with-matplotlib-python#comment56913201_21789908
import matplotlib
matplotlib.use('TkAgg')
import matplotlib.pyplot as plt

print("--------------------")
print("LOADING DATA...")
print("--------------------")

# h/t: https://keras.io/datasets/#mnist-database-of-handwritten-digits
(x_train, y_train), (x_test, y_test) = mnist.load_data()

print(f"  + TRAINING/INPUTS: {type(x_train)} {x_train.shape} {x_train.dtype}") #> <class 'numpy.ndarray'> (60000, 28, 28) uint8
print(f"  + TRAINING/OUTPUTS: {type(y_train)} {y_train.shape} {y_train.dtype}") #> <class 'numpy.ndarray'> (60000,) uint8
print(f"  + TESTING/INPUTS: {type(x_test)} {x_test.shape} {x_test.dtype}") #> <class 'numpy.ndarray'> (10000, 28, 28) uint8
print(f"  + TESTING/OUTPUTS: {type(y_test)} {y_test.shape} {y_test.dtype}") #> <class 'numpy.ndarray'> (10000,) uint8

print("--------------------")
print("EXAMPLE TRAINING DATA INAGES...")
print("--------------------")

# h/t: https://topicfly.io/classify-mnist-digits-keras/
plt.subplot(221)
plt.imshow(x_train[0], cmap=plt.get_cmap('gray'))
plt.subplot(222)
plt.imshow(x_train[1], cmap=plt.get_cmap('gray'))
plt.subplot(223)
plt.imshow(x_train[2], cmap=plt.get_cmap('gray'))
plt.subplot(224)
plt.imshow(x_train[3], cmap=plt.get_cmap('gray'))

plt.show() # h/t: https://stackoverflow.com/a/35085705/670433

#pdb.set_trace()

print("--------------------")
print("PROCESSING DATA...")
print("--------------------")

# todo

print("--------------------")
print("CREATING MODEL...")
print("--------------------")

model = Sequential()

print(type(model)) #> <class 'keras.engine.sequential.Sequential'>