CANN - Compression Algorithms for Neural Networks

Directory for tracking progress in the thesis project on compression algorithms for neural networks.

OVERVIEW

1. basic_python : learning Python.
2. neural_network : building simple neural networks.
3. tensorflow : understanding TensorFlow and machine learning.
4. compression_algorithms : compression algorithms for neural network.

REQUIREMENTS

Python 3 [3.6.5] and PIP 3 [10.0]
brew install python3

Python modules from PIP 3

pip3 install <module_name>

H5py [2.7.1]
iPython [6.3]
iPhyton notebook
Jupyter [4.4]
Keras [2.1.5]
MatPlotLib [2.2.2]
NumPy [1.14.2]
OpenCV [3.4.0.12] (opencv-python)
Pandas [0.22]
SciKit-Learn [0.19.1]
SciPy [1.0.1]
Seaborn [0.8.1]
TensorFlow [1.7]
TFLearn [0.3.2]
TQDM [4.23]

Anaconda and Environment

Downloading and Installing Anaconda
Download the package from the official site and follow the instruction.

Setting up the environment

# Creating the environment
conda create -n cann python=3.6.5 pandas scikit-learn jupyter matplotlib

# Activate the environment
conda activate cann

# Install Tensorflow and Keras
conda install tensorflow
conda install keras

# Launch Jupyter notebook
jupyter notebook

# Deactivate the environment
conda deactivate