/deeplearning-solutions

Hands on experience with neural network and CNN coding with Keras and finding solution for business problem.

Primary LanguagePythonMIT LicenseMIT

Deep Learning by Examples: Creating neural network and CNN using Keras Python library.

In this project, We have solved two different business problems using Artificial Neural Network and Covolution Neural Network.

From bank dataset We are trying to predict whether user will subscribe term deposit or not using Artificial Neural Network.

For image classification, we have image dataset of driver images captured at cross-roads. Using Convolution Neural Network we will classify that driver is driving safely or looking by side or talking on mobile. We will classify image into 9 different classes with simple CNN architecture which can be run on your CPU.

For classification task you can also use saved CNN model.

  • ANN for client will subscribe term deposit or not?
  • CNN for classification of driver images into 9 different categories for safe driving.

Installation

To start with project just follow the few steps

$ git clone https://github.com/keyurr2/deeplearning-solutions.git
$ pip install -r requirements.txt
$ cd into algo folder
$ python ann.py 

This will install python libraries required to start with Deep Learning like Tensorflow and Keras

NOTE: We are using Python 3 in this project.

Set up the project in anaconda

The first step is to download the anaconda for your system.

Now activate anaconda virtual environment and then launch anaconda navigator. Find below command for this.

$ source anaconda3/bin/activate
$ anaconda-navigator  

This will launch a web UI for anaconda from where you can open Spyder IDE to run code examples in this project.

Authors

License

This project is licensed under the MIT License - see the LICENSE.md file for details