Skip to content

Comparative analysis of Stochastic Optimization methods with binary classification using a CNN through Python, Tensorflow, Keras, and GPU.

Notifications You must be signed in to change notification settings

hchen98/dtsc615_ImageClassification

Repository files navigation

DTSC 615 - Cat vs Dog Image Classification

Course: Optimization Method for Data Science (DTSC 615/M01) - Fall 2021
Instructor: Dr. Jerry Cheng
Team Memebers: Michael Trzaskoma, Hui (Henry) Chen, Bofan He, Ephraim Hallford

Comparative analysis of Stochastic Optimization methods with binary classification using a CNN through Python, Tensorflow, Keras, and GPU.


File Structure

.
├── assets
├── dataset
│   ├── test
│   │   ├── cat
│   │   └── dog
│   ├── test1
│   ├── train
│   │   ├── cat
│   │   └── dog
│   └── val
│       ├── cat
│       └── dog
└── saved_models


Get Started

In this project, we analysis the different Stochastic Optimization methods such as Adagrad, Adadelta, RMSprop, Adam and Adamax with a Convolutional Neural Network (CNN) application.


Results

Optimizer Classification Accuracy

CNN Accuracy


Binary Cross Entropy Loss

CNN Loss


Adagrad Confusion Matrix

Adagrad Confusion Matrix


Adadelta Confusion Matrix

Adadelta Confusion Matrix


RMSprop Confusion Matrix

RMSprop Confusion Matrix


Adam Confusion Matrix

Adam Confusion Matrix


Adamax Confusion Matrix

Adamax Confusion Matrix

About

Comparative analysis of Stochastic Optimization methods with binary classification using a CNN through Python, Tensorflow, Keras, and GPU.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published