Skip to content

"A collection of deep learning projects from my PGD in Data Science at NED University, including CNN-based image classification and sequence processing with ConvNets, applied to datasets like Cat-Dog, Netflix, and Jena Climate."

Notifications You must be signed in to change notification settings

SheemaMasood381/Deep-Learning

Repository files navigation

Deep Learning Repository

This repository contains my deep learning projects from my Post Graduate Diploma (PGD) in Data Science at NED University. The projects involve applying deep learning techniques to different datasets, implementing Convolutional Neural Networks (CNN) for classification tasks, and experimenting with sequence processing using ConvNets.

CNN Architecture

This is the Convolutional Neural Network architecture used in the project.

CNN Architecture

Projects

1. Cat-Dog CNN Classifier

A Convolutional Neural Network (CNN) classifier to distinguish between images of cats and dogs. The model achieves 97% accuracy and uses visualizations like confusion matrices and F1 scores for performance analysis. The notebook contains the full implementation of the model.

2. Sequence Processing with ConvNets (Chollet DL Book - Chapter 6)

This project focuses on implementing sequence processing models using ConvNets, inspired by Chapter 6 of François Chollet’s Deep Learning with Python. The assignment involves experimenting with two datasets:

  • Netflix Dataset: I replaced the IMDB dataset used in the original notebook and applied Conv1D sequence processing for multi-class classification (3 classes).
  • Jena Climate Dataset (2009-2016): This dataset required 2D ConvNets and a GRU-based model for better sequence processing. I modified the architecture and hyperparameters to improve performance.

Key Modifications:

  • Updated the input layers to fit the dataset format.

  • Changed the output layer to softmax for multi-class classification (Netflix data).

  • Used GRU layers for sequence processing (Jena Climate data) instead of the original Conv1D.

  • Adjusted the architecture and hyperparameters to improve results.

  • View Project Notebook

CNN Images

This folder contains various visualizations related to CNN architectures, model performance (confusion matrix, etc.), and other resources related to sequence processing with ConvNets.

Getting Started

To start working with these projects, follow these steps:

  1. Clone this repository:
    git clone https://github.com/SheemaMasood381/Deep-Learning.git
    

Dependencies

  • Python 3.x
  • TensorFlow/Keras
  • NumPy
  • Pandas
  • Matplotlib
  • Scikit-learn

Author

Sheema Masood
GitHub Profile

Feel free to explore, contribute, and provide feedback on the projects in this repository!

About

"A collection of deep learning projects from my PGD in Data Science at NED University, including CNN-based image classification and sequence processing with ConvNets, applied to datasets like Cat-Dog, Netflix, and Jena Climate."

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published