tomarakanksha / CNN_Model

Hello world model of Convolution Neural Network!

Convolutional Neural Network (CNN) Project

This project is focused on learning and implementing Convolutional Neural Networks (CNNs) using PyTorch for image classification. I have used the MNIST dataset for this task, which consists of hand-written digits. The project is divided into two files, each representing a different stage of the CNN implementation.

File 1(CNN_Model_Breakdown): Exploring CNN Basics

• In "CNN_Model_Breakdown.ipynb," I have created the breakdown of all the fundamental components of a CNN. The following key steps are performed:

Step 1: Data Preparation

• Used PyTorch and the torchvision library to download and prepare the MNIST dataset for training and testing. The data is converted into tensors.

Step 2: Data Loading

• Created DataLoader objects for both the training and test datasets, enabling us to efficiently load batches of data for training and evaluation.

Step 3: Defining the CNN Model

• Defined a simple CNN model with two convolutional layers (conv1 and conv2) to learn feature representations from the input images.

Step 4: Understanding Convolution and Pooling

• Explained the concept of convolution and pooling in CNNs, demonstrating their impact on the dimensions of the feature maps.

Step 5: Processing a Single Image

• Performed the initial processing of a single MNIST image using the defined model, showing the changes in the image dimensions after each layer.

File 2(CNN_Model): Additional Experiments

• In "CNN_Model.ipynb," I have created a model for exploring CNN concepts and demonstrate additional experiments or code snippets.

Model Structure

• Created a sample Model class with 2 convolutional layer along with maxpooling layers 3 Fully-connected layers. And then feed the training data to the model and implemented an end to end CNN model.

Visualizing Data

• Vizualized the data or display intermediate results to enhance understanding.

Prerequisites

• To run these scripts, you should have Python installed along with the required libraries, including PyTorch, NumPy, and Matplotlib.

Getting Started

• To get started, clone this repository and execute the Jupyter notebooks "CNN_Model_Breakdown.ipynb" and "CNN_Model.ipynb." Make sure you have the necessary dependencies installed.

git clone https://github.com/your-username/cnn-project.git
cd CNN-MODEL
jupyter notebook file1.ipynb