Simple Neural Network for Cat vs Dog Classification

This script is designed to train a simple neural network to classify images as either cats or dogs using a dataset of images. It utilizes TensorFlow and Keras for model creation and training.

Usage

Prerequisites

• Python 3.x
• TensorFlow
• Keras
• scikit-learn
• NumPy
• matplotlib
• imutils
• OpenCV

Installation

1. Clone this repository:

   git clone https://github.com/Sebastian352/Simple_Binary_NN

2. Navigate to the directory:

   cd repository

3. Install dependencies:

   pip install -r requirements.txt

Running the Script

Use the following command to run the script:

python script.py --dataset path/to/dataset --model path/to/model --label-bin path/to/label/binarizer --plot path/to/plot

Replace path/to/dataset, path/to/model, path/to/label/binarizer, and path/to/plot with relevant file paths.

Arguments

• -d, --dataset: Path to the input dataset of images (required).
• -m, --model: Path to the output trained model (required).
• -l, --label-bin: Path to the output label binarizer (required).
• -p, --plot: Path to the output accuracy/loss plot (required).

Dataset

You can download a relevant dataset from this link.

Script Details

This script performs the following steps:

1. Loads the images from the dataset.
2. Preprocesses the images and labels.
3. Splits the dataset into training and testing sets.
4. Creates a simple neural network model using Keras.
5. Trains the model on the training data.
6. Evaluates the trained model on the testing data.
7. Generates a plot showing training and validation loss/accuracy.
8. Saves the trained model, label binarizer, and plot to specified paths.

Acknowledgements

• This script is adapted from a tutorial on Kaggle.
• The dataset used for training the model can be found on Kaggle.

License

MIT License