Neural Network Implementations in PyTorch and TensorFlow

Welcome to the Neural Network Implementations repository. This repository contains various implementations of common neural network architectures using both PyTorch and TensorFlow frameworks. Each implementation is organized into a separate file for ease of use and understanding.

Repository Structure

• Autoencoder
• Image Classification using ConvNets
• GAN
• K-Folds Autotuning
• Transfer Learning
• More on CNN's
• Object Detection (YOLO)

Directories and Files

• going_modular/: This directory is reserved for future modular code development.
• LOADING_DATA.py: A script for loading datasets used across different implementations.
• autoencoder_tf.py: Implementation of an autoencoder using TensorFlow.
• autoencoder_torch.py: Implementation of an autoencoder using PyTorch.
• catsDogs_torch.py: PyTorch implementation for classifying cats and dogs.
• cats_tf.py: TensorFlow implementation for classifying cats and dogs.
• cifar100_torch.py: PyTorch implementation for classifying CIFAR-100 dataset.
• cifar10_torch.py: PyTorch implementation for classifying CIFAR-10 dataset.
• gan_pytorch_cifar.py: Implementation of a GAN using PyTorch for the CIFAR dataset.
• gan_torch_mnist.py: Implementation of a GAN using PyTorch for the MNIST dataset.
• helper_functions.py: Contains helper functions used across different scripts.
• mnist_kfolds_torch.py: PyTorch implementation for MNIST classification with k-fold cross-validation.
• mnist_torch.py: Basic PyTorch implementation for MNIST classification.
• transfer_cifar10.py: TensorFlow implementation for transfer learning on CIFAR-10 dataset.
• transfer_learning_torch_extra.py: Extra transfer learning experiments using PyTorch.
• visualizing_convolutions_torch.py: PyTorch script for visualizing convolutional layers.
• yolo_torch.py: PyTorch implementation of the YOLO object detection algorithm.

Getting Started

To get started with these implementations, clone the repository to your local machine:

git clone https://github.com/your_username/neural-networks-implementations.git
cd neural-networks-implementations`

Requirements

Make sure you have the following libraries installed:

• Python 3.x
• PyTorch
• TensorFlow
• NumPy
• Matplotlib

You can install the required libraries using pip:

pip install torch tensorflow numpy matplotlib

Usage

---

Each script is standalone and can be run individually. For example, to run the PyTorch autoencoder implementation:

python autoencoder_torch.py

Or to run the TensorFlow cats vs. dogs classifier:

python cats_tf.py

Contributing

---

Contributions are welcome! If you have any improvements or additional implementations, feel free to open a pull request. Please make sure to follow the existing code style and include appropriate comments and documentation.

License

---

This project is unlicensed.

Acknowledgements

---

Thanks to the contributors and the open-source community for their invaluable support and contributions.