Deep Learning Journey

Welcome to my Deep Learning Journey repository! This repository documents my exploration and learning in the field of deep learning. It includes various projects, experiments, and notes on key concepts and techniques in deep learning.

Table of Contents

• Introduction
• Projects
• Installation
• Usage
• Contributing
• License

Introduction

This repository is a collection of my deep learning projects and experiments. It includes implementations of different neural network architectures, solutions to common deep learning tasks, and notes on theoretical concepts. The goal is to build a comprehensive understanding of deep learning through hands-on practice and experimentation.

Projects

1. MNIST Digit Classification

Implementation of a neural network to classify handwritten digits from the MNIST dataset using TensorFlow and Keras.

• Folder: mnist-digit-classification
• Description: Trains a neural network to classify images of handwritten digits.
• Technologies: TensorFlow, Keras, Python
• Notebook: MNIST_Digit_Classification.ipynb

2. MNIST Image Classification

Implementation of a neural network to classify fashion items from the MNIST dataset using TensorFlow and Keras.

• Folder: mnist-image-classification
• Description: Trains a neural network to classify images of fashion items.
• Technologies: TensorFlow, Keras, Python
• Notebook: image-classifer.ipynb

3. Loading Datasets using tensorflow API

Exploring tf.data API for creating efficient and scalable data input pipelines, including functions for reading, transforming, shuffling, batching, and prefetching datasets.

• Folder: Loading-the-data-using-tf
• Description: Going through a tf.data API.
• Technologies: TensorFlow, Keras, Python
• Notebook: Loading_data_with_tf.ipynb

4. CNN Architectures

Exploring and implementing various CNN architectures using keras library

• Folder: CNN-Architectures
• Description: Implementing various types of CNN architectures.
• Technologies: TensorFlow, Keras, Python, numpy

5. Implementing CNN from scratch

Tried to implement the CNN model from scratch to understand it better.

• Folder: Implementing-CNN-from-Scratch
• Description: Trains a convolutinonal neural network to classify images or any pictorial data.
• Technologies: numpy
• Notebook: CNN_test.ipynb

5. Solving GTSRB dataset using CNN model

Developed a Convolutional Neural Network (CNN) model to classify German traffic signs using the GTSRB dataset.

• Folder: GTSRB_German-Traffic-Sign-Classification
• Description: Trains a Convolutional Neural Network to classify traffic signs into their respective categories using the GTSRB dataset.
• Technologies: numpy, tensorflow, flask, pandas,
• Notebook: GTSRB.ipynb

6. Cats vs. Dogs Classification using CNN

A CNN model that can classify the images of cats and dogs

• Folder: cats-and-dogs
• Description: Trains a Convolutional Neural Network to classify dogs and cats.
• Technologies: numpy, tensorflow, PIL, matplotlib
• Notebook: cats-and-dogs.ipynb

7. Vision Transformer (ViT) from Scratch

A Vision Transformer model built from scratch for image classification.

• Folder: vit-from-scratch
• Description: Implements a Vision Transformer model from scratch for image classification tasks. The project includes data preprocessing, model training, and evaluation on benchmark datasets.
• Technologies: numpy, tensorflow, matplotlib, scikit-learn, pytorch
• Pytorch Notebook: ViT.ipynb
• TensorFlow Notebook: ViT-tf.ipynb

8. Fake News Classifier using LSTM

A LSTM Model built for classifiying whether the article is real or fake

• Folder: Fake-News-Classifier
• Description: Implemented a Long Short-Term Memory (LSTM) deep learning model to classify news articles as real or fake. The project involves text preprocessing, tokenization, vectorization, model training, and evaluation.
• Technologies: TensorFlow, pandas, scikit-learn, NLTK
• Notebook: fake-news.ipynb

9. Crypto Currency Price Prediction

A LSTM Model built for predict the price of cryptocurrency using historical data

• Folder: CryptoCurrency-Prediction
• Description: Built a deep learning model for Crypto Price Prediction using LSTM networks to predict future cryptocurrency prices based on historical data and market trends
• Technologies: TensorFlow, pandas, scikit-learn,LSTM, numpy
• Notebook: crypto-pred.ipynb )

10. Character-Level Recurrent Sequence-to-Sequence Model (Eng -> Deu words translation)

A seq2seq model built for predict the translation of words from English to German

• Folder: Seq-to-Seq-model`
• Description: Built a deep learning model for character level translation
• Technologies: TensorFlow, pandas, keras, LSTM, numpy
• Notebook: seq-to-seq.ipynb

Installation

To set up the environment and run the notebooks locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/anson10/deep-learning-journey.git
   cd deep-learning-journey

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows, use `venv\Scripts\activate`

3. Install the required dependencies:

   pip install -r requirements.txt

Usage

To run the notebooks:

1. Activate the virtual environment (if not already activated):

   source venv/bin/activate  # On Windows, use `venv\Scripts\activate`

2. Launch Jupyter Notebook:

   jupyter notebook

3. Open the desired notebook from the Jupyter interface and run the cells.

Contributing

Contributions are welcome! If you have suggestions for improvements, new projects, or bug fixes, please open an issue or submit a pull request. Make sure to follow the existing code style and include relevant tests.

License

This repository is licensed under the MIT License. See the LICENSE file for more details.

---

Happy Learning!