Introduction to Deep Learning - Course

This repository contains all my homework assignments' answers and final project work done during the course. Each assignment and project phase has been meticulously documented and includes both the Persian and English slides provided by the instructor.

Homework Assignments

HW1:

• Topics Covered:
  • Multi-Layer Perceptrons (MLPs)
  • Various Activation Functions

HW2:

• Topics Covered:
  • Overfitting
  • Dropout Layer
  • Regularization Methods
  • Transfer Learning

HW3:

• Topics Covered:
  • Optimizers
  • Learning Rate Schedules
  • Convolutional Neural Networks (CNNs)
  • Convolutional Layers
  • Inception Modules

HW4:

• Topics Covered:
  • CNN Networks using KerasTuner
  • Pooling Layers
  • Recurrent Neural Networks (RNNs)
  • Long Short-Term Memory Networks (LSTMs)
  • Batch Normalization
  • Gradient-weighted Class Activation Mapping (Grad-CAM)

HW5:

• Topics Covered:
  • Different Types of RNNs
  • Backpropagation Formulas
  • Attention Mechanism

HW6:

• Topics Covered:
  • Evaluation Metrics: Accuracy, Recall, Precision, and F1 Score
  • Generative Adversarial Networks (GANs)

Each homework folder contains the respective Jupyter notebooks with my solutions and detailed explanations.

Final Project

Phase One:

• Implemented a multi-label classification model using the ARMAN-EMO dataset.

Phase Two:

• Developed a multi-modal classification model combining ResNet for feature extraction and a customized MLP for classification.

Submodules

• Existing Repo

Course Materials

This repository also includes the instructor's slides in both Persian and English, which were used throughout the course for lectures and explanations.