This repository documents a complete Deep Learning introductory concepts, covering everything from basic neural networks to advanced Transfer Learning using pre-trained CNNs.
Each part builds upon the previous one, introducing new TensorFlow and Keras concepts through hands-on experiments and well-structured notebooks.
| Part | Topic | Key Concepts |
|---|---|---|
| Part 1 | Neural Network Basics | Dense layers, activation functions, backpropagation |
| Part 2 | Regularization & Dropout | Overfitting control, L1/L2 regularization, Dropout |
| Part 3 | Optimizers & Loss Functions | SGD, Adam, RMSprop, learning rate tuning |
| Part 4 | Convolutional Neural Networks (CNNs) | Conv2D, MaxPooling, Flatten, feature extraction |
| Part 5 | Data Augmentation | Image transformations, preventing overfitting |
| Part 6 | Model Saving & Checkpointing | .h5, .weights.h5, .keras, checkpoints |
| Part 7 | Model APIs | Sequential, Functional, and Subclassing APIs |
| Part 8 | Transfer Learning | MobileNetV2 (achieved >85% accuracy) , feature extraction, fine-tuning |
- Language: Python 3.x
- Framework: TensorFlow / Keras
- Libraries: NumPy, Pandas, Matplotlib, scikit-learn
- Environment: Jupyter Notebook