A hands-on collection of Deep Learning resources —> perfect for beginners and intermediate learners looking to build real intuition. If you're learning PyTorch or need quick refreshers while working on deep learning projects, this repo's designed to be a useful companion.
This repository documents my daily progress as I learn and build with PyTorch, exploring core deep learning concepts. This repo serves as a hands-on learning journal and will be covering everything from tensors to neural networks , cv and beyond that i learn from Daniel Bourke's course
| Day | Topic Covered | Description |
|---|---|---|
| 1 | Tensors | Creation, indexing, operations, reshaping , viewing and stacking ,Squeezing, unsqueezing and permuting |
| 2 | Fundamentals | PyTorch and NumPy, Reproducibility, Accessing a GPU, Setting up device agnostic code , Exercise Questions |
| 3 | Workflow | Creating a dataset with linear regression,Creating training and test sets ,Creating first PyTorch model, Making predictions with our model ,Training a model with PyTorch , Setting up a loss function and optimizer, PyTorch training loop intuition, Running our training loop epoch by epoch , Writing testing loop code ,Saving/loading a model , Putting everything |
| 3.1 | Exercises | Exercises related to topics covered in 3 |
| 4 | Neural Network | NN with PyTorch -> Linear Regression , Binary Classification , Multiclass Classification models |
| 4.1 | Exercises | Exercises related to topics covered in 4 |
| 5 | Computer Vision | TorchVision , Convolutional neural networks , FashionMNIST , predictions with a confusion matrix |
| 5.1 | Exercises | MNIST , Exercises related to topics covered in 5 |
| 6 | Custom Datasets | Downloading a custom dataset of pizza, steak and sushi images, Creating image DataLoaders, Creating a custom dataset class (overview), Writing a custom dataset class from scratch , Turning custom datasets into DataLoaders , Data augmentation, Building a baseline model, Predicting on custom data |
| 7 | PyTorch going modular | |
| 8 | Transfer Learning | Getting a pretrained model , Which pretrained model should you use? , Setting up a pretrained model , Getting a summary of our model with , Freezing the base model and changing the output layer to suit our needs , Train model |
| 9 | Experiment Tracking | Train model and track results , Adjust function to track results with, View our model's results in TensorBoard ,Create a helper function to build instances , Update the function to include a parameter , Setting up a series of modelling experiments , View experiments in TensorBoard, Load in the best model and make predictions with it |
This contains relevant .ipynb files along with my notes and experiments.
- Python 3.x
- PyTorch
- Jupyter Notebook / Google colab
- NumPy (where needed)
- Tensors & operations
- Automatic differentiation (autograd)
- Neural network architecture basics
- Training loops, optimizers, loss functions
- CNNs, RNNs, and model evaluation (coming soon)
- GitHub: mahidhiman12
- LinkedIn: Mahi Dhiman
“Consistency over perfection - one model, one bug, one step at a time.” 🔁