souravdutta91/pytorch-tutorial

A comprehensive tutorial in Pytorch

Pytorch A-Z

This is a comprehensive tutorial for Pytorch. Each and every topic is implemented as a separate Jupyter Notebook. If you are interested in the topics, download/clone/fork this repository and then try out the code in the different notebooks. This repository was inspired from various online courses and tutorials on Pytorch, without which it probably would not have been possible to build this. We recommend you learn from the notebook files first, and then try to implement the code by yourself in your local notebooks or python scripts. And if you like the content, feel free to star the repository and share it with others. Now, let's learn Pytorch together!

---

Requirements

You will need to install different python libraries like torch, jupyter notebook, numpy, pandas, matplotlib to name a few. However, we recommend you to rather install a complete bundle like Anaconda which comes with all the necessary libraries pre-installed.

Seems like a lot of work? If you do not want to install the necessary libraries in your system, you could still practice all the topics online through Colab.

---

NOTE: We will start with the basic concept of tensors, go through various operations in torch framework, and then gradually learn how to contruct deep neural networks in Pytorch. If you are beginner, we recommend you to start with the first chapter Introduction to tensors and continue sequentially. However, you already have some knwoledge of the basics, you could choose any of the topics you want to explore.

---

Contents

-- coming soon --

1. Introduction to tensors
2. 2D tensors and basic operations
3. Advanced operations
4. Linear Regression
5. Gradient Descent & Optimization
6. Classification
7. Simple Neural Networks, Backpropagation & Activations
8. Deep Neural Networks, Dropout
9. Batch Normalization
10. Convolutional Neural Networks
11. Sequential Neural Networks

Feedback

Your feedback is important and is always welcome. Feel free to let us know about any mistake in the code or text by creating a new issue. We will get back to you and work on it.

Contribution

If you feel that there are more topics that could be added to this tutorial, feel free to create a new issue for it. I will try to respond as soon as possible, and if everything goes well, it will be updated in the tutorial. Thank you and happy learning! :)