Course duration: **Nov 21, 2020 - Jan 2, 2021**

"Deep Learning with PyTorch: Zero to GANs" is a beginner-friendly online course offering a practical and coding-focused introduction to deep learning using the PyTorch framework. Enroll now to start learning.

* Watch live hands-on tutorials on YouTube
* Train models on cloud Jupyter notebooks
* Build an end-to-end real-world course project
* Earn a verified certificate of accomplishment

Add the lecture schedule to your calendar ([English](https://calendar.google.com/calendar/u/2?cid=Y19jOWYxdHZlMHM3aGppNnY3YXZ2bDcxNmxxY0Bncm91cC5jYWxlbmRhci5nb29nbGUuY29t), [Hindi](https://calendar.google.com/calendar/u/2?cid=Y184bWk0MnRkdGVpNnVoNnU5bHBpbjJjYjVxZ0Bncm91cC5jYWxlbmRhci5nb29nbGUuY29t)) and browse the [Course Community Forum](https://jovian.ai/forum/c/pytorch-zero-to-gans/18).

Deep Learning with PyTorch: Zero to GANs

**Submission Deadline**: Dec 5, 2020, 11:59 PM GMT

The objective of this assignment is to develop a solid understanding of PyTorch tensors. In this assignment you will:
* Pick 5 interesting functions related to PyTorch tensors by [reading the documentation](https://pytorch.org/docs/stable/torch.html)
* Create a Jupyter notebook using a starter template to illustrate their usage
* Upload and showcase your Jupyter notebook on your Jovian profile
* (optional) Write a blog post to accompany and showcase your Jupyter notebook 
* Share your work with the community and exchange feedback with other participants

* Explore the PyTorch documentation website
* Demonstrate usage of some tensor operations
* Publish your Jupyter notebook & share your work

Notebook Link

Assignment 1 - All About torch.Tensor

**Submission Deadline**: Dec 12, 2020, 11:59 PM GMT

In this assignment, you’re going to use information like a person’s age, sex, BMI, no. of children, and smoking habit to predict the price of yearly medical bills. You will train a model with the following steps:

* Download and explore the dataset
* Prepare the dataset for training
* Create a linear regression model
* Train the model to fit the data
* Make predictions using the trained model

* Download and explore a real-world dataset
* Create a linear regression model using PyTorch
* Train multiple models and make predictions

Assignment 2 - Train Your First Model

The ability to try many different neural network architectures to address a problem is what makes deep learning really powerful, especially compared to shallow learning techniques like linear regression, logistic regression etc. In this assignment, you will:
* Explore the [CIFAR10 dataset](https://www.cs.toronto.edu/~kriz/cifar.html)
* Set up a training pipeline to train a neural network on a GPU
* Experiment with different network architectures & hyperparameters

* Explore the CIFAR10 image dataset
* Create a pipeline for training on GPUs
* Hyperparameter tuning & optimization

Assignment 3 - Feed Forward Neural Networks

For the course project, students will create an image classification model using Convolutional neural networks, on a real-world dataset of their choice. The project will allow students to experiment with different types of models and regularization techniques. Students will also present their work at the end of the course and publish a blog post describing their approach and results.

* Discover & explore a large real-world dataset
* Train a convolutional neural network from scratch
* Document, present, and publish your work online

Blog Link

Description

Course Project

Hindi version: https://youtu.be/7yRkvPxE7Hs

We start this lesson with the basics of PyTorch: tensors, gradients, and Autograd. We then proceed to implement linear regression and gradient descent from scratch, first using matrix operations and then using PyTorch built-ins. Notebooks used in this lesson:

* PyTorch Basics: [jovian.ai/aakashns/01-pytorch-basics](https://jovian.ai/aakashns/01-pytorch-basics)
* Linear Regression: [jovian.ai/aakashns/02-linear-regression](https://jovian.ai/aakashns/02-linear-regression)
* Machine Learning: [jovian.ai/aakashns/machine-learning-intro](https://jovian.ai/aakashns/machine-learning-intro)

Ask questions on the [course discussion forum](https://jovian.ai/forum/c/pytorch-zero-to-gans/lecture-1-pytorch-basics-linear-regression/63).

* PyTorch basics: tensors, gradients, and autograd
* Linear regression & gradient descent from scratch
* Using PyTorch modules: nn.Linear & nn.functional

Lesson 1 - PyTorch Basics and Gradient Descent

Hindi Version: https://youtu.be/qmzgJvVP9XA

This lesson covers how to work with images from the MNIST dataset, creating training & validation datasets to avoid overfitting, and training a logistic regression model using the softmax activation & the cross-entropy loss. Notebooks used in this lesson: 

* Logistic regression: https://jovian.ai/aakashns/03-logistic-regression
* Logistic regression starter: https://jovian.ai/aakashns/mnist-logistic-minimal
* Linear regression starter: https://jovian.ai/aakashns/housing-linear-minimal

Ask questions on the [community discussion forum](https://jovian.ai/forum/t/lecture-2-working-with-images-logistic-regression/13735).

* Training-validation split on the MNIST dataset
* Logistic regression, softmax & cross-entropy
* Model training, evaluation & sample predictions

Lesson 2 - Working with Images and Logistic Regression

* Working with cloud GPU platforms like Kaggle & Colab
* Creating a multilayer neural network using nn.Module
* Activation function, non-linearity and universal approximation theorem
* Moving with datasets and models to the GPU for faster training

* Multilayer neural networks using nn.Module
* Activation functions, non-linearity & backprop
* Training models faster using cloud GPUs

Lesson 3: Training Deep Neural Networks on a GPU

* Working with the 3-channel RGB images from the CIFAR10 dataset
* Introduction to Convolutions, kernels & features maps
* Underfitting, overfitting and techniques to improve model performance

* Working with 3-channel RGB images 
* Convolutions, kernels & features maps
* Training curve, underfitting & overfitting

Lesson 4: Image Classification with Convolutional Neural Networks

* Improving the dataset using data normalization and data augmentation
* Improving the model using residual connections and batch normalization
* Improving the training loop using learning rate annealing, weight decay and gradient clip
* Training a state of the art image classifier from scratch in 10 minutes

* Adding residual layers with batchnorm to CNNs
* Learning rate annealing, weight decay & more
* Training a state-of-the-art model in 5 minutes

Lesson 5: Data Augmentation, Regularization, and ResNets

* Introduction to generative modeling and application of GANs
* Creating generator and discriminator neural networks
* Generating and evaluating fake images of handwritten digits
* Training the generator and discriminator in tandem and visualizing results

* Generative modeling and applications of GANs
* Training generator and discriminator networks
* Generating fake digits & anime faces with GANs

Lesson 6: Image Generation using Generative Adversarial Networks (GANs)

Aakash is the co-founder and CEO of [Jovian.ml](https://jovian.ml), a project management and collaboration platform for machine learning. Prior to starting [Jovian.ml](https://jovian.ml), Aakash worked as a software engineer (APIs & Data Platforms) at Twitter in Ireland & San Francisco and graduated from Indian Institute of Technology, Bombay. He’s also an avid blogger, open source contributor and online educator. 
https://www.linkedin.com/in/aakashns/

Instructor - Aakash N S