Deep Learning with Google Colab

Name: Deep Learning with Google Colab
Rating: 4.5 (119 reviews)

Implementing and training deep learning models in a free, integrated environment

Created byBPB Online + 100 Million Books Sold

Last updated 2/2020

English

What you'll learn

This course covers the general workflow of a deep learning project, implemented using PyTorch in Google Colab. At the end of the course, students will be proficient at using Google Colab as well as PyTorch in their own projects. Students will also learn about the theoretical foundations for various deep learning models and techniques, as well as how to implement them using PyTorch. Finally, the course ends by offering an overview on general deep learning and how to think about problems in the field; students will gain a high-level understanding of the role deep learning plays in the field of AI.
Learn how to utilize Google Colab as an online computing platform in deep learning projects, including running Python code, using a free GPU, and working with external files and folders
Understand the general workflow of a deep learning project
Examine the various APIs (datasets, modeling, training) PyTorch offers to facilitate deep learning
Learn about the theoretical basis for various deep learning models such as convolutional networks or residual networks and what problems they address
Gain an overview understanding of deep learning in the context of the artificial intelligence field and its best practices

Course content

8 sections • 61 lectures • 5h 42m total length

Introduction2:52
1. This lecture discusses the initial process of creating a Google account. As most Google applications are tied to this account, students will also have access to Google Drive, Gmail, etc.
- Know how to register for a Google account
- Know how to navigate to the Colab application
Registering for a Google account1:17
A hands-on tutorial on how to register for a Google account.
Navigating to Google Colab1:50
How to navigate to the Google Colab application within the Google workspace.
Exploring your Google Colab Notebook2:06
Exploration of various features in a Google Colab notebook.
The definition of notebooks1:07
Introduction to the concept of a computer notebook.
Running your first Google Colab code cell4:20
Executing Python code in a Colab notebook.
The markup language Markdown2:00
Introduction to the markup language Markdown.
Writing Markdown in Google Colab2:58
How to write Markdown code in a Colab notebook.
Writing LaTeX in Google Colab1:40
How to write LaTeX code in a Colab notebook.
Section conclusion1:01

Installing packages in Google Colab4:49
How to install external packages in a Colab notebook.
Working with files using Google Drive4:53
How to work with files via Google Drive in a Colab notebook.
Working with files directly in Google Colab5:22
How to work with files via Python code in a Colab notebook.
Sharing files via Google Drive4:04
How to share files with other users.
Introduction to version control with Git and GitHub4:15
Introduction to the concept of version control with Git and GitHub.
Sending Google Colab notebooks to GitHub7:00
How to facilitate version control with a Colab notebooks.

Creating a tensor7:48
How to create a tensor object in PyTorch.
Tensor operations6:44
How to apply operations on tensor objects in PyTorch.
GPUs in the context of deep learning5:30
Introduction to Graphical Processing Units and why they can be used in deep learning.
Turning on your Colab GPU4:55
How to utilize the free GPU provided in each Colab notebook.
Limits of the Colab GPU3:22
Various limitations that the free GPU has.
Neural network basics3:30
Introduction to neural networks.
Gradients and backpropagation8:19
Gradients and how neural networks learn.
Automatic differentiation in PyTorch7:08
How to facilitate automatic differentiation in PyTorch.
Training a model8:23
How to train a PyTorch model from scratch.
Saving and loading models7:43
How to save a trained model and load it back in a Python program.
Problem statement and setup3:31
Introduction to a sample curve-fitting problem.
Approaches and solutions7:17
Discussion of potential solutions to the curve-fitting problem.

Downloading a built-in dataset5:41
How to download a built-in dataset with PyTorch.
Working with PyTorch datasets7:04
Introduction to API provided by PyTorch datasets.
Loading a dataset into Colab4:42
The procedure of loading a custom dataset into Google Colab.
Building a PyTorch dataset9:03
The procedure of building a custom PyTorch dataset.
Image augmentation fundamentals7:10
Introduction to image augmentation methods.
Image augmentation in PyTorch9:42
How to utilize PyTorch’s API to facilitate image augmentation.

Downloading the dataset4:07
How to import in the dataset used for this project.
Understanding the dataset5:38
Exploration of various characteristics of the dataset.
Implementing a starting solution8:12
The design of a starting neural network to recognize handwritten digits.
Training and evaluating5:00
The process of training and evaluating deep learning models.
Choosing the size of input and output layers4:54
The intuition behind choosing the size of input and output layers of a neural network.
Choosing the size of hidden layers9:40
The intuition behind choosing the size of hidden layers of a neural network.
Loss functions6:32
Discussion regarding and comparisons between loss functions.
Activation functions and weight initialization7:47
Discussion regarding activation functions and weight initialization to avoid the vanishing gradient problem.
Optimizers8:14
Discussion regarding and comparisons between optimizers.

Downloading the dataset6:21
How to import in the dataset used for this project.
Understanding the dataset6:40
Exploration of various characteristics of the dataset.
What is transfer learning?8:20
Explanation of the practice of transfer learning.
The transfer learning workflow6:57
Detailed procedure of a transfer learning workflow.
Training and evaluating6:25
The process of training and evaluating deep learning models.
Pretrained models for transfer learning6:13
Discussion regarding and comparisons between pretrained models provided in PyTorch.

Downloading the dataset5:16
How to import in the dataset used for this project.
Understanding the dataset5:30
Exploration of various characteristics of the dataset.
Convolutional network fundamentals8:41
Introduction of convolutional neural networks and the problems they try to address.
Implementation in PyTorch7:26
The procedure of implementing a convolutional neural network in PyTorch.
Residual network fundamentals5:38
Introduction of residual neural networks and the problems they try to address.
Residual blocks in convolutional networks5:09
Discussions regarding and comparisons between different residual blocks.
Implementation in PyTorch5:28
The procedure of implementing a residual neural network in PyTorch.

General ensembling in machine learning6:33
Introduction to the concept of ensembling in general machine learning.
Ensembling in deep learning8:02
Unique methods of ensembling in deep learning with neural networks.
Data versioning5:20
Introduction to data version in deep learning and why it is important.
Reproducibility3:20
Introduction to reproducibility in deep learning and why it is important.
When not to use deep learning6:28
Discussions of various situations where deep learning is not desirable.

Requirements

Familiarity with Python programming (including classes, functions, context managers)
Basic linear algebra and calculus (briefly used during the discussions on various deep learning models and techniques)

Description

This course covers the general workflow of a deep learning project, implemented using PyTorch in Google Colab. At the end of the course, students will be proficient at using Google Colab as well as PyTorch in their own projects. Students will also learn about the theoretical foundations for various deep learning models and techniques, as well as how to implement them using PyTorch. Finally, the course ends by offering an overview on general deep learning and how to think about problems in the field; students will gain a high-level understanding of the role deep learning plays in the field of AI.

Learn how to utilize Google Colab as an online computing platform in deep learning projects, including running Python code, using a free GPU, and working with external files and folders
Understand the general workflow of a deep learning project
Examine the various APIs (datasets, modeling, training) PyTorch offers to facilitate deep learning
Learn about the theoretical basis for various deep learning models such as convolutional networks or residual networks and what problems they address
Gain an overview understanding of deep learning in the context of the artificial intelligence field and its best practices

Who this course is for:

AI enthusiasts interested in getting started on deep learning
Programmers familiar with deep learning looking to gain a comprehensive understanding of various deep learning models and techniques

Deep Learning with Google Colab

What you'll learn

Explore related topics

Course content

Getting started in Google Colab10 lectures • 21min

The ecosystem of Google Colab6 lectures • 30min

Introduction to PyTorch12 lectures • 1hr 14min

Working with datasets6 lectures • 43min

Recognizing handwritten digits9 lectures • 1hr

Transfer learning for object recognition6 lectures • 41min

Recognizing fashion items7 lectures • 43min

Deep learning best practices5 lectures • 30min

Requirements

Description

Who this course is for: