Expediting Deep Learning with Transfer Learning: PyTorch Playbook

This course covers the important design choices that a data professional must make while leveraging pre-trained models using Transfer Learning. It also covers the implementation aspects of different Transfer Learning approaches in PyTorch.
Course info
Level
Advanced
Updated
Jul 31, 2019
Duration
1h 45m
Table of contents
Description
Course info
Level
Advanced
Updated
Jul 31, 2019
Duration
1h 45m
Your 10-day individual free trial includes:

Expert-led courses

Keep up with the pace of change with thousands of expert-led, in-depth courses.
Description

Transfer learning refers to the re-use of a trained machine learning model for a similar problem, keeping the model architecture unchanged, but potentially altering the model’s weights.

In this course, Expediting Deep Learning with Transfer Learning: PyTorch Playbook, you will gain the ability to identify the right approach to transfer learning, and implement it using PyTorch.

First, you will learn how different forms of transfer learning - such as inductive, transductive, and unsupervised transfer learning - can be applied to different combinations of source and target domains. Next, you will discover how transfer learning solutions leverage the fact that lower layers of pre-trained models typically extract feature information and are data-specific, while later layers tend to be more problem-specific.

Finally, you will explore how to design and implement the correct strategy for freezing and fine-tuning the appropriate layers of your pre-trained model. You will round out the course by seeing how various powerful architectures are made available, in pre-trained form, in PyTorch’s suite of transfer learning solutions.

When you’re finished with this course, you will have the skills and knowledge to choose the right transfer learning approach to your specific problem, and design and implement it using PyTorch.

About the author
About the author

A problem solver at heart, Janani has a Masters degree from Stanford and worked for 7+ years at Google. She was one of the original engineers on Google Docs and holds 4 patents for its real-time collaborative editing framework.

More from the author
Summarizing Data and Deducing Probabilities
Intermediate
2h 50m
Jul 8, 2021
More courses by Janani Ravi
Section Introduction Transcripts
Section Introduction Transcripts

Course Overview
Hi. My name is Janani Ravi, and welcome to this course on Expediting Deep Learning with Transfer Learning. A little about myself, I have a masters degree in electrical engineering from Stanford and have worked at companies such as Microsoft, Google, and FlipKart. At Google, I was one of the first engineers working on realtime collaborative editing in Google Docs, and I hold four patents for its underlying technologies. I currently work on my own startup, Loonycorn, a studio for high-quality video content. Transfer learning refers to the reuse of a trained machine-learning model for a similar problem, keeping the model architecture unchanged, but potentially altering the model's weights. In this course, you will gain the ability to identify the right approach to transfer learning and implement it using PyTorch. First, you will learn how different forms of transfer learning, such as inductive, transductive, and unsupervised transfer learning can be applied to different combinations of source and target domains. Next, you will discover how transfer learning solutions leverage the fact that lower layers of pretrained models typically extract feature information and are data specific, while later layers tend to be more problem specific. Finally, you will explore how to design and implement the correct strategy for freezing and fine-tuning the appropriate layers of your pretrained model. You will round out the course by seeing the various powerful architectures that are made available in pretrained for in PyTorch's suite of suite of transfer learning solutions. When you're finished with this course, you will have the skills and knowledge to choose the right transfer learning approach to your specific problem, and design and implement it using PyTorch.