Coursera – Generative Adversarial Networks (GANs) Specialization 2021-2

Description

Generative Adversarial Networks (GANs) Specialization is a training course on Enemy Generating Networks (GANs). GANs are powerful machine learning models that are capable of producing realistic images, videos, and sounds. Although these models originated from game theory, today they have a wide range of applications, from improving cyber security and anonymizing data for privacy to producing artistic images, coloring black and white images, increasing resolution, creating avatars. , Convert 2D photos to 3D and many more. This course introduces you to the production of images by GANs and enhances your level of knowledge from basic concepts to advanced techniques, and for software engineers, students and researchers in any field interested in machine learning and understanding of GAN functionality. Are suitable.

This course is divided into 3 sub-courses. In the first part, you will understand the basics of GANs, build a simple GAN using the PyTorch module, use torsion layers to build advanced DCGANs capable of processing images and applying the W-Loss function, and Learn how to build conditional GANs. The second part is dedicated to the challenges of evaluating GANs, during which you will learn how to compare different GAN models, use the FID method to evaluate the accuracy and diversity of the model, identify biased sources, and implement various techniques related to StyleGAN. The last section is devoted to the practical use of GANs in enhancing data, privacy, building Pix2Pix and CycleGAN for translating images and other applications.

What are you learning?

Understand the components of GANs, build simple GANs with advanced PyTorch and DCGANs

Comparison of manufacturer models, use of Fréchet Inception Distance - FID method, bias detection and implementation of StyleGAN techniques

Use GANs to enhance data, privacy, mapping applications, and test and build Pix2Pix and CycleGAN for image translation

What skills do you acquire?

Generative Adversarial Networks (GANs)

Photo to photo generator and translator

Controlled and conditioned production

WGANs, DCGANs and StyleGANs

Bias in GANs

And …

Specifications Generative Adversarial Networks (GANs) Specialization

Publisher: Coursera

Instructors: Sharon Zhou, Eda Zhou, Eric Zelikman

English language

Level of education: Intermediate

Number: 3 courses

Course duration: with a suggested time of 9 hours per week, approximately 3 months

Prerequisites

Learners should have a working knowledge of AI, deep learning, and convolutional neural networks. They should have intermediate Python skills as well as some experience with any deep learning framework (TensorFlow, Keras, or PyTorch). Learners should be proficient in basic calculus, linear algebra, and statistics.

We highly recommend that you complete the Deep Learning Specialization prior to starting the GANs Specialization.

Installation

After Extract, watch with your favorite Player.

English subtitle

Quality: 720p