sukanyasaha007 / MS-Independent-study

This contains learning outcome on generative advarsarial networks

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Independent Study Learning progress

In this repo I will be tracking my progress on independent study. I will start with Coursera and open ai materials on GANs and RL, google scholar papers, paper on Learning by Watching: Physical Imitation, Robot Learns To Imitate Human Demonstrations Using GAN Networks by AVID etc.

Resources -

GAN -

Coursera GAN Speciallization,

GANs in Action by Jakub Langer,

RL -

Coursera RL Speciallization,

Deep Learning by Ian Goodfellow and Yoshua Bengio

Independent Study Goal

Weekly IS journal (Github Readme, Google Slides) and document of learning progress

Mini projects on GANs

Read one-two related papers a week

Brainstorm ideas for Spring 2022 IS project

Checklist for the semester

  • Weekly report on learning outcome
  • Final github repo of all the assignments and projects done during the semester
  • Finalize plan for Spring 2022 IS project, maintain github repo for the same

Weekly Progress

Week-1 - Basic Introduction on How GAN works

Week-2 - Batch Normalization, Pooling, Upsampling, DCGAN

Week-3 - Mode Collapse, Binary crossentropy, WGAN

Week-4 - Conditional GANs

Week-5 - Feature extraction and Evaluation

Week-6 - GAN Disadvantages and Bias

Week-7 - StyleGAN and Advancements

Week-8 - GANs for Data Augmentation and Privacy

Week-9 - Image-to-Image Translation with Pix2Pix

Week-10 - Unpaired Translation with CycleGAN

Week-11-14 - Project on Virtual Stylist

Final Project Report

Virtual Stylist.pdf

Paper Readings

Paper 1 - Imagenet

Paper 2 - An Introduction to Variational Autoencoders

Paper 3 - StyleGAN

Paper 4 - Pix2Pix and CycleGAN

Paper 5 - text-2-image

Paper 6 - DCGAN

Paper 7 - TryOnGAN

Paper 8 - ShineON

Paper 9 - Wasserstein Loss

Paper 9 - Self-Attention GAN

Paper 10 - Are GANs Created Equal

Paper 11 - Buy Me That Look

About

This contains learning outcome on generative advarsarial networks


Languages

Language:Jupyter Notebook 99.7%Language:Python 0.3%Language:Cuda 0.0%Language:C++ 0.0%Language:Shell 0.0%Language:Dockerfile 0.0%