Deep Learning代写 Computer Vision代写

ECS795P Deep Learning and Computer Vision, 2023

Deep Learning代写 How do D-loss and G-loss change during training? Visualize how the D-loss and   G-loss change during training and explain the reason.

Course Work 2 Guideline: Unsupervised Learning by Generative Adversarial Nets (GAN)

Introduction

Teamwork:

In groups of 2 students (grouped automatically by surname alphabetical order)

Aim:  Deep Learning代写

The assignment is to implement a deep learning model Generative Adversarial Nets (GAN) using PyTorch for unsupervised learning. The objectives are

(1) to obtain practical knowledge and hands-on understanding of the basic concepts in Generative Adversarial Nets(GAN);

(2) to obtain practical experience on how to implement basic GAN using PyTorch.

Start: Download and install PyTorch from its official website:

For Linux: https://pytorch.org/get-started/locally/#linux-installation .;

For Mac: https://pytorch.org/get-started/locally/#mac-installation ;

For Windows: https://pytorch.org/get-started/locally/#windows-installation .

Tasks: three subtasks are involved:   Deep Learning代写

1. Coding (group submission):to add your code blocks in the required sections based on Sec.2.2 in this guideline and annotations in coding files; (35% of this CW)
2. Report (group submission – up to 4 pages):to state the team assignment; to complete the questions in report; (50% of this CW)
3. Self reflection report (individual submission – up to 300 words): to describe the work that each member of the team did in relation to the questions and exercises in the group report and coding; to describe the challenge and your contribution to the team-work; to describe what you have learned from the team-work; (15% of this CW)

The group submission will be assessed based on contributions in a team collaboration (requires a brief declaration on who contributed what in each group), clarity, explanations, references, and coding. The individual submissions will be assessed for clarity and consistency of each member’s contribution to the team effort.

Platform: python +  PyTorch

Basic material:

Some of online materials for PyTorch-code may help you better complete this coursework (if you are not familiar with PyTorch, you can follow this step by step)

https://pytorch.org/tutorials/

https://github.com/yunjey/pytorch-tutorial

https://github.com/MorvanZhou/PyTorch-Tutorial/blob/master/tutorial-contents-notebooks/404_autoencoder.ipynb

1.Understanding basic concepts in GAN models  Deep Learning代写

Objective: To become familiar with the basic knowledge of the GAN model and its basic usages.

1.1 The questions to refresh (no need to write down in your report):

Reference: https://towardsdatascience.com/understanding-generative-adversarial-networks-4dafc963f2ef

Read the above blog and answer the below questions.:

1. What are the two basic parts in GAN?
2. What are the specific objectives of these two parts?
3. What is the basic loss function of GAN?
4. What is the training process of GAN?

2.Generative Adversarial Networks with PyTorch

Objective: To become familiar with GAN and re-implement the original GAN model.

2.1 The questions to refresh (no need to write down in your report):

How do D-loss and G-loss change during training? Visualize how the D-loss and   G-loss change during training and explain the reason.

2.2 The exercises to conduct with GAN_pytorch.py:

[coding] Modify the codes in GAN_pytorch.py to have the architecture of discriminator and generator as follow, then complete the rest part of the codes:

 

更多代写：北美exam代考  gre在家考vpn  北美代写作业  海外留学Essay代写  北美dissertation代写  Stochastic Modeling代考

合作平台：essay代写 论文代写 写手招聘 英国留学生代写