Introduction

this is a project that I have developed for my final project at high school.
I have strated from scracth with no knowlage about machine learning and moved my way up during research and development.
I am setisfied with my final result but I will still try to improve things.

Usage

Neural network usage

in order to use the gans network:

python main.py <mode> <batch_size>

there are 4 avilable modes:

• train mode: used for training the model
• test mode: used for testing the model
• eval mode: used for colorizing single or few images
• standby mode: a mode where the script waits for files from the stdin, colorizes them, rinse and repeat.
  extra usage

python main.py <test dataset> <train dataset>
<epochs>
<learning rate> <visdom ip and port> <decy lr> <checkpoint> <save weights> <save location> <beat0>  <eval images>

Server usage

the server is a simple server built with flask. Its purpose is to make the network accessible and easy to use.

python server.py <host> <port> <upload loc> <save loc> <checkpoint>

Built with

PyTorch - a machine learning framework
Flask - open source web application framework
torchvision - a package that consists of popular datasets, model architectures, and common image transformations for computer vision.

Exampales

Credits

Thanks to @Gangana3 for helping me with the amazing designs for this projects.

I also want to thank my mentor Shai who helped me through the whole project.

During my project I have taken insperation from a number of projects and I would like to mention them here:

https://arxiv.org/pdf/1803.05400.pdf
https://github.com/jantic/DeOldify