Skip to content
2D fully convolutional neural network for unsupervised super resolution
Branch: master
Clone or download
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
ZSSR_Images added Apr 3, 2019
.gitignore updated req main and gitignore Feb 24, 2019
.ml_recipe.yaml Update .ml_recipe.yaml Jun 4, 2019
.ml_tracking_repo first commit Dec 4, 2018
ZSSR_1.png Add files via upload Feb 28, 2019
main.py Update main.py May 11, 2019
readme.md fix: readme - change git clone instruction to use https Jun 10, 2019
requirements.txt Update requirements.txt May 11, 2019

readme.md

alt text

Zero Shot Super Resolution

This repository contains a Keras implementation of the ZSSR project. It produces state of the art (SOTA) single image super resolution (SISR) outputs for in-the-wild images with real world degragation and compression artifacts.

Getting Started

These instructions will get you a copy of the project up and running on your local machine and on MissingLink's cloud. for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisites

You need Python 3.6.8 on your system to run this example.

First, updating pip is advised:

pip install -U pip

Installing

Clone this repo:

git clone https://github.com/missinglinkai/ZSSR.git

Change directory:

cd ZSSR

You are strongly recommended to use virtualenv to create a sandboxed environment for individual Python projects:

pip install virtualenv

Create and activate the virtual environment inside the project directory:

virtualenv .venv
source .venv/bin/activate

Install dependency libraries:

pip install -r requirements.txt
  • To run local, uncomment "tensorflow" in the requiremnets.txt file before installing.

Run

Sign up to MissingLink and follow through the process to view the project on our UI.

Authenticate your username from the CLI:

ml auth init

Local

python main.py

Running config for example:

python main.py --epochs 2000 --subdir 001

MissingLink with Resource Management

Follow instructions here: Resource Management

After setting up Resource Management: Create a data volume through the UI and use its ID number to sync the local dataset:

ml data sync yourDataVolumeID --data-path ~/ZSSR_Images

Edit the ".ml_recipe.yaml" file and fill in your Data Volume ID, Data Volume Version and Organization Name:

command: 'python main.py'
data_volume: yourDataVolumeID
data_query: '@version:yourDataVolumeVersion @path:002/*'
gpu: true
org: 'yourOrganizationName'

Then simply run the project:

ml run xp

To change the input image (data query) open the ".ml_recipe.yaml" file and edit the DirName after path:

data_query: '@version:yourDataVolumeVersion @path:DirName/*'

Optional image directories: 001 002 003 004 005 007 016 019 020 032 034 039 047 052 067 071 098 100

You can’t perform that action at this time.