Implementation of Image Upscaling using TensorFlow
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
notebooks code update May 22, 2017
.gitignore ignored MNIST_data directory Mar 22, 2017
Dockerfile Dockerfile fix Apr 11, 2017
docker-compose.yml docker-compose uses now local Dockerfile Mar 18, 2017
nvidia-docker-compose.yml added nvidia docker compose Apr 11, 2017

Neural Image Upscaling

This project aims to implement Image Upscaling using Neural Networks with Python3 and TensorFlow

Currently WIP


Project is designed to run using Docker-Compose. First install it, and then type:

$ docker-compose up
Starting neuralimageupscaling_anakonda_1
Attaching to neuralimageupscaling_anakonda_1
anakonda_1  | Fetching package metadata .........
anakonda_1  | Solving package specifications: .
anakonda_1  | 
anakonda_1  | # All requested packages already installed.
anakonda_1  | # packages in environment at /opt/conda:
anakonda_1  | #
anakonda_1  | jupyter                   1.0.0                    py36_3  
anakonda_1  | [W 17:27:18.439 NotebookApp] WARNING: The notebook server is listening on all IP addresses and not using encryption. This is not recommended.
anakonda_1  | [I 17:27:18.444 NotebookApp] Serving notebooks from local directory: /opt/notebooks
anakonda_1  | [I 17:27:18.444 NotebookApp] 0 active kernels 
anakonda_1  | [I 17:27:18.444 NotebookApp] The Jupyter Notebook is running at: http://[all ip addresses on your system]:8888/?token=f147dd59eff55fafa7a693e52d35052b69e2603cde52202c
anakonda_1  | [I 17:27:18.444 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
anakonda_1  | [C 17:27:18.445 NotebookApp] 
anakonda_1  |     
anakonda_1  |     Copy/paste this URL into your browser when you connect for the first time,
anakonda_1  |     to login with a token:
anakonda_1  |         http://localhost:8888/?token=f147dd59eff55fafa7a693e52d35052b69e2603cde52202c

Now Jupyter notebook should be available under localhost:8888 address.

GPU support

You need to install nvidia-docker and nvidia-docker-compose first. Then, just run:

$ nvidia-docker-compose up