Skip to content

Latest commit

 

History

History
50 lines (41 loc) · 1.92 KB

README.md

File metadata and controls

50 lines (41 loc) · 1.92 KB
Raw

Reproducibility

In this sub-directory you can find all scripts and jupyter notebook which we used to produce the results reported in the manuscript. If you want to run the exact code we used, you have to use release v0.1.7. You will see the name noise2seg, the original name of the project, in these files. We only renamed the project to DenoiSeg once we started writing it up :)

Note: These files are here for completeness and not expected to run out of the box on any given setup.

./cluster_execution

The scripts in this sub-directory were used to run all our experiments on the MPI-CBG cluster. All experiments were run via singularity. The singularity image was built with ./cluster_exectuion/singularity/noise2seg.Singularity.

Build Singularity Container

  1. Install Singularity
  2. sudo singularity build denoiseg.sim singularity/noise2seg.Singularity

Run Singularity Container

Now you can run a jupyter-server with this container: singularity run --nv -B singularity/user:/run/user -B ./:/notebooks singularity/noise2seg.simg

Run Singularity Headless

  1. Install PyInquirer
  2. python3 run_noise2seg.py

For the headless to work you have to set following parameters in the file run_noise2seg.py:

singularity_path = "/projects/Noise2Seg/singularity/"
base_dir = "/projects/Noise2Seg"
gitrepo_path = join(base_dir, 'Noise2Seg')
base_path_data = join(base_dir, "data")
base_path_exp = join(base_dir, "experiments")

./dataprep

This sub-directory contains all the jupyter notebooks which we used to prepare the training datasets.

./figures

This sub-directory contains all the jupyter notebooks which we used to create the figures. These notebooks use TeX to render the fonts. If you want to run them, make sure that you have TeX installed.

./tables

This sub-directory contains all the jupyter notebooks which we used to create the tables.