Skip to content

ducha-aiki/wisv2019-competition

Repository files navigation

Starter pack for participating in CAIP 2019 local descriptor contest

This is python started pack depending on numpy and pytorch for particiation in CAIP 2019 contest

Official starter pack is in matlab, so I have prepared python version.

How to use it:

  1. Download the dataset and unpack, following the official instructions

  2. Put this repo into "code" directory

  3. Run

        python extract_hardnets.py
        python match_hardnets.py
  1. Now you have output_data folder, which you may zip and send to organizers.

  2. You might want to explore the data with EDA.ipynb script and visually check a validity of your submission via sanity_check_submission.ipynb

Releases

No releases published

Packages

No packages published