1. 23 Sep, 2020 1 commit
  2. 21 Sep, 2020 5 commits
  3. 15 Sep, 2020 7 commits
  4. 14 Sep, 2020 3 commits
  5. 08 Sep, 2020 1 commit
  6. 07 Sep, 2020 2 commits
  7. 04 Sep, 2020 5 commits
  8. 03 Sep, 2020 2 commits
  9. 31 Aug, 2020 5 commits
  10. 30 Aug, 2020 2 commits
    • Patrick Scheibe's avatar
      Make multi-label training independent of NN Training. · 3c20131a
      Patrick Scheibe authored
      For testing, it's nicer when you don't need to train the single label NN before training multi-label. So setting trainFNN to False and enableMultiLabel to True will only train the multi labels.
      3c20131a
    • Patrick Scheibe's avatar
      Fix data types used for training · e98c2888
      Patrick Scheibe authored
      - all types used for training are now consistently specified in settings.py
      - Remove verbose>0 conditions for logging. We can influence the logging globally be specifying the log-level and the conditions in all places are no longer required
      - Add separate callback for NN training. Now it can have settings independent of the AC.
      - Change several prints with logging
      e98c2888
  11. 28 Aug, 2020 2 commits
  12. 27 Aug, 2020 5 commits