It's a very simple version for visualizing the training result produced by mmdetection.


The program supports drawing six training result and the most important evaluation tool:PR curve(only for VOC now)

  1. loss_rpn_bbox

  2. loss_rpn_cls

  3. loss_bbox

  4. loss_cls

  5. loss

  6. acc

  7. PR_curve

  8. F-measure


  1. Clone it
    git clone

There will be total 5 files(json directory, output directory,,,

  • put under /mmdetection/tools/

  • put under mmdetection/mmdet/core/evaluation/

How to use

six training result

  1. After training finished, you will have work_dir directory in your mmdetection directory
  2. take the latest json file and put into json directory in mmditection_visualize directory
  3. command python json/xxxxxxxlog.json in terminal
  4. check the output directory, Done !

PR curve and F-measure

  1. make sure and settled down
  2. command as usual like python tools/ {your pkl file} {your network configs file}
    • example python tools/ result.pkl ./configs/
  3. check the /mmdetection main directory, you will see the PR_curve_each_class.png thereļ¼Œ Done !