How to contribute
Help building the visionAI wiki, because you probably know it better anyway
We just are just getting started with the wiki and it's far from perfect. If you happen to know something about terms mentioned in the wiki and would like to contribute to this wiki, please be our guest! We appreciate every help and would send you a small thank you if you'd like to disclose your identity to us.
To contribute to the wiki, please reach out to [email protected] and we'll add you as an author to this Gitbook.
Some of the ideas we have to expand the wiki:
  • adding more frameworks to the code-examples
  • adding more sophisticated theoretical explanations
  • linking relevant papers and blog posts
  • adding more terms (see the list below)
But, we're always open to your ideas as well!

Terms we plan to include

Below, you find all the terms which we'd like to add to the wiki. The ones crossed out are the ones, which are already written. So, if you want to contribute just choose one which isn't crossed out yet. If you feel like something is missing, just add it as a pull request ๐ŸŒž

Model families

  • Classifier / Tagger
  • Semantic Segmentor
  • Object Detector
  • Instance Segmentor
  • Attributor

Model architectures

  • DeeplabV3+
  • ResNet
  • Faster R-CNN
  • Mask R-CNN


  • Confusion matrix
  • Accuracy
  • Hamming Score
  • IoU
  • mAP
  • Precision
  • Recall


  • Combined Segmentation Loss
  • Combined Detection Loss
  • Cross-Entropy Loss
  • Binary Cross-Entropy Loss

Solvers / optimizers

  • SGD
  • Adam
  • AdamW


  • Warm-up

Training parameters

  • Iterations
  • Epochs
  • Batch size
  • Evaluation period
  • Patience
  • Min delta
  • Seed