Skip to content

Print stats on the dataset before training #58

Description

@starride-teklia

It could be useful to have some stats on the dataset before training. Something like:

  • number of images (training/validation sets)
  • number of characters/words (training/validation sets)
  • character distribution (training set)
  • unknown characters

Metadata

Metadata

Labels

No labels
No labels

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions