Gallery of short examples

This example gallery contains small code-snippets you can use as a starting point for specific tasks. They are not meant as complete analyses, and many examples are based on synthetic toy examples. For a more in-depth example, see Worked example: Inspecting an HTR model.

• Computing the CER and WER
• Toy emoji OCR example
• Overview of tokenizations
• Resolving confusables and ligatures with custom lists
• Custom tokenization with Norwegian stemmer
• Evaluating token-specific performance
• Exploring common errors in a transcription
• Uncertainty metrics for token-specific statistics
• Visualizing alignments
• Customise the alignment visualisation with CSS
• Reading PAGE- and ALTO-XML