Pimlico modules

Pimlico modules for symbol embedding experiments. These are used in the Pimlico pipelines in the pipelines/ directory.

• Fake language tools
  • Corrupt text
  • Inspect corrupted text
• Input readers
  • Est Ref normalization
  • Europarl corpus reader
  • Ylilauta VRT files
• Symbol embedding methods
  • Corruption results
  • Learned embedding analysis
  • Embeddings from model
  • Language-specific embeddings
  • Neural sixgram (Xsym) trainer, v1
  • Neural sixgram (Xsym) trainer, v2
  • Neural sixgram samples prep
  • Plots of neural sixgram models
  • Store in TSV format
  • Validation criterion correlation