segmentation # Knowledge liquefaction by [BERT image

  • Default “copy” is fine.
    • see copy mechanism
    • In that case, the circuit would be chopped at the appropriate length with a token string
    • You can give them a combined bulleted section from Scrapbox, excluding code regions, etc., and train them to split it correctly.
  • Delete unneeded tokens
    • There is no suitable data set here.

This page is auto-translated from /nishio/BERTによる分節化 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.