Training on a new corpus

[loretoparisi]

Hello,
I did not find a notebook or training recipe to train wtpsplit on a new corpus.
In the issues there were some colab and links related to fine-tune LoRa adapters, but it seemes outdated and some related to nnsplit.

[bminixhofer]

Hi!

Does this do what you need? https://github.com/segment-any-text/wtpsplit?tab=readme-ov-file#adapt-to-your-own-corpus-via-lora