gklezd

3 ms·

1. 1.▲
   by gklezd 2y ago

   That is a great point! I can certainly think of cases where you might want to go with an LLM instead and we have definitely experimented with that approach. Here are some reasons why we think TreeSeg is more suitable for us: 1. A more algor
2. 2.▲
   by gklezd 2y ago

   Very interesting indeed!
3. 3.▲
   by gklezd 2y ago

   Thanks for the reference! Note that we have open-sourced TreeSeg. You can find the repo here: https://github.com/AugmendTech/treeseg You can see its inner workings and use it directly or adapt it to your case. Excited
4. 4.▲
   by gklezd 2y ago

   Here is a link to the preprint for more details: https://www.arxiv.org/abs/2407.12028
5. 5.▲
   TreeSeg: Hierarchical Topic Segmentation of Large Transcripts(augmend.com)

   102 points by gklezd 2y ago|8 comments