3 Commits

Author	SHA1	Message	Date
Dylan De Faoite	4abbd0643e	perf: use gpu on topic AI & move Model Init into functions ... By passing model initialisation into the function itself, the model is unloaded from memory after completion which avoids OOM errors	2026-02-05 19:11:51 +00:00
Dylan De Faoite	ba1501e493	feat: add nlp topic processing	2026-02-05 17:23:25 +00:00
Dylan De Faoite	d4db7bec24	refactor: move NLP processing out of server file	2026-02-05 17:10:48 +00:00