feat(ocr): per-sender specialized Kurrent models with automatic active-learning retraining #263

Merged
marcel merged 32 commits from feat/issue-253-sender-models into main 2026-04-18 12:30:57 +02:00
Showing only changes of commit 57ffb7d751 - Show all commits

View File

@@ -97,7 +97,7 @@ services:
DETECTOR_BATCH_SIZE: "8"
OCR_CLAHE_CLIP_LIMIT: "2.0" # CLAHE contrast limit (multiplier of average histogram frequency)
OCR_CLAHE_TILE_SIZE: "8" # CLAHE tile grid size (NxN tiles per page)
OCR_MAX_CACHED_MODELS: "5" # LRU cache size for per-sender Kraken models
OCR_MAX_CACHED_MODELS: "2" # LRU cache; each model ~500 MB, so 2 = ~1 GB resident
networks:
- archive-net
healthcheck: