Horatiu George TodoranandCodex f418967d55 perf: RKNN half RAM model usage + no weight duplication in multi-core (#23958) ...

* perf(ml): share RKNN weights across NPU cores and halve model RAM

* perf(ml): default rknn threads to npu cores and cache shape setup

* perf(ml): serve rknn runs in arrival order

(cherry picked from commit e659fb62f2bb3f6b1db08ed643d3d054a539c415)

2026-10-28 20:40:53 +03:00

..

clip

fix(ml): stricter typing and request validation (#31856)

2026-09-28 14:14:01 -04:00

facial_recognition

fix(ml): stricter typing and request validation (#31856)

2026-09-28 14:14:01 -04:00

ocr

feat(ml): ppocr v6 (#32151)

2026-10-06 08:44:13 +00:00

__init__.py

fix(ml): stricter typing and request validation (#31856)

2026-09-28 14:14:01 -04:00

base.py

fix(ml): stricter typing and request validation (#31856)

2026-09-28 14:14:01 -04:00

cache.py

fix(ml): preloaded model loaded twice (#32610)

2026-10-22 05:23:22 +03:00

constants.py

perf: RKNN half RAM model usage + no weight duplication in multi-core (#23958)

2026-10-28 20:40:53 +03:00

transforms.py

feat(server): add jxl as target format (#32732)

2026-10-27 23:12:52 +03:00