4 Commits

Author SHA1 Message Date
b379faebdb change scheduling strategy to hibrid priority queue with aging 2026-05-19 20:57:38 +02:00
13fb341354 feat: model unload 2026-05-18 00:34:27 +02:00
1dbb1c7f6f add per model limits 2026-05-17 23:07:21 +02:00
7344aa4ef4 first commit 2026-05-17 09:54:18 +02:00