num parallel embed

2024-07-26 15:18:35 -07:00 · 2024-07-26 15:18:35 -07:00 · 2647a0e443
commit 2647a0e443
parent ec4c35fe99
1 changed files with 2 additions and 0 deletions
--- a/server/sched.go
+++ b/server/sched.go
@ -132,6 +132,8 @@ func (s *Scheduler) processPending(ctx context.Context) {
 			if len(pending.model.ProjectorPaths) > 0 && numParallel != 1 {
 				numParallel = 1
 				slog.Warn("multimodal models don't support parallel requests yet")
+			} else if strings.Contains(pending.model.Config.ModelFamily, "bert") {
+				numParallel = runtime.NumCPU()
 			}

 			for {