Lets keep those models loaded for a bit longer.

2024-07-17 14:10:25 -05:00 · 2024-07-17 14:10:25 -05:00 · 09642a8de9
commit 09642a8de9
parent ba9afaad72
1 changed files with 2 additions and 0 deletions
--- a/kubernetes/apps/ai/ollama/app/helmrelease.yaml
+++ b/kubernetes/apps/ai/ollama/app/helmrelease.yaml
@ -43,6 +43,8 @@ spec:
                value: "*"
              - name: OLLAMA_MODELS
                value: &modelPath "/models"
+              - name: OLLAMA_KEEP_ALIVE
+                value: "24h"
            resources:
              requests:
                nvidia.com/gpu: 1 # requesting 1 GPU