perf: always keep one midnight rose (OpenRouterTeam#83)

mojowebs · Mar 23, 2024 · af26dd5 · af26dd5
1 parent 936f731
commit af26dd5
Showing 1 changed file with 1 addition and 2 deletions.
diff --git a/modal/runner/containers/vllm_unified.py b/modal/runner/containers/vllm_unified.py
@@ -174,8 +174,7 @@ def __init__(self):
     model_name=_midnight_rose,
     gpu=modal.gpu.H100(count=1),
     concurrent_inputs=4,
-    # TODO: Keep it warm once launched
-    # keep_warm=1,
+    keep_warm=1,
     max_containers=1,
     quantization="GPTQ",
 )