remove option

2026-05-11 11:35:29 -05:00
parent 08106b6693
commit ce65c6435b
1 changed files with 0 additions and 1 deletions
--- a/clusters/k3s-dgx/nim-service/qwen.yaml
+++ b/clusters/k3s-dgx/nim-service/qwen.yaml
@@ -68,7 +68,6 @@ spec:
    - "float"                # Explicit FP8 encoding
    - --quantization
    - "fbgemm-fp8"            # Explicit quantization backend
-    - --sv2-transformer-bindings  # Skip SA checks (small speedup)
  authSecret: hf-api-secret
  storage:
    sharedMemorySizeLimit: 64Gi