Remove memory resources constraint in A4X serving-launcher.yaml and change default number of GPUs in A4 TRTLLM inference recipe

Author: hmhv1222

inference/a4/single-host-serving/tensorrt-llm/values.yaml

@@ -37,7 +37,7 @@ service:
 workload:
   model:
     name:
-  gpus: 4
+  gpus: 8
   image:
   framework:
   configFile: serving-args.yaml

src/helm-charts/a4x/inference-templates/deployment/templates/serving-launcher.yaml

@@ -278,7 +278,5 @@ spec:
           resources:
             requests:
               nvidia.com/gpu: {{ $gpusPerNode }}
-              memory: "800Gi"
             limits:
-              nvidia.com/gpu: {{ $gpusPerNode }}
-              memory: "1000Gi"
+              nvidia.com/gpu: {{ $gpusPerNode }}