Try to set --tensor-parallel-size based on GPU count

Author: JohnGarbutt

charts/azimuth-llm/templates/api/deployment.yml

@@ -47,6 +47,10 @@ spec:
           {{- if .Values.api.extraArgs -}}
           {{- .Values.api.extraArgs | toYaml | nindent 10 }}
           {{- end -}}
+          {{- if gt .Values.api.gpus 1 }}
+          - --tensor-parallel-size
+          - {{ .Values.api.gpus }}
+          {{- end }}
         {{- if .Values.huggingface.secretName -}}
         envFrom:
         - secretRef: