Enable dedicate endpoint for TGI Gemma2 predict and chat completion

PiperOrigin-RevId: 684623757

notebooks/community/model_garden/model_garden_huggingface_tgi_deployment.ipynb

@@ -134,10 +134,7 @@
         "\n",
         "models, endpoints = {}, {}\n",
         "\n",
-        "HF_TOKEN = \"\"  # @param {type:\"string\", isTemplate: true}\n",
-        "\n",
-        "# Dedicated endpoint not supported yet\n",
-        "use_dedicated_endpoint = False"
+        "HF_TOKEN = \"\"  # @param {type:\"string\", isTemplate: true}\n"
       ]
     },
     {
@@ -159,6 +156,10 @@
         "# The pre-built serving docker image for TGI.\n",
         "TGI_DOCKER_URI = \"us-docker.pkg.dev/deeplearning-platform-release/gcr.io/huggingface-text-generation-inference-cu121.2-2.ubuntu2204.py310\"\n",
         "\n",
+        "# @markdown Set use_dedicated_endpoint to False if you don't want to use [dedicated endpoint](https://cloud.google.com/vertex-ai/docs/general/deployment#create-dedicated-endpoint).\n",
+        "use_dedicated_endpoint = True  # @param {type:\"boolean\"}\n",
+        "\n",
+        "\n",
         "machine_type = \"g2-standard-8\"  # @param {type: \"string\", isTemplate: true}\n",
         "accelerator_type = \"NVIDIA_L4\"  # @param [\"NVIDIA_L4\"] {isTemplate: true}\n",
         "accelerator_count = 1  # @param {type: \"integer\", isTemplate: true}\n",
@@ -350,6 +351,8 @@
         "# Send streaming chat completions request.\n",
         "endpoint = endpoints[\"tgi\"]\n",
         "url = f\"https://{endpoint.location}-aiplatform.googleapis.com/v1beta1/{endpoint.resource_name}/chat/completions\"\n",
+        "if use_dedicated_endpoint:\n",
+        "    url = f\"https://{endpoint.gca_resource.dedicated_endpoint_dns}/v1beta1/{endpoint.resource_name}/chat/completions\"\n",
         "\n",
         "payload = {\n",
         "    \"messages\": [{\"role\": \"user\", \"content\": prompt}],\n",