Fix llama notebooks

PiperOrigin-RevId: 761112151

notebooks/community/model_garden/model_garden_advanced_features.ipynb

@@ -6,7 +6,7 @@
       "id": "DZ1j6RRg-Td6",
       "metadata": {
         "cellView": "form",
-        "id": "DZ1j6RRg-Td6"
+        "id": "f705f4be70e9"
       },
       "outputs": [],
       "source": [
@@ -29,7 +29,7 @@
       "cell_type": "markdown",
       "id": "99c1c3fc2ca5",
       "metadata": {
-        "id": "99c1c3fc2ca5"
+        "id": "778cc1227be8"
       },
       "source": [
         "# Vertex AI Model Garden - Advanced Features\n",
@@ -52,7 +52,7 @@
       "cell_type": "markdown",
       "id": "f9-tJ6RfDLIs",
       "metadata": {
-        "id": "f9-tJ6RfDLIs"
+        "id": "0779b48f654e"
       },
       "source": [
         "## Overview\n",
@@ -90,7 +90,7 @@
       "cell_type": "markdown",
       "id": "47GcOrZjosOx",
       "metadata": {
-        "id": "47GcOrZjosOx"
+        "id": "69453bf7230e"
       },
       "source": [
         "## Before you begin"
@@ -100,7 +100,7 @@
       "cell_type": "markdown",
       "id": "1D_pWejJPHP3",
       "metadata": {
-        "id": "1D_pWejJPHP3"
+        "id": "bf3706e69f61"
       },
       "source": [
         "### Request for quota\n",
@@ -118,10 +118,9 @@
       "cell_type": "code",
       "execution_count": null,
       "id": "L3dqbxovo5t6",
-      "language": "python",
       "metadata": {
         "cellView": "form",
-        "id": "L3dqbxovo5t6"
+        "id": "2b585189a670"
       },
       "outputs": [],
       "source": [
@@ -149,6 +148,8 @@
         "# Install and import the necessary packages\n",
         "! pip install -q openai google-auth requests\n",
         "\n",
+        "! pip3 install --upgrade --quiet 'google-cloud-aiplatform>=1.93.1'\n",
+        "\n",
         "! git clone https://github.com/GoogleCloudPlatform/vertex-ai-samples.git\n",
         "\n",
         "import datetime\n",
@@ -224,7 +225,7 @@
       "cell_type": "markdown",
       "id": "SeGqxuMfRBS5",
       "metadata": {
-        "id": "SeGqxuMfRBS5"
+        "id": "4782dd003acb"
       },
       "source": [
         "### Access Llama 3.1, 3.2, and 3.3 models on Vertex AI for serving"
@@ -236,7 +237,7 @@
       "id": "BxlzWU2KQqmw",
       "metadata": {
         "cellView": "form",
-        "id": "BxlzWU2KQqmw"
+        "id": "798068fc0355"
       },
       "outputs": [],
       "source": [
@@ -275,7 +276,7 @@
       "cell_type": "markdown",
       "id": "JpNBJJgjWL7j",
       "metadata": {
-        "id": "JpNBJJgjWL7j"
+        "id": "10ed490e28e5"
       },
       "source": [
         "## Prefix Caching <a name=\"prefix-caching\"></a>\n",
@@ -303,7 +304,7 @@
       "cell_type": "markdown",
       "id": "9gZJ8cB27e1m",
       "metadata": {
-        "id": "9gZJ8cB27e1m"
+        "id": "30ddb93fdd7b"
       },
       "source": [
         "### Try out Prefix Caching with Hex-LLM\n",
@@ -319,7 +320,7 @@
       "id": "RpmoA2nXjdCd",
       "metadata": {
         "cellView": "form",
-        "id": "RpmoA2nXjdCd"
+        "id": "b56d82c1aa6f"
       },
       "outputs": [],
       "source": [
@@ -508,7 +509,7 @@
       "id": "5QoK8c0R9U3B",
       "metadata": {
         "cellView": "form",
-        "id": "5QoK8c0R9U3B"
+        "id": "96c5afed49b4"
       },
       "outputs": [],
       "source": [
@@ -574,7 +575,7 @@
       "id": "29rn5ATmB2YC",
       "metadata": {
         "cellView": "form",
-        "id": "29rn5ATmB2YC"
+        "id": "9a95c9f90358"
       },
       "outputs": [],
       "source": [
@@ -647,7 +648,7 @@
       "cell_type": "markdown",
       "id": "KjbM8E9DGuuR",
       "metadata": {
-        "id": "KjbM8E9DGuuR"
+        "id": "12ad6d1ff725"
       },
       "source": [
         "#### Delete the models and endpoints"
@@ -659,7 +660,7 @@
       "id": "JpLU7GRQGuuR",
       "metadata": {
         "cellView": "form",
-        "id": "JpLU7GRQGuuR"
+        "id": "1ab4e3bb74b4"
       },
       "outputs": [],
       "source": [
@@ -685,7 +686,7 @@
       "cell_type": "markdown",
       "id": "XZ33HhYmOxCS",
       "metadata": {
-        "id": "XZ33HhYmOxCS"
+        "id": "7a8a9a1b2ddf"
       },
       "source": [
         "### Try out Prefix Caching with vLLM\n",
@@ -708,7 +709,7 @@
       "id": "E8OiHHNNE_wj",
       "metadata": {
         "cellView": "form",
-        "id": "E8OiHHNNE_wj"
+        "id": "4425cc0bdedc"
       },
       "outputs": [],
       "source": [
@@ -912,7 +913,7 @@
       "id": "zex1oXl36A70",
       "metadata": {
         "cellView": "form",
-        "id": "zex1oXl36A70"
+        "id": "bcbafec839cd"
       },
       "outputs": [],
       "source": [
@@ -975,7 +976,7 @@
       "id": "gDOC_nfsJeUR",
       "metadata": {
         "cellView": "form",
-        "id": "gDOC_nfsJeUR"
+        "id": "e984f43422d5"
       },
       "outputs": [],
       "source": [
@@ -1048,7 +1049,7 @@
       "cell_type": "markdown",
       "id": "GdGxaTirJeUR",
       "metadata": {
-        "id": "GdGxaTirJeUR"
+        "id": "dff0d10dcc20"
       },
       "source": [
         "#### Delete the models and endpoints"
@@ -1060,7 +1061,7 @@
       "id": "OgoqXE-VJeUR",
       "metadata": {
         "cellView": "form",
-        "id": "OgoqXE-VJeUR"
+        "id": "5b8751773e7f"
       },
       "outputs": [],
       "source": [
@@ -1086,7 +1087,7 @@
       "cell_type": "markdown",
       "id": "w4Guijaw_NEs",
       "metadata": {
-        "id": "w4Guijaw_NEs"
+        "id": "863775857a46"
       },
       "source": [
         "### Best practices\n",
@@ -1101,7 +1102,7 @@
       "cell_type": "markdown",
       "id": "ml8fgoIQWSbY",
       "metadata": {
-        "id": "ml8fgoIQWSbY"
+        "id": "565cbdc3a06b"
       },
       "source": [
         "## Speculative Decoding <a name=\"spec-decoding\"></a>\n",
@@ -1143,7 +1144,7 @@
       "cell_type": "markdown",
       "id": "NmWRro8Q-Td6",
       "metadata": {
-        "id": "NmWRro8Q-Td6"
+        "id": "94eaa9050abb"
       },
       "source": [
         "### Try out Speculative Decoding with vLLM"
@@ -1155,7 +1156,7 @@
       "id": "72d1GlrYifKU",
       "metadata": {
         "cellView": "form",
-        "id": "72d1GlrYifKU"
+        "id": "5f358cc230a6"
       },
       "outputs": [],
       "source": [
@@ -1472,7 +1473,7 @@
       "id": "CNiItf5hdVFU",
       "metadata": {
         "cellView": "form",
-        "id": "CNiItf5hdVFU"
+        "id": "be3170e0e05a"
       },
       "outputs": [],
       "source": [
@@ -1547,7 +1548,7 @@
       "cell_type": "markdown",
       "id": "WahYGAZyq6Gl",
       "metadata": {
-        "id": "WahYGAZyq6Gl"
+        "id": "30c5d2535df3"
       },
       "source": [
         "## Clean up resources"
@@ -1557,7 +1558,7 @@
       "cell_type": "markdown",
       "id": "bV5Yjkgav9BZ",
       "metadata": {
-        "id": "bV5Yjkgav9BZ"
+        "id": "63c10917ff95"
       },
       "source": [
         "### Delete the models and endpoints"
@@ -1569,7 +1570,7 @@
       "id": "qsks36cOH9rb",
       "metadata": {
         "cellView": "form",
-        "id": "qsks36cOH9rb"
+        "id": "92892e1b1730"
       },
       "outputs": [],
       "source": [

notebooks/community/model_garden/model_garden_axolotl_finetuning.ipynb

@@ -231,6 +231,8 @@
         "import yaml\n",
         "from google.cloud import aiplatform\n",
         "\n",
+        "! pip3 install --upgrade --quiet 'google-cloud-aiplatform>=1.93.1'\n",
+        "\n",
         "common_util = importlib.import_module(\n",
         "    \"vertex-ai-samples.community-content.vertex_model_garden.model_oss.notebook_util.common_util\"\n",
         ")\n",

notebooks/community/model_garden/model_garden_pytorch_llama3_1_deployment.ipynb

@@ -1072,7 +1072,7 @@
         "            vllm_args.append(\"--tool-call-parser=vertex-llama-3\")\n",
         "        else:\n",
         "            vllm_args.append(\"--enable-auto-tool-choice\")\n",
-        "            vllm_args.append(\"--llama3_json\")\n",
+        "            vllm_args.append(\"--tool-call-parser=llama3_json\")\n",
         "\n",
         "    env_vars = {\n",
         "        \"MODEL_ID\": base_model_id,\n",

notebooks/community/model_garden/model_garden_pytorch_llama3_2_deployment.ipynb

@@ -1052,7 +1052,7 @@
         "            vllm_args.append(\"--tool-call-parser=vertex-llama-3\")\n",
         "        else:\n",
         "            vllm_args.append(\"--enable-auto-tool-choice\")\n",
-        "            vllm_args.append(\"--llama3_json\")\n",
+        "            vllm_args.append(\"--tool-call-parser=llama3_json\")\n",
         "\n",
         "    env_vars = {\n",
         "        \"MODEL_ID\": base_model_id,\n",

notebooks/community/model_garden/model_garden_pytorch_llama3_3_deployment.ipynb

@@ -406,7 +406,7 @@
         "            vllm_args.append(\"--tool-call-parser=vertex-llama-3\")\n",
         "        else:\n",
         "            vllm_args.append(\"--enable-auto-tool-choice\")\n",
-        "            vllm_args.append(\"--llama3_json\")\n",
+        "            vllm_args.append(\"--tool-call-parser=llama3_json\")\n",
         "\n",
         "    env_vars = {\n",
         "        \"MODEL_ID\": base_model_id,\n",
@@ -697,14 +697,10 @@
         "# @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). Note that [dedicated endpoint does not support VPC Service Controls](https://cloud.google.com/vertex-ai/docs/predictions/choose-endpoint-type), uncheck the box if you are using VPC-SC.\n",
         "use_dedicated_endpoint = True  # @param {type:\"boolean\"}\n",
         "\n",
-        "# @markdown Find Vertex AI prediction supported accelerators and regions at https://cloud.google.com/vertex-ai/docs/predictions/configure-compute.\n",
-        "accelerator_type = \"NVIDIA_H100_80GB\"  # @param [\"NVIDIA_H100_80GB\"] {isTemplate:true}\n",
+        "accelerator_type = \"NVIDIA_H100_80GB\"\n",
         "accelerator_count = 8\n",
-        "if accelerator_type == \"NVIDIA_H100_80GB\":\n",
-        "    machine_type = \"a3-highgpu-8g\"\n",
-        "    multihost_gpu_node_count = 1\n",
-        "else:\n",
-        "    raise ValueError(\"Only NVIDIA_H100_80GB is supported.\")\n",
+        "machine_type = \"a3-highgpu-8g\"\n",
+        "multihost_gpu_node_count = 1\n",
         "\n",
         "common_util.check_quota(\n",
         "    project_id=PROJECT_ID,\n",

notebooks/community/model_garden/model_garden_pytorch_llama4_deployment.ipynb

@@ -380,7 +380,7 @@
         "            vllm_args.append(\"--tool-call-parser=vertex-llama-3\")\n",
         "        else:\n",
         "            vllm_args.append(\"--enable-auto-tool-choice\")\n",
-        "            vllm_args.append(\"--llama3_json\")\n",
+        "            vllm_args.append(\"--tool-call-parser=llama3_json\")\n",
         "\n",
         "    env_vars = {\n",
         "        \"MODEL_ID\": base_model_id,\n",
@@ -891,7 +891,7 @@
         "\n",
         "\n",
         "models[\"hexllm_tpu\"], endpoints[\"hexllm_tpu\"] = deploy_model_hexllm(\n",
-        "    model_name=common_util.get_job_name_with_datetime(prefix=MODEL_ID),\n",
+        "    model_name=common_util.get_job_name_with_datetime(prefix=\"llama4-serve-hexllm\"),\n",
         "    model_id=model_id,\n",
         "    publisher=\"meta\",\n",
         "    publisher_model_id=\"llama4\",\n",