ai-ml/gke-ray/rayserve/llm/llama-3-8b/ray-service.yaml

# Copyright 2024 Google LLC
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

# [START gke_ai_ml_gke_ray_rayserve_llm_llama_3_8b]
apiVersion: ray.io/v1
kind: RayService
metadata:
  name: llama-3-8b
spec:
  serveConfigV2: |
    applications:
    - name: llm
      route_prefix: /
      import_path:  ai-ml.gke-ray.rayserve.llm.serve_chat_completion:model
      deployments:
      - name: VLLMDeployment
        num_replicas: 1
        ray_actor_options:
          num_cpus: 2
          # NOTE: num_gpus is set automatically based on TENSOR_PARALLELISM
      runtime_env:
        working_dir: "https://github.com/GoogleCloudPlatform/kubernetes-engine-samples/archive/main.zip"
        pip: ["vllm==0.4.3","transformers[torch]==4.40.2"]
        env_vars:
          MODEL_ID: "meta-llama/Meta-Llama-3-8B-Instruct"
          TENSOR_PARALLELISM: "2"
  rayClusterConfig:
    headGroupSpec:
      rayStartParams:
        dashboard-host: "0.0.0.0"
      template:
        metadata:
          labels:
            ai.gke.io: rayserve
        spec:
          containers:
            - name: ray-head
              image: rayproject/ray-ml:2.9.0-py310
              resources:
                limits:
                  cpu: "2"
                  memory: "8Gi"
                  ephemeral-storage: "10Gi"
                requests:
                  cpu: "2"
                  memory: "8Gi"
                  ephemeral-storage: "10Gi"
              ports:
                - containerPort: 6379
                  name: gcs-server
                - containerPort: 8265
                  name: dashboard
                - containerPort: 10001
                  name: client
                - containerPort: 8000
                  name: serve
              env:
                - name: HUGGING_FACE_HUB_TOKEN
                  valueFrom:
                    secretKeyRef:
                      name: hf-secret
                      key: hf_api_token
    workerGroupSpecs:
      - replicas: 1
        minReplicas: 0
        maxReplicas: 4
        groupName: gpu-group
        rayStartParams: {}
        template:
          metadata:
            labels:
              ai.gke.io: rayserve
          spec:
            containers:
              - name: llm
                image: rayproject/ray-ml:2.9.0-py310
                env:
                  - name: HUGGING_FACE_HUB_TOKEN
                    valueFrom:
                      secretKeyRef:
                        name: hf-secret
                        key: hf_api_token
                resources:
                  limits:
                    cpu: "8"
                    memory: "20Gi"
                    nvidia.com/gpu: "2"
                  requests:
                    cpu: "8"
                    memory: "20Gi"
                    nvidia.com/gpu: "2"
            nodeSelector:
              cloud.google.com/gke-accelerator: nvidia-l4
# [END gke_ai_ml_gke_ray_rayserve_llm_llama_3_8b]