ai-ml/gke-ray/rayserve/ap_falcon-40b.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_ap_falcon_40b]
apiVersion: ray.io/v1
kind: RayService
metadata:
  name: rayllm
spec:
  serviceUnhealthySecondThreshold: 2400 # Config for the health check threshold for service. Default value is 60.
  deploymentUnhealthySecondThreshold: 2400 # Config for the health check threshold for deployments. Default value is 60.
  # setting the MODEL_ID env var on the workerGroupSpec did not work, it couldn't find it during exec
  serveConfigV2: |
    applications:
    - name: falcon
      route_prefix: /
      import_path: model_nf4_mg:chat_app_nf4_mg
      runtime_env:
        env_vars:
          MODEL_ID: "tiiuae/falcon-40b-instruct"
          PYTHONPATH: "${HOME}/models"
        pip: 
        - bitsandbytes==0.42.0          
      deployments:
      - name: Chat
        num_replicas: 1
  rayClusterConfig:
    # Ray head pod template
    headGroupSpec:
      # The `rayStartParams` are used to configure the `ray start` command.
      # See https://github.com/ray-project/kuberay/blob/master/docs/guidance/rayStartParams.md for the default settings of `rayStartParams` in KubeRay.
      # See https://docs.ray.io/en/latest/cluster/cli.html#ray-start for all available options in `rayStartParams`.
      rayStartParams:
        resources: '"{\"accelerator_type_cpu\": 2}"'
        dashboard-host: '0.0.0.0'
        block: 'true'
      #pod template
      template:
        spec:
          containers:
          - name: ray-head
            image: anyscale/ray-llm:0.5.0
            resources:
              limits:
                cpu: "2"
                memory: "8Gi"
              requests:
                cpu: "2"
                memory: "8Gi"
            volumeMounts:
            - mountPath: /home/ray/models
              name: model                
            ports:
            - containerPort: 6379
              name: gcs-server
            - containerPort: 8265 # Ray dashboard
              name: dashboard
            - containerPort: 10001
              name: client
            - containerPort: 8000
              name: serve
          volumes:
          - name: model
            configMap:
              name: quantized-model
              items:
              - key: model_nf4_mg.py
                path: model_nf4_mg.py
    workerGroupSpecs:
    # the pod replicas in this group typed worker
    - replicas: 1
      minReplicas: 0
      maxReplicas: 1
      # logical group name, for this called small-group, also can be functional
      groupName: gpu-group
      rayStartParams:
        block: 'true'
        resources: '"{\"accelerator_type_cpu\": 22, \"accelerator_type_l4\": 2}"'
      # pod template
      template:
        spec:
          securityContext:
            runAsUser: 1000
            fsGroup: 100        
          containers:
          - name: llm
            image: anyscale/ray-llm:0.5.0
            env:
            - name: HUGGING_FACE_HUB_TOKEN
              valueFrom:
                secretKeyRef:
                  name: hf-secret
                  key: hf_api_token
            lifecycle:
              preStop:
                exec:
                  command: ["/bin/sh","-c","ray stop"]
            resources:
              limits:
                cpu: "22"
                memory: "40Gi"
                nvidia.com/gpu: "2"
              requests:
                cpu: "22"
                memory: "40Gi"
                nvidia.com/gpu: "2"
            volumeMounts:
            - mountPath: "/home/ray/.cache"
              name: worker-disk-vol
          # Please ensure the following taint has been applied to the GPU node in the cluster.
          tolerations:
            - key: "ray.io/node-type"
              operator: "Equal"
              value: "worker"
              effect: "NoSchedule"
          nodeSelector:
            cloud.google.com/gke-accelerator: nvidia-l4
          volumes:
          - name: worker-disk-vol
            persistentVolumeClaim:
              claimName: worker-disk
# [END gke_ai_ml_gke_ray_rayserve_ap_falcon_40b]