mirror of
https://github.com/hanzo-ml/kserve.git
synced 2026-10-30 20:53:49 +00:00
45 lines
1.1 KiB
YAML
45 lines
1.1 KiB
YAML
apiVersion: serving.kserve.io/v1alpha2
|
|
kind: LLMInferenceService
|
|
metadata:
|
|
name: facebook-opt-125m-single
|
|
spec:
|
|
model:
|
|
uri: "hf://facebook/opt-125m"
|
|
name: facebook/opt-125m
|
|
replicas: 1
|
|
router:
|
|
scheduler: {}
|
|
route: {}
|
|
gateway: {}
|
|
template:
|
|
containers:
|
|
- name: main
|
|
image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.19.0
|
|
securityContext:
|
|
runAsNonRoot: true
|
|
runAsUser: 1000
|
|
env:
|
|
- name: USER
|
|
value: nonroot
|
|
- name: VLLM_LOGGING_LEVEL
|
|
value: DEBUG
|
|
- name: VLLM_CPU_KVCACHE_SPACE
|
|
value: "1"
|
|
resources:
|
|
limits:
|
|
cpu: "2"
|
|
memory: 7Gi
|
|
requests:
|
|
cpu: 200m
|
|
memory: 2Gi
|
|
livenessProbe:
|
|
initialDelaySeconds: 180
|
|
periodSeconds: 30
|
|
timeoutSeconds: 30
|
|
failureThreshold: 8
|
|
readinessProbe:
|
|
initialDelaySeconds: 30
|
|
periodSeconds: 10
|
|
timeoutSeconds: 5
|
|
failureThreshold: 3
|