mirror of
https://github.com/hanzo-ml/kserve.git
synced 2026-10-30 20:53:49 +00:00
51 lines
1.1 KiB
YAML
51 lines
1.1 KiB
YAML
apiVersion: serving.kserve.io/v1alpha1
|
|
kind: ClusterServingRuntime
|
|
metadata:
|
|
name: kserve-predictiveserver
|
|
annotations:
|
|
serving.kserve.io/server-type: predictiveserver
|
|
spec:
|
|
multiModel: false
|
|
annotations:
|
|
prometheus.kserve.io/port: '8080'
|
|
prometheus.kserve.io/path: "/metrics"
|
|
supportedModelFormats:
|
|
- name: sklearn
|
|
version: "1"
|
|
autoSelect: false
|
|
priority: 3
|
|
- name: xgboost
|
|
version: "2"
|
|
autoSelect: false
|
|
priority: 3
|
|
- name: lightgbm
|
|
version: "4"
|
|
autoSelect: false
|
|
priority: 3
|
|
protocolVersions:
|
|
- v1
|
|
- v2
|
|
containers:
|
|
- name: kserve-container
|
|
image: kserve-predictiveserver:replace
|
|
args:
|
|
- --model_name={{.Name}}
|
|
- --model_dir=/mnt/models
|
|
- --http_port=8080
|
|
- --framework={{.Annotations.modelFormat}}
|
|
- --nthread=1
|
|
securityContext:
|
|
allowPrivilegeEscalation: false
|
|
privileged: false
|
|
runAsNonRoot: true
|
|
capabilities:
|
|
drop:
|
|
- ALL
|
|
resources:
|
|
requests:
|
|
cpu: "1"
|
|
memory: 2Gi
|
|
limits:
|
|
cpu: "1"
|
|
memory: 2Gi
|