dip/service-catalog
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

kserve model image add

Browse Source
This commit is contained in:
ychangkim
2025-07-21 17:17:52 +09:00
parent 5f49acbb74
commit eb17388ee4
4 changed files with 6 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
charts/kserve/CUSTOM-README.md
+1
View File
@@ -29,6 +29,7 @@
| `inferenceService.model.args[1]` | 사용할 모델 이름 | `--model_name=phi-1.5` | | `inferenceService.model.args[1]` | 사용할 모델 이름 | `--model_name=phi-1.5` |
| `inferenceService.model.args[2]` | 데이터 타입 설정 | `--dtype=float16` | | `inferenceService.model.args[2]` | 데이터 타입 설정 | `--dtype=float16` |
| `inferenceService.model.storageUri` | 모델 파일 저장 위치 | `pvc://model-volume/phi-1_5` | | `inferenceService.model.storageUri` | 모델 파일 저장 위치 | `pvc://model-volume/phi-1_5` |
| `inferenceService.model.image` | 모델 서빙 이미지 | `kserve/huggingfaceserver:v0.15.0-gpu` |
### 3) 리소스 설정 ### 3) 리소스 설정
charts/kserve/custom-values.yaml
+1
View File
@@ -8,6 +8,7 @@ inferenceService:
- --model_name=phi-1_5 - --model_name=phi-1_5
- --dtype=float16 - --dtype=float16
storageUri: "pvc://model-volume/phi-1_5" storageUri: "pvc://model-volume/phi-1_5"
image: kserve/huggingfaceserver:v0.15.0-gpu
resources: resources:
limits: limits:
charts/kserve/templates/inferenceservice.yaml
+3
View File
@@ -17,5 +17,8 @@ spec:
{{- toYaml . | nindent 8 }} {{- toYaml . | nindent 8 }}
{{- end }} {{- end }}
storageUri: {{ .Values.inferenceService.model.storageUri }} storageUri: {{ .Values.inferenceService.model.storageUri }}
{{- if .Values.inferenceService.model.image }}
image: {{ .Values.inferenceService.model.image }}
{{- end }}
resources: resources:
{{- toYaml .Values.inferenceService.resources | nindent 8 }} {{- toYaml .Values.inferenceService.resources | nindent 8 }}
charts/kserve/values.yaml
+1
View File
@@ -22,6 +22,7 @@ inferenceService:
- --model_name=phi-1.5 - --model_name=phi-1.5
- --dtype=float16 - --dtype=float16
storageUri: "pvc://model-volume/phi-1_5" storageUri: "pvc://model-volume/phi-1_5"
image: kserve/huggingfaceserver:v0.15.0-gpu
# 리소스 설정 # 리소스 설정
resources: resources: