Merge pull request #1005 from Lanture1064/dev

feat: add llama-factory chart

deploy/llama-factory/Chart.yaml

@@ -0,0 +1,15 @@
+apiVersion: v2
+name: llama-factory
+version: 0.1.1
+appVersion: 0.1.0
+description: llama-factory
+type: application
+
+keywords:
+  - llama-factory
+  - model tune
+sources:
+  - https://github.com/huangqg/LLaMA-Factory
+maintainers:
+  - name: huangqg
+    url: https://github.com/huangqg

deploy/llama-factory/README.md

@@ -0,0 +1 @@
+# llama-factory

deploy/llama-factory/templates/deployment.yaml

@@ -0,0 +1,43 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: llama-factory
+  namespace: {{ .Release.Namespace }}
+  labels:
+    app: llama-factory
+spec:
+  selector:
+    matchLabels:
+      app: llama-factory
+  replicas: 1
+  template:
+    metadata:
+      labels:
+        app: llama-factory
+    spec:
+      containers:
+      - image: {{ .Values.image }}
+        imagePullPolicy: {{ .Values.imagePullPolicy }}
+        name: llama-factory
+        ports:
+        - containerPort: 7860
+          protocol: TCP
+        resources: {{ toYaml .Values.resources | nindent 10 }}
+        volumeMounts:
+          - name: model-volume
+            mountPath: /root/.cache/huggingface
+          - name: data-volume
+            mountPath: /app/data
+          - name: output-volume
+            mountPath: /app/output
+      volumes:
+        - name: model-volume
+          hostPath:
+            path: {{ .Values.volume.model }}
+        - name: data-volume
+          hostPath:
+            path: {{ .Values.volume.data }}
+        - name: output-volume
+          hostPath:
+            path: {{ .Values.volume.output }}
+      terminationGracePeriodSeconds: 30

deploy/llama-factory/templates/ingress.yaml

@@ -0,0 +1,43 @@
+{{- if .Values.ingress.enable }}
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  annotations:
+    description: llama-factory-ingress
+    displayName: llama-factory-ingress
+    httpSend: /
+    ingress-lb: {{ .Values.ingress.ingressClassName }}
+    kubernetes.io/ingress.class: {{ .Values.ingress.ingressClassName }}
+    nginx.ingress.kubernetes.io/configuration-snippet: |
+      proxy_cache static-cache;
+      proxy_cache_valid 404 10m;
+      proxy_cache_use_stale error timeout updating http_404 http_500 http_502 http_503 http_504;
+      proxy_cache_bypass $http_x_purge;
+      add_header X-Cache-Status $upstream_cache_status; # check X-Cache-Status to see if it's HIT
+      rewrite ^/(llama-factory-apis)(/|$)(.*)$ /$3 break;
+    nginx.ingress.kubernetes.io/enable-access-log: "false"
+    nginx.ingress.kubernetes.io/enable-rewrite-log: "false"
+    nginx.ingress.kubernetes.io/load-balance: round_robin
+    nginx.ingress.kubernetes.io/proxy-body-size: ""
+    nginx.ingress.kubernetes.io/proxy-buffering: "on"
+    nginx.ingress.kubernetes.io/proxy-connect-timeout: "60"
+    nginx.ingress.kubernetes.io/server-alias: ""
+    nginx.ingress.kubernetes.io/ssl-redirect: "true"
+    nginx.ingress.kubernetes.io/upstream-vhost: $host
+  labels:
+    ingress-lb: {{ .Values.ingress.ingressClassName }}
+  name: llama-factory
+  namespace: {{ .Release.Namespace }}
+spec:
+  rules:
+    - host: portal.{{ .Values.ingress.ingressDomain }}
+      http:
+        paths:
+          - backend:
+              service:
+                name: llama-factory
+                port:
+                  number: 7860
+            path: /
+            pathType: ImplementationSpecific
+{{- end }}

deploy/llama-factory/templates/service.yaml

@@ -0,0 +1,17 @@
+apiVersion: v1
+kind: Service
+metadata:
+  labels:
+    app: llama-factory
+  name: llama-factory
+  namespace: {{ .Release.Namespace }}
+spec:
+  ports:
+    - name: http
+      port: 7860
+      protocol: TCP
+      targetPort: 7860
+  selector:
+    app: llama-factory
+  sessionAffinity: None
+  type: ClusterIP

deploy/llama-factory/values.yaml

@@ -0,0 +1,23 @@
+image: kubeagi/llama-factory:v0.6.1
+imagePullPolicy: IfNotPresent
+resources:
+  limits:
+    cpu: "4"
+    memory: 12Gi
+    nvidia.com/gpu: "1"
+  requests:
+    cpu: "1"
+    memory: 1Gi
+    nvidia.com/gpu: "1"
+
+volume:
+  model: /path/to/model/or/cache
+  data: /path/to/data
+  output: /path/to/output
+
+# ingress configurations for component
+ingress:
+  # set enable to `true` to enable ingress
+  enable: true
+  ingressClassName: portal-ingress
+  ingressDomain: your.domain.nip.io

deploy/llms/Dockerfile.fastchat-server

@@ -17,7 +17,8 @@ RUN apt-get update -y && apt-get install -y python3-distutils curl python3-pip p
 ARG PYTHON_INDEX_URL=https://pypi.mirrors.ustc.edu.cn/simple/
 
 WORKDIR /git
-RUN git clone https://github.com/lm-sys/FastChat.git
+RUN git clone https://github.com/lanture1064/FastK8ts.git
 
-WORKDIR /git/FastChat
-RUN pip install --upgrade pip -i ${PYTHON_INDEX_URL} && pip install -e . -i ${PYTHON_INDEX_URL}
+WORKDIR /git/FastK8ts
+RUN pip install --upgrade pip -i ${PYTHON_INDEX_URL} && pip install -e . -i ${PYTHON_INDEX_URL}
+RUN pip install pydantic==1.10.11 -i ${PYTHON_INDEX_URL}