workbench + pipelines + models update

- update of Python libraries in workbenches and pipelines
- change Flan-T5 for Qwen-2.5-0.5B quantized on vLLM
- update pipelines code
- rebuild workbench and pipeline images

Author: guimou

bootstrap/ic-rhoai-configuration/workbench-imagestream.yaml

@@ -22,17 +22,17 @@ spec:
   lookupPolicy:
     local: true
   tags:
-    - name: "1.2"
+    - name: "3.0.4"
       annotations:
         opendatahub.io/notebook-python-dependencies: >-
-          [{"name":"PyTorch","version":"2.2.2"},{"name":"Langchain","version":"0.3.1"},{"name":"Ultralytics","version":"8.2.100"},]
+          [{"name":"PyTorch","version":"2.6.0"},{"name":"Langchain","version":"0.3.25"},{"name":"Ultralytics","version":"8.3.151"},]
         opendatahub.io/notebook-software: >-
-          [{"name":"Python","version":"v3.9"}]
-        openshift.io/imported-from: quay.io/rh-aiservices-bu/rhoai-lab-parasol-insurance
+          [{"name":"Python","version":"v3.11"}]
+        openshift.io/imported-from: quay.io/rh-aiservices-bu/rhoai-lab-insurance-claim-workbench
       from:
         kind: DockerImage
         name: >-
-          quay.io/rh-aiservices-bu/rhoai-lab-parasol-insurance:1.2
+          quay.io/rh-aiservices-bu/rhoai-lab-insurance-claim-workbench:3.0.4
       importPolicy:
         importMode: Legacy
       referencePolicy:
@@ -61,15 +61,15 @@ spec:
   lookupPolicy:
     local: true
   tags:
-    - name: "2024.1"
+    - name: "2025.1"
       annotations:
-        opendatahub.io/notebook-python-dependencies: '[{"name":"code-server","version":"4.22"}]'
-        opendatahub.io/notebook-software: '[{"name":"Python","version":"v3.9"}]'
+        opendatahub.io/notebook-python-dependencies: '[{"name":"code-server","version":"4.98"}]'
+        opendatahub.io/notebook-software: '[{"name":"Python","version":"v3.11"}]'
         openshift.io/imported-from: quay.io/modh/codeserver
       from:
         kind: DockerImage
         name: >-
-          quay.io/modh/codeserver@sha256:4b6b563e75f9e642bac9f21c4a2d8d72d77923fc391dd5dc985d4489c7258ef5
+          quay.io/modh/codeserver@sha256:38460e3b5536d903add40869714512e2c3a993c3dfe4a8010688e4e3b79bf090
       importPolicy:
         importMode: Legacy
       referencePolicy:

bootstrap/ic-shared-llm/Qwen2.5-0.5B-quantized.w8a8-modelcar.Containerfile

@@ -0,0 +1,17 @@
+FROM quay.io/redhat-ai-services/huggingface-modelcar-builder:latest as base
+
+# Set the HF_TOKEN with --build-arg HF_TOKEN="hf_..." at build time
+ARG HF_TOKEN
+
+# The model repo to download
+ENV MODEL_REPO="RedHatAI/Qwen2.5-0.5B-quantized.w8a8"
+
+# Download the necessary model files
+RUN python3 download_model.py --model-repo ${MODEL_REPO}
+
+# Final image containing only the essential model files
+FROM registry.access.redhat.com/ubi9/ubi-micro:9.5
+
+COPY --from=base /models /models
+
+USER 1001

bootstrap/ic-shared-llm/inference-service-qwen-modelcar.yaml

@@ -0,0 +1,43 @@
+apiVersion: serving.kserve.io/v1beta1
+kind: InferenceService
+metadata:
+  annotations:
+    openshift.io/display-name: qwen
+    serving.knative.openshift.io/enablePassthrough: 'true'
+    sidecar.istio.io/inject: 'true'
+    sidecar.istio.io/rewriteAppHTTPProbers: 'true'
+    argocd.argoproj.io/sync-wave: "2"
+    serving.kserve.io/deploymentMode: RawDeployment
+    argocd.argoproj.io/compare-options: IgnoreExtraneous
+    argocd.argoproj.io/sync-options: Prune=false
+  name: qwen
+  namespace: ic-shared-llm
+  labels:
+    opendatahub.io/dashboard: 'true'
+spec:
+  predictor:
+    maxReplicas: 2
+    minReplicas: 2
+    model:
+      args:
+        - '--port=8080'
+        - '--model=/mnt/models'
+        - '--served-model-name=qwen2.5'
+        - '--max-model-len=2048'
+        - '--tensor-parallel-size=1'
+      modelFormat:
+        name: vLLM
+      name: ''
+      resources:
+        limits:
+          cpu: '4'
+          memory: 6Gi
+        requests:
+          cpu: '4'
+          memory: 5Gi
+      runtime: vllm-cpu
+      storageUri: oci://quay.io/rh-aiservices-bu/qwen2.5-0.5b-quantized.w8a8-modelcar:0.0.1
+    tolerations:
+      - effect: NoSchedule
+        key: nvidia.com/gpu
+        operator: Exists

bootstrap/ic-shared-llm/kustomization.yaml

@@ -13,8 +13,10 @@ resources:
   - namespace.yaml
   # wave 1
   - serving-runtime-vllm-granite-modelcar.yaml
+  - serving-runtime-vllm-cpu-qwen-modelcar.yaml
   # wave 2
   - inference-service-granite-modelcar.yaml
+  - inference-service-qwen-modelcar.yaml
   - pvc-hftgi.yaml
   - deployment-hftgi.yaml
   - service-hftgi.yaml

bootstrap/ic-shared-llm/serving-runtime-vllm-cpu-qwen-modelcar.yaml

@@ -0,0 +1,38 @@
+apiVersion: serving.kserve.io/v1alpha1
+kind: ServingRuntime
+labels:
+  opendatahub.io/dashboard: "true"
+metadata:
+  annotations:
+    openshift.io/display-name: vLLM-CPU
+    opendatahub.io/apiProtocol: REST
+    opendatahub.io/recommended-accelerators: '[]'
+    opendatahub.io/template-display-name: vLLM ServingRuntime for KServe
+    opendatahub.io/template-name: vllm-runtime-cpu
+    argocd.argoproj.io/sync-wave: "1"
+  name: vllm-cpu
+spec:
+  builtInAdapter:
+    modelLoadingTimeoutMillis: 90000
+  containers:
+    - command:
+        - python
+        - '-m'
+        - vllm.entrypoints.openai.api_server
+      image: public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.9.1
+      env:
+        - name: VLLM_CPU_KVCACHE_SPACE
+          value: "2"
+        - name: OMP_NUM_THREADS
+          value: "1"
+        - name: VLLM_CPU_OMP_THREADS_BIND
+          value: "all"
+      name: kserve-container
+      ports:
+        - containerPort: 8080
+          name: http1
+          protocol: TCP
+  multiModel: false
+  supportedModelFormats:
+    - autoSelect: true
+      name: vLLM

bootstrap/ic-user-projects/images-puller.yaml

@@ -19,7 +19,7 @@ spec:
     spec:
       containers:
       - name: ic-workbench
-        image: image-registry.openshift-image-registry.svc:5000/redhat-ods-applications/ic-workbench:1.2
+        image: image-registry.openshift-image-registry.svc:5000/redhat-ods-applications/ic-workbench:3.0.4
         command: ["tail"]
         args: ["-f", "/dev/null"]
         resources:
@@ -29,7 +29,7 @@ spec:
             cpu: 10m
             memory: 10Mi
       - name: ic-code-server
-        image: quay.io/modh/codeserver@sha256:4b6b563e75f9e642bac9f21c4a2d8d72d77923fc391dd5dc985d4489c7258ef5
+        image: image-registry.openshift-image-registry.svc:5000/redhat-ods-applications/ic-code-server:2025.1
         command: ["tail"]
         args: ["-f", "/dev/null"]
         resources:
@@ -57,7 +57,7 @@ spec:
     spec:
       containers:
       - name: oauth-proxy
-        image: registry.redhat.io/openshift4/ose-oauth-proxy@sha256:4bef31eb993feb6f1096b51b4876c65a6fb1f4401fee97fa4f4542b6b7c9bc46
+        image: registry.redhat.io/openshift4/ose-oauth-proxy@sha256:8507daed246d4d367704f7d7193233724acf1072572e1226ca063c066b858ecf
         command: ["tail"]
         args: ["-f", "/dev/null"]
         resources:
@@ -67,7 +67,7 @@ spec:
             cpu: 10m
             memory: 10Mi
       - name: model-server
-        image: quay.io/modh/openvino_model_server@sha256:9086c1ba1ba30d358194c534f0563923aab02d03954e43e9f3647136b44a5daf
+        image: quay.io/modh/openvino_model_server@sha256:53b7fcf95de9b81e4c8652d0bf4e84e22d5b696827a5d951d863420c68b9cfe8
         command: ["tail"]
         args: ["-f", "/dev/null"]
         resources:
@@ -77,7 +77,7 @@ spec:
             cpu: 10m
             memory: 10Mi
       - name: runtime-adapter
-        image: registry.redhat.io/rhoai/odh-modelmesh-runtime-adapter-rhel8@sha256:7d085942e4da73c2ff4885c62f10b904dc70fd08e819e62f925eb1b4ddfa1554
+        image: registry.redhat.io/rhoai/odh-modelmesh-runtime-adapter-rhel8@sha256:12d75776a60c119d938dec28625f574f5d55a2616c49bb8773e5b87ba3141280
         command: ["tail"]
         args: ["-f", "/dev/null"]
         resources:
@@ -87,7 +87,7 @@ spec:
             cpu: 10m
             memory: 10Mi
       - name: modelmesh
-        image: registry.redhat.io/rhoai/odh-modelmesh-rhel8@sha256:bf25fce160dd7b8cce11e68a501006e4fa81a7858645038cdefdbba03df7ecdb
+        image: registry.redhat.io/rhoai/odh-modelmesh-rhel8@sha256:43d98c254ad76c2eefc48a56e84e4083281c72e8b9ae38e49dfb6d5751bdb895
         command: ["tail"]
         args: ["-f", "/dev/null"]
         resources:
@@ -115,7 +115,7 @@ spec:
     spec:
       containers:
         - name: mariadb
-          image: registry.redhat.io/rhel8/mariadb-103@sha256:3d30992e60774f887c4e7959c81b0c41b0d82d042250b3b56f05ab67fd4cdee1
+          image: registry.redhat.io/rhel8/mariadb-103@sha256:f0ee0d27bb784e289f7d88cc8ee0e085ca70e88a5d126562105542f259a1ac01
           command: ["tail"]
           args: ["-f", "/dev/null"]
           resources:
@@ -125,7 +125,7 @@ spec:
               cpu: 10m
               memory: 10Mi
         - name: ds-pipeline-scheduledworkflow-pipelines-definition
-          image: registry.redhat.io/rhoai/odh-ml-pipelines-scheduledworkflow-v2-rhel8@sha256:c2bd493e9345eccd124bbe0440e45cce8cf57588827d8eb9533f8bbadaa3ebbe
+          image: registry.redhat.io/rhoai/odh-ml-pipelines-scheduledworkflow-v2-rhel8@sha256:99ca5a981184f0dd3f94105172b20494328c5c7a762fb027905e108710b7b5d4
           command: ["tail"]
           args: ["-f", "/dev/null"]
           resources:
@@ -135,7 +135,7 @@ spec:
               cpu: 10m
               memory: 10Mi
         - name: ds-pipeline-pipelines-definition
-          image: registry.redhat.io/rhoai/odh-ml-pipelines-api-server-v2-rhel8@sha256:5d6354e65106460ab9285b42124faf288298387a8b16317565bd591f009c39a6
+          image: registry.redhat.io/rhoai/odh-ml-pipelines-api-server-v2-rhel8@sha256:da649491f5112763d9f563d64ca3455ed5e4309b15adab0886ae16e0bb367b0c
           command: ["tail"]
           args: ["-f", "/dev/null"]
           resources:
@@ -145,7 +145,7 @@ spec:
               cpu: 10m
               memory: 10Mi
         - name: oauth-proxy
-          image: registry.redhat.io/openshift4/ose-oauth-proxy@sha256:ab112105ac37352a2a4916a39d6736f5db6ab4c29bad4467de8d613e80e9bb33
+          image: registry.redhat.io/openshift4/ose-oauth-proxy@sha256:8ce44de8c683f198bf24ba36cd17e89708153d11f5b42c0a27e77f8fdb233551
           command: ["tail"]
           args: ["-f", "/dev/null"]
           resources:
@@ -155,7 +155,7 @@ spec:
               cpu: 10m
               memory: 10Mi
         - name: ds-pipeline-persistenceagent-pipelines-definition
-          image: registry.redhat.io/rhoai/odh-ml-pipelines-persistenceagent-v2-rhel8@sha256:b90b63e3c5de629eb89798d818218104158a5401220397ddbc7072a390e46bdd
+          image: registry.redhat.io/rhoai/odh-ml-pipelines-persistenceagent-v2-rhel8@sha256:4e0c858d68f925a307ee53498aa6019563254c5b27f5d7635debc0195db2c5bd
           command: ["tail"]
           args: ["-f", "/dev/null"]
           resources:
@@ -165,7 +165,7 @@ spec:
               cpu: 10m
               memory: 10Mi
         - name: ds-pipeline-runtime-image-confidence
-          image: quay.io/rh-aiservices-bu/rhoai-lab-insurance-claim-confidence-pipeline:1.2
+          image: quay.io/rh-aiservices-bu/rhoai-lab-insurance-claim-confidence-pipeline:2.0
           command: ["tail"]
           args: ["-f", "/dev/null"]
           resources:

bootstrap/workbench-image/Containerfile

@@ -1,4 +1,5 @@
-FROM quay.io/modh/odh-generic-data-science-notebook:v2-2024a-20240829
+FROM quay.io/modh/odh-generic-data-science-notebook:v3-20250320-3fa2d83
+
 
 ##########################
 # Deploy Python packages #
@@ -17,6 +18,11 @@ RUN echo "Installing packages" && \
     # Install Python packages \
     pip install --no-cache-dir --no-dependencies -r requirements.txt
 
+# Copy pycodestyle configuration
+COPY --chown=1001:0 pycodestyle /opt/app-root/etc/pycodestyle
+# Copy the custom start script
+COPY --chown=1001:0 start-notebook.sh /opt/app-root/bin/start-notebook.sh
+
 WORKDIR /opt/app-root/src
 
 ENTRYPOINT ["start-notebook.sh"]