Katib official release v0.14.0

.github/workflows/publish-trial-images.yaml

@@ -31,8 +31,10 @@ jobs:
         include:
           - trial-name: mxnet-mnist
             dockerfile: examples/v1beta1/trial-images/mxnet-mnist/Dockerfile
-          - trial-name: pytorch-mnist
-            dockerfile: examples/v1beta1/trial-images/pytorch-mnist/Dockerfile
+          - trial-name: pytorch-mnist-cpu
+            dockerfile: examples/v1beta1/trial-images/pytorch-mnist/Dockerfile.cpu
+          - trial-name: pytorch-mnist-gpu
+            dockerfile: examples/v1beta1/trial-images/pytorch-mnist/Dockerfile.gpu
           - trial-name: tf-mnist-with-summaries
             dockerfile: examples/v1beta1/trial-images/tf-mnist-with-summaries/Dockerfile
           - trial-name: enas-cnn-cifar10-gpu

.github/workflows/pytorch-mnist-e2e-test.yaml

@@ -24,7 +24,7 @@ jobs:
           experiments: ${{ matrix.experiments }}
           training-operator: true
           # Comma Delimited
-          trial-images: pytorch-mnist
+          trial-images: pytorch-mnist-cpu
 
     strategy:
       fail-fast: false

docs/images-location.md

@@ -273,13 +273,24 @@ The following table shows images for training containers which are used in the
     </tr>
     <tr align="center">
       <td>
-        <code>docker.io/kubeflowkatib/pytorch-mnist</code>
+        <code>docker.io/kubeflowkatib/pytorch-mnist-cpu</code>
       </td>
       <td>
-        PyTorch MNIST example with printing metrics to the file or StdOut
+        PyTorch MNIST example with printing metrics to the file or StdOut with CPU support
       </td>
       <td>
-        <a href="https://github.com/kubeflow/katib/blob/master/examples/v1beta1/trial-images/pytorch-mnist/Dockerfile">Dockerfile</a>
+        <a href="https://github.com/kubeflow/katib/blob/master/examples/v1beta1/trial-images/pytorch-mnist/Dockerfile.cpu">Dockerfile</a>
+      </td>
+    </tr>
+    <tr align="center">
+      <td>
+        <code>docker.io/kubeflowkatib/pytorch-mnist-gpu</code>
+      </td>
+      <td>
+        PyTorch MNIST example with printing metrics to the file or StdOut with GPU support
+      </td>
+      <td>
+        <a href="https://github.com/kubeflow/katib/blob/master/examples/v1beta1/trial-images/pytorch-mnist/Dockerfile.gpu">Dockerfile</a>
       </td>
     </tr>
     <tr align="center">

examples/v1beta1/argo/argo-workflow.yaml

@@ -76,7 +76,7 @@ spec:
                 - name: num-examples
             container:
               name: model-training
-              image: docker.io/kubeflowkatib/mxnet-mnist:latest
+              image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
               command:
                 - "python3"
                 - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/early-stopping/median-stop-with-json-format.yaml

@@ -62,7 +62,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/pytorch-mnist:latest
+                image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.14.0
                 command:
                   - "python3"
                   - "/opt/pytorch-mnist/mnist.py"

examples/v1beta1/early-stopping/median-stop.yaml

@@ -54,7 +54,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/hp-tuning/bayesian-optimization.yaml

@@ -57,7 +57,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/hp-tuning/cma-es.yaml

@@ -57,7 +57,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/hp-tuning/grid.yaml

@@ -55,7 +55,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/hp-tuning/hyperband.yaml

@@ -69,7 +69,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/hp-tuning/multivariate-tpe.yaml

@@ -54,7 +54,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/hp-tuning/random.yaml

@@ -54,7 +54,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/hp-tuning/simple-pbt.yaml

@@ -43,7 +43,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/simple-pbt:latest
+                image: docker.io/kubeflowkatib/simple-pbt:v0.14.0
                 command:
                   - "python3"
                   - "/opt/pbt/pbt_test.py"

examples/v1beta1/hp-tuning/sobol.yaml

@@ -54,7 +54,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/hp-tuning/tpe.yaml

@@ -54,7 +54,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/kubeflow-training-operator/pytorchjob-mnist.yaml

@@ -46,7 +46,7 @@ spec:
               spec:
                 containers:
                   - name: pytorch
-                    image: docker.io/kubeflowkatib/pytorch-mnist:latest
+                    image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.14.0
                     command:
                       - "python3"
                       - "/opt/pytorch-mnist/mnist.py"
@@ -61,7 +61,7 @@ spec:
               spec:
                 containers:
                   - name: pytorch
-                    image: docker.io/kubeflowkatib/pytorch-mnist:latest
+                    image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.14.0
                     command:
                       - "python3"
                       - "/opt/pytorch-mnist/mnist.py"

examples/v1beta1/kubeflow-training-operator/tfjob-mnist-with-summaries.yaml

@@ -56,7 +56,7 @@ spec:
               spec:
                 containers:
                   - name: tensorflow
-                    image: docker.io/kubeflowkatib/tf-mnist-with-summaries:latest
+                    image: docker.io/kubeflowkatib/tf-mnist-with-summaries:v0.14.0
                     command:
                       - "python"
                       - "/opt/tf-mnist-with-summaries/mnist.py"

examples/v1beta1/metrics-collector/custom-metrics-collector.yaml

@@ -67,7 +67,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/pytorch-mnist:latest
+                image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.14.0
                 command:
                   - "python3"
                   - "/opt/pytorch-mnist/mnist.py"

examples/v1beta1/metrics-collector/file-metrics-collector-with-json-format.yaml

@@ -52,7 +52,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/pytorch-mnist:latest
+                image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.14.0
                 command:
                   - "python3"
                   - "/opt/pytorch-mnist/mnist.py"

examples/v1beta1/metrics-collector/file-metrics-collector.yaml

@@ -54,7 +54,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/pytorch-mnist:latest
+                image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.14.0
                 command:
                   - "python3"
                   - "/opt/pytorch-mnist/mnist.py"

examples/v1beta1/metrics-collector/metrics-collection-strategy.yaml

@@ -59,7 +59,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/nas/darts-cpu.yaml

@@ -60,7 +60,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/darts-cnn-cifar10-cpu:latest
+                image: docker.io/kubeflowkatib/darts-cnn-cifar10-cpu:v0.14.0
                 command:
                   - python3
                   - run_trial.py

examples/v1beta1/nas/darts-gpu.yaml

@@ -77,7 +77,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/darts-cnn-cifar10-gpu:latest
+                image: docker.io/kubeflowkatib/darts-cnn-cifar10-gpu:v0.14.0
                 command:
                   - python3
                   - run_trial.py

examples/v1beta1/nas/enas-cpu.yaml

@@ -139,7 +139,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/enas-cnn-cifar10-cpu:latest
+                image: docker.io/kubeflowkatib/enas-cnn-cifar10-cpu:v0.14.0
                 command:
                   - python3
                   - -u

examples/v1beta1/nas/enas-gpu.yaml

@@ -136,7 +136,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/enas-cnn-cifar10-gpu:latest
+                image: docker.io/kubeflowkatib/enas-cnn-cifar10-gpu:v0.14.0
                 command:
                   - python3
                   - -u

examples/v1beta1/resume-experiment/from-volume-resume.yaml

@@ -55,7 +55,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/resume-experiment/never-resume.yaml

@@ -55,7 +55,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/tekton/pipeline-run.yaml

@@ -89,7 +89,7 @@ spec:
                     description: Number of training examples
                 steps:
                   - name: model-training
-                    image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                    image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                     command:
                       - "python3"
                       - "/opt/mxnet-mnist/mnist.py"

examples/v1beta1/trial-images/pytorch-mnist/Dockerfile.gpu

@@ -0,0 +1,15 @@
+FROM pytorch/pytorch:1.11.0-cuda11.3-cudnn8-runtime
+
+ADD examples/v1beta1/trial-images/pytorch-mnist /opt/pytorch-mnist
+WORKDIR /opt/pytorch-mnist
+
+# Add folder for the logs.
+RUN mkdir /katib
+RUN pip install --no-cache-dir -r requirements.txt
+
+RUN chgrp -R 0 /opt/pytorch-mnist \
+  && chmod -R g+rwX /opt/pytorch-mnist \
+  && chgrp -R 0 /katib \
+  && chmod -R g+rwX /katib
+
+ENTRYPOINT ["python3", "/opt/pytorch-mnist/mnist.py"]

examples/v1beta1/trial-template/trial-metadata-substitution.yaml

@@ -60,7 +60,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

manifests/v1beta1/components/controller/katib-config.yaml

@@ -8,13 +8,13 @@ data:
   metrics-collector-sidecar: |-
     {
       "StdOut": {
-        "image": "docker.io/kubeflowkatib/file-metrics-collector:latest"
+        "image": "docker.io/kubeflowkatib/file-metrics-collector:v0.14.0"
       },
       "File": {
-        "image": "docker.io/kubeflowkatib/file-metrics-collector:latest"
+        "image": "docker.io/kubeflowkatib/file-metrics-collector:v0.14.0"
       },
       "TensorFlowEvent": {
-        "image": "docker.io/kubeflowkatib/tfevent-metrics-collector:latest",
+        "image": "docker.io/kubeflowkatib/tfevent-metrics-collector:v0.14.0",
         "resources": {
           "limits": {
             "memory": "1Gi"
@@ -25,31 +25,31 @@ data:
   suggestion: |-
     {
       "random": {
-        "image": "docker.io/kubeflowkatib/suggestion-hyperopt:latest"
+        "image": "docker.io/kubeflowkatib/suggestion-hyperopt:v0.14.0"
       },
       "tpe": {
-        "image": "docker.io/kubeflowkatib/suggestion-hyperopt:latest"
+        "image": "docker.io/kubeflowkatib/suggestion-hyperopt:v0.14.0"
       },
       "grid": {
-        "image": "docker.io/kubeflowkatib/suggestion-chocolate:latest"
+        "image": "docker.io/kubeflowkatib/suggestion-chocolate:v0.14.0"
       },
       "hyperband": {
-        "image": "docker.io/kubeflowkatib/suggestion-hyperband:latest"
+        "image": "docker.io/kubeflowkatib/suggestion-hyperband:v0.14.0"
       },
       "bayesianoptimization": {
-        "image": "docker.io/kubeflowkatib/suggestion-skopt:latest"
+        "image": "docker.io/kubeflowkatib/suggestion-skopt:v0.14.0"
       },
       "cmaes": {
-        "image": "docker.io/kubeflowkatib/suggestion-goptuna:latest"
+        "image": "docker.io/kubeflowkatib/suggestion-goptuna:v0.14.0"
       },
       "sobol": {
-        "image": "docker.io/kubeflowkatib/suggestion-goptuna:latest"
+        "image": "docker.io/kubeflowkatib/suggestion-goptuna:v0.14.0"
       },
       "multivariate-tpe": {
-        "image": "docker.io/kubeflowkatib/suggestion-optuna:latest"
+        "image": "docker.io/kubeflowkatib/suggestion-optuna:v0.14.0"
       },
       "enas": {
-        "image": "docker.io/kubeflowkatib/suggestion-enas:latest",
+        "image": "docker.io/kubeflowkatib/suggestion-enas:v0.14.0",
         "resources": {
           "limits": {
             "memory": "200Mi"
@@ -57,10 +57,10 @@ data:
         }
       },
       "darts": {
-        "image": "docker.io/kubeflowkatib/suggestion-darts:latest"
+        "image": "docker.io/kubeflowkatib/suggestion-darts:v0.14.0"
       },
       "pbt": {
-        "image": "docker.io/kubeflowkatib/suggestion-pbt:latest",
+        "image": "docker.io/kubeflowkatib/suggestion-pbt:v0.14.0",
         "persistentVolumeClaimSpec": {
           "accessModes": [
             "ReadWriteMany"
@@ -76,6 +76,6 @@ data:
   early-stopping: |-
     {
       "medianstop": {
-        "image": "docker.io/kubeflowkatib/earlystopping-medianstop:latest"
+        "image": "docker.io/kubeflowkatib/earlystopping-medianstop:v0.14.0"
       }
     }

manifests/v1beta1/components/controller/trial-templates.yaml

@@ -15,7 +15,7 @@ data:
         spec:
           containers:
             - name: training-container
-              image: docker.io/kubeflowkatib/mxnet-mnist:latest
+              image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
               command:
                 - "python3"
                 - "/opt/mxnet-mnist/mnist.py"
@@ -33,7 +33,7 @@ data:
         spec:
           containers:
             - name: training-container
-              image: docker.io/kubeflowkatib/enas-cnn-cifar10-cpu:latest
+              image: docker.io/kubeflowkatib/enas-cnn-cifar10-cpu:v0.14.0
               command:
                 - python3
                 - -u
@@ -54,7 +54,7 @@ data:
             spec:
               containers:
                 - name: pytorch
-                  image: docker.io/kubeflowkatib/pytorch-mnist:latest
+                  image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.14.0
                   command:
                     - "python3"
                     - "/opt/pytorch-mnist/mnist.py"
@@ -68,7 +68,7 @@ data:
             spec:
               containers:
                 - name: pytorch
-                  image: docker.io/kubeflowkatib/pytorch-mnist:latest
+                  image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.14.0
                   command:
                     - "python3"
                     - "/opt/pytorch-mnist/mnist.py"

manifests/v1beta1/installs/katib-cert-manager/kustomization.yaml

@@ -22,13 +22,13 @@ resources:
 images:
   - name: docker.io/kubeflowkatib/katib-controller
     newName: docker.io/kubeflowkatib/katib-controller
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-db-manager
     newName: docker.io/kubeflowkatib/katib-db-manager
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-ui
     newName: docker.io/kubeflowkatib/katib-ui
-    newTag: latest
+    newTag: v0.14.0
 
 patchesStrategicMerge:
   - patches/katib-cert-injection.yaml

manifests/v1beta1/installs/katib-external-db/kustomization.yaml

@@ -20,16 +20,16 @@ resources:
 images:
   - name: docker.io/kubeflowkatib/katib-controller
     newName: docker.io/kubeflowkatib/katib-controller
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-db-manager
     newName: docker.io/kubeflowkatib/katib-db-manager
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-ui
     newName: docker.io/kubeflowkatib/katib-ui
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/cert-generator
     newName: docker.io/kubeflowkatib/cert-generator
-    newTag: latest
+    newTag: v0.14.0
 patchesStrategicMerge:
   - patches/db-manager.yaml
 # Modify katib-mysql-secrets with parameters for the DB.

manifests/v1beta1/installs/katib-openshift/kustomization.yaml

@@ -30,13 +30,13 @@ resources:
 images:
   - name: docker.io/kubeflowkatib/katib-controller
     newName: docker.io/kubeflowkatib/katib-controller
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-db-manager
     newName: docker.io/kubeflowkatib/katib-db-manager
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-ui
     newName: docker.io/kubeflowkatib/katib-ui
-    newTag: latest
+    newTag: v0.14.0
 
 patchesJson6902:
   # Annotate Service to delegate TLS-secret generation to OpenShift service controller

manifests/v1beta1/installs/katib-standalone/kustomization.yaml

@@ -22,13 +22,13 @@ resources:
 images:
   - name: docker.io/kubeflowkatib/katib-controller
     newName: docker.io/kubeflowkatib/katib-controller
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-db-manager
     newName: docker.io/kubeflowkatib/katib-db-manager
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-ui
     newName: docker.io/kubeflowkatib/katib-ui
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/cert-generator
     newName: docker.io/kubeflowkatib/cert-generator
-    newTag: latest
+    newTag: v0.14.0

manifests/v1beta1/installs/katib-with-kubeflow/kustomization.yaml

@@ -10,13 +10,13 @@ resources:
 images:
   - name: docker.io/kubeflowkatib/katib-controller
     newName: docker.io/kubeflowkatib/katib-controller
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-db-manager
     newName: docker.io/kubeflowkatib/katib-db-manager
-    newTag: latest
+    newTag: v0.14.0
   - name: docker.io/kubeflowkatib/katib-ui
     newName: docker.io/kubeflowkatib/katib-ui
-    newTag: latest
+    newTag: v0.14.0
 
 patchesStrategicMerge:
   - patches/remove-namespace.yaml

operators/katib-controller/src/defaultTrialTemplate.yaml

@@ -5,7 +5,7 @@ spec:
     spec:
       containers:
         - name: training-container
-          image: docker.io/kubeflowkatib/mxnet-mnist:latest
+          image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
           command:
             - "python3"
             - "/opt/mxnet-mnist/mnist.py"

operators/katib-controller/src/enasCPUTemplate.yaml

@@ -5,7 +5,7 @@ spec:
     spec:
       containers:
         - name: training-container
-          image: docker.io/kubeflowkatib/enas-cnn-cifar10-cpu:latest
+          image: docker.io/kubeflowkatib/enas-cnn-cifar10-cpu:v0.14.0
           command:
             - python3
             - -u

operators/katib-controller/src/pytorchJobTemplate.yaml

@@ -9,7 +9,7 @@ spec:
         spec:
           containers:
             - name: pytorch
-              image: docker.io/kubeflowkatib/pytorch-mnist:latest
+              image: docker.io/kubeflowkatib/pytorch-mnist:v0.14.0-rc.0
               imagePullPolicy: Always
               command:
                 - "python3"
@@ -24,7 +24,7 @@ spec:
         spec:
           containers:
             - name: pytorch
-              image: docker.io/kubeflowkatib/pytorch-mnist:latest
+              image: docker.io/kubeflowkatib/pytorch-mnist:v0.14.0-rc.0
               imagePullPolicy: Always
               command:
                 - "python3"

scripts/v1beta1/build.sh

@@ -123,8 +123,11 @@ else
   echo -e "\nBuilding mxnet mnist training container example...\n"
   docker build --platform linux/amd64 -t "${REGISTRY}/mxnet-mnist:${TAG}" -f examples/${VERSION}/trial-images/mxnet-mnist/Dockerfile .
 
-  echo -e "\nBuilding PyTorch mnist training container example...\n"
-  docker build --platform linux/amd64 -t "${REGISTRY}/pytorch-mnist:${TAG}" -f examples/${VERSION}/trial-images/pytorch-mnist/Dockerfile .
+  echo -e "\nBuilding PyTorch mnist training container example with CPU support...\n"
+  docker build --platform linux/amd64 -t "${REGISTRY}/pytorch-mnist-cpu:${TAG}" -f examples/${VERSION}/trial-images/pytorch-mnist/Dockerfile.cpu .
+
+  echo -e "\nBuilding PyTorch mnist training container example with GPU support...\n"
+  docker build --platform linux/amd64 -t "${REGISTRY}/pytorch-mnist-gpu:${TAG}" -f examples/${VERSION}/trial-images/pytorch-mnist/Dockerfile.gpu .
 
   echo -e "\nBuilding Keras CIFAR-10 CNN training container example for ENAS with GPU support...\n"
   docker build --platform linux/amd64 -t "${REGISTRY}/enas-cnn-cifar10-gpu:${TAG}" -f examples/${VERSION}/trial-images/enas-cnn-cifar10/Dockerfile.gpu .

scripts/v1beta1/push.sh

@@ -98,8 +98,11 @@ docker push "${REGISTRY}/mxnet-mnist:${TAG}"
 echo -e "\nPushing Tensorflow with summaries mnist training container example...\n"
 docker push "${REGISTRY}/tf-mnist-with-summaries:${TAG}"
 
-echo -e "\nPushing PyTorch mnist training container example...\n"
-docker push "${REGISTRY}/pytorch-mnist:${TAG}"
+echo -e "\nPushing PyTorch mnist training container example with CPU support...\n"
+docker push "${REGISTRY}/pytorch-mnist-cpu:${TAG}"
+
+echo -e "\nPushing PyTorch mnist training container example with GPU support...\n"
+docker push "${REGISTRY}/pytorch-mnist-gpu:${TAG}"
 
 echo -e "\nPushing Keras CIFAR-10 CNN training container example for ENAS with GPU support...\n"
 docker push "${REGISTRY}/enas-cnn-cifar10-gpu:${TAG}"
@@ -107,8 +110,11 @@ docker push "${REGISTRY}/enas-cnn-cifar10-gpu:${TAG}"
 echo -e "\nPushing Keras CIFAR-10 CNN training container example for ENAS with CPU support...\n"
 docker push "${REGISTRY}/enas-cnn-cifar10-cpu:${TAG}"
 
-echo -e "\nPushing PyTorch CIFAR-10 CNN training container example for DARTS...\n"
-docker push "${REGISTRY}/darts-cnn-cifar10:${TAG}"
+echo -e "\nPushing PyTorch CIFAR-10 CNN training container example for DARTS with CPU support...\n"
+docker push "${REGISTRY}/darts-cnn-cifar10-cpu:${TAG}"
+
+echo -e "\nPushing PyTorch CIFAR-10 CNN training container example for DARTS with GPU support...\n"
+docker push "${REGISTRY}/darts-cnn-cifar10-gpu:${TAG}"
 
 echo -e "\nPushing dynamic learning rate training container example for PBT...\n"
 docker push "${REGISTRY}/simple-pbt:${TAG}"

scripts/v1beta1/update-images.sh

@@ -83,7 +83,8 @@ update_yaml_files "${CONFIG_PATH}" ":[^[:space:]].*\"" ":${TAG}\""
 
 # Postfixes for the each Trial image.
 MXNET_MNIST="mxnet-mnist"
-PYTORCH_MNIST="pytorch-mnist"
+PYTORCH_MNIST_CPU="pytorch-mnist-cpu"
+PYTORCH_MNIST_GPU="pytorch-mnist-gpu"
 TF_MNIST_WITH_SUMMARIES="tf-mnist-with-summaries"
 ENAS_GPU="enas-cnn-cifar10-gpu"
 ENAS_CPU="enas-cnn-cifar10-cpu"
@@ -93,7 +94,8 @@ SIMPLE_PBT="simple-pbt"
 
 echo -e "Update Katib Trial training container images\n"
 update_yaml_files "./" "${OLD_PREFIX}${MXNET_MNIST}:.*" "${NEW_PREFIX}${MXNET_MNIST}:${TAG}"
-update_yaml_files "./" "${OLD_PREFIX}${PYTORCH_MNIST}:.*" "${NEW_PREFIX}${PYTORCH_MNIST}:${TAG}"
+update_yaml_files "./" "${OLD_PREFIX}${PYTORCH_MNIST_CPU}:.*" "${NEW_PREFIX}${PYTORCH_MNIST_CPU}:${TAG}"
+update_yaml_files "./" "${OLD_PREFIX}${PYTORCH_MNIST_GPU}:.*" "${NEW_PREFIX}${PYTORCH_MNIST_GPU}:${TAG}"
 update_yaml_files "./" "${OLD_PREFIX}${TF_MNIST_WITH_SUMMARIES}:.*" "${NEW_PREFIX}${TF_MNIST_WITH_SUMMARIES}:${TAG}"
 update_yaml_files "./" "${OLD_PREFIX}${ENAS_GPU}:.*" "${NEW_PREFIX}${ENAS_GPU}:${TAG}"
 update_yaml_files "./" "${OLD_PREFIX}${ENAS_CPU}:.*" "${NEW_PREFIX}${ENAS_CPU}:${TAG}"

sdk/python/v1beta1/setup.py

@@ -19,7 +19,7 @@ with open('requirements.txt') as f:
 
 setuptools.setup(
     name='kubeflow-katib',
-    version='0.13.0',
+    version='0.14.0',
     author="Kubeflow Authors",
     author_email='premnath.vel@gmail.com',
     license="Apache License Version 2.0",

test/e2e/v1beta1/scripts/gh-actions/build-load.sh

@@ -30,7 +30,7 @@ REGISTRY="docker.io/kubeflowkatib"
 TAG="e2e-test"
 VERSION="v1beta1"
 CMD_PREFIX="cmd"
-SPECIFIED_DEVICE_TYPE_IMAGES=("enas-cnn-cifar10-cpu" "darts-cnn-cifar10-cpu")
+SPECIFIED_DEVICE_TYPE_IMAGES=("enas-cnn-cifar10-cpu" "darts-cnn-cifar10-cpu" "pytorch-mnist-cpu")
 
 IFS="," read -r -a TRIAL_IMAGE_ARRAY <<< "$TRIAL_IMAGES"
 IFS="," read -r -a EXPERIMENT_ARRAY <<< "$EXPERIMENTS"
@@ -51,7 +51,7 @@ _build_containers() {
   docker build --platform "$(uname -m)" -t "$REGISTRY/$CONTAINER_NAME:$TAG" -f "../../../../../$DOCKERFILE" ../../../../../
 }
 
-_load_kind_cluster() {
+_load_minikube_cluster() {
   CONTAINER_NAME=${1:-"katib-controller"}
 
   echo -e "\n\nLoading $CONTAINER_NAME image...\n\n"
@@ -99,7 +99,7 @@ run() {
     for s in "${suggestions[@]}"; do
       if [ "$s" == "$CONTAINER_NAME" ]; then
         _build_containers "$CONTAINER_NAME" "$DOCKERFILE"
-        _load_kind_cluster "$CONTAINER_NAME"
+        _load_minikube_cluster "$CONTAINER_NAME"
         break
       fi
     done
@@ -126,7 +126,7 @@ run() {
     for e in "${earlystoppings[@]}"; do
       if [ "$e" == "$CONTAINER_NAME" ]; then
         _build_containers "$CONTAINER_NAME" "$DOCKERFILE"
-        _load_kind_cluster "$CONTAINER_NAME"
+        _load_minikube_cluster "$CONTAINER_NAME"
         break
       fi
     done
@@ -134,7 +134,7 @@ run() {
   # Others
   else
     _build_containers "$CONTAINER_NAME" "$DOCKERFILE"
-    _load_kind_cluster "$CONTAINER_NAME"
+    _load_minikube_cluster "$CONTAINER_NAME"
   fi
 }
 

test/e2e/v1beta1/testdata/invalid-experiment.yaml

@@ -52,7 +52,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"

test/e2e/v1beta1/testdata/valid-experiment.yaml

@@ -52,7 +52,7 @@ spec:
           spec:
             containers:
               - name: training-container
-                image: docker.io/kubeflowkatib/mxnet-mnist:latest
+                image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0
                 command:
                   - "python3"
                   - "/opt/mxnet-mnist/mnist.py"