Update tf job examples to tf v2 (kubeflow#2270)

* mnist with summaries updaetd to TF v2 Signed-off-by: yelias <[email protected]> * tf_sample updaetd to TF v2 Signed-off-by: yelias <[email protected]> * Add mnist_utils and update dist-mnist Signed-off-by: yelias <[email protected]> * Add mnist_utils and update dist-mnist Signed-off-by: yelias <[email protected]> * Remove old example - estimator-API, this example has been replaced by distribution_strategy Signed-off-by: yelias <[email protected]> * Small fix Signed-off-by: yelias <[email protected]> * Remove unsupported powerPC dockerfiles Signed-off-by: yelias <[email protected]> * Fix typo in copyright Signed-off-by: yelias <[email protected]> --------- Signed-off-by: yelias <[email protected]> Co-authored-by: yelias <[email protected]>
andreyvelich · Nov 4, 2024 · 95be3c0 · 95be3c0
1 parent 9e46f9d
commit 95be3c0
Show file tree

Hide file tree

Showing 23 changed files with 397 additions and 784 deletions.
diff --git a/.github/workflows/publish-example-images.yaml b/.github/workflows/publish-example-images.yaml
@@ -24,12 +24,9 @@ jobs:
           - component-name: tf-dist-mnist-test
             platforms: linux/amd64,linux/arm64
             dockerfile: examples/tensorflow/dist-mnist/Dockerfile
-          - component-name: tf-distributed-worker
-            platforms: linux/amd64,linux/arm64
-            dockerfile: examples/tensorflow/distribution_strategy/estimator-API/Dockerfile
           - component-name: tf-multi-worker-strategy
             platforms: linux/amd64,linux/arm64
-            dockerfile: examples/tensorflow/distribution_strategy/keras-API/Dockerfile
+            dockerfile: examples/tensorflow/distribution_strategy/Dockerfile
           - component-name: tf-mnist-with-summaries
             platforms: linux/amd64,linux/arm64
             dockerfile: examples/tensorflow/mnist_with_summaries/Dockerfile

diff --git a/examples/tensorflow/dist-mnist/Dockerfile b/examples/tensorflow/dist-mnist/Dockerfile
@@ -12,7 +12,13 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-FROM tensorflow/tensorflow:1.5.0
+FROM tensorflow/tensorflow:2.17.0
+
+# Using keras-2.17 because of bug on keras-3.4.1
+# which used by default by TF-2.17 (https://github.com/tensorflow/tensorflow/issues/72388)
+ENV TF_USE_LEGACY_KERAS 1
+RUN pip install tf_keras
 
 ADD examples/tensorflow/dist-mnist/ /var/tf_dist_mnist
+ADD examples/tensorflow/mnist_utils.py /var/tf_dist_mnist
 ENTRYPOINT ["python", "/var/tf_dist_mnist/dist_mnist.py"]
diff --git a/examples/tensorflow/dist-mnist/Dockerfile.ppc64le b/examples/tensorflow/dist-mnist/Dockerfile.ppc64le
diff --git a/examples/tensorflow/dist-mnist/README.md b/examples/tensorflow/dist-mnist/README.md
@@ -10,14 +10,9 @@ To build this image on x86_64:
 ```shell
 docker build -f Dockerfile -t kubeflow/tf-dist-mnist-test:1.0 ./
 ```
-To build this image on ppc64le:
-```shell
-docker build -f Dockerfile.ppc64le -t kubeflow123/tf-dist-mnist-test:1.0 ./
-```
 
 **Create TFJob YAML**
 
 ```
 kubectl create -f ./tf_job_mnist.yaml
 ```
-  * If on ppc64le, please update tf_job_mnist.yaml to use the image of ppc64le firstly.