kubeflow · k8s-ci-robot · Jul 25, 2018 · Jun 27, 2018 · Jun 28, 2018 · Jul 11, 2018
diff --git a/object_detection/README.md b/object_detection/README.md
@@ -8,4 +8,5 @@ This example is based on the TensorFlow [Pets tutorial](https://github.com/tenso
 1. [Setup a Kubeflow cluster](setup.md)
 2. [Submit a distributed object detection training job](submit_job.md)
 3. [Monitor your training job](monitor_job.md)
-4. [Serve your model with TensorFlow serving](export_tf_graph.md)
+4. [Export model](export_tf_graph.md)
+5. [Serve the model with GPU](tf_serving_gpu.md)
diff --git a/object_detection/export_tf_graph.md b/object_detection/export_tf_graph.md
@@ -33,30 +33,3 @@ kubectl -n kubeflow exec -it pets-training-master-r1hv-0-i6k7c sh
 mkdir /pets_data/exported_graphs/saved_model/1  
 cp /pets_data/exported_graphs/saved_model/* /pets_data/exported_graphs/saved_model/1  
 ```  
-
-## Serve the model using TF-Serving  
-
-Apply the manifest file under [tf-serving](./tf-serving) directory:  
-```  
-kubectl -n kubeflow apply -f ./tf-serving/tf-serving.yaml  
-```  
-
-After that you should see pets-model pod. Run:  
-```  
-kubectl -n kubeflow get pods | grep pets-model  
-```  
-That will output something like this:
-```  
-pets-model-v1-57674c8f76-4qrqp      1/1       Running     0          4h  
-```  
-Take a look at the logs:  
-```  
-kubectl -n kubeflow logs pets-model-v1-57674c8f76-4qrqp  
-```  
-And you should see:  
-```  
-2018-06-21 19:20:32.325406: I tensorflow_serving/core/loader_harness.cc:86] Successfully loaded servable version {name: pets-model version: 1}  
-E0621 19:20:34.134165172       7 ev_epoll1_linux.c:1051]     grpc epoll fd: 3  
-2018-06-21 19:20:34.135354: I tensorflow_serving/model_servers/main.cc:288] Running ModelServer at 0.0.0.0:9000 ...  
-```
-Now you can use a gRPC client to run inference using your trained model!
diff --git a/object_detection/ks-app/.gitignore b/object_detection/ks-app/.gitignore
@@ -0,0 +1,5 @@
+/lib
+/.ksonnet/registries
+/app.override.yaml
+/.ks_environment
+/environments
diff --git a/object_detection/ks-app/app.yaml b/object_detection/ks-app/app.yaml
@@ -0,0 +1,37 @@
+apiVersion: 0.1.0
+environments:
+  default:
+    destination:
+      namespace: kubeflow
+      server: https://35.193.147.35
+    k8sVersion: v1.9.6
+    path: default
+  kai4:
+    destination:
+      namespace: kubeflow
+      server: https://35.227.118.244
+    k8sVersion: v1.9.6
+    path: kai4
+kind: ksonnet.io/app
+libraries:
+  tf-serving:
+    gitVersion:
+      commitSha: 430d4f7ae16af389ac570182af519f6571f160cf
+      refSpec: master
+    name: tf-serving
+    registry: kubeflow
+name: object-detection-app
+registries:
+  incubator:
+    gitVersion:
+      commitSha: 40285d8a14f1ac5787e405e1023cf0c07f6aa28c
+      refSpec: master
+    protocol: github
+    uri: github.com/ksonnet/parts/tree/master/incubator
+  kubeflow:
+    gitVersion:
+      commitSha: 430d4f7ae16af389ac570182af519f6571f160cf
+      refSpec: master
+    protocol: github
+    uri: github.com/kubeflow/kubeflow/tree/master/kubeflow
+version: 0.0.1
diff --git a/object_detection/ks-app/components/model1.jsonnet b/object_detection/ks-app/components/model1.jsonnet
@@ -0,0 +1,21 @@
+local env = std.extVar("__ksonnet/environments");
+local params = std.extVar("__ksonnet/params").components["model1"];
+
+local k = import "k.libsonnet";
+
+// ksonnet appears to require name be a parameter of the prototype which is why we handle it differently.
+local name = params.name;
+
+// updatedParams includes the namespace from env by default.
+// We can override namespace in params if needed
+local updatedParams = env + params;
+
+local tfServingBase = import "kubeflow/tf-serving/tf-serving.libsonnet";
+local tfServing = tfServingBase {
+  // Override parameters with user supplied parameters.
+  params+: updatedParams {
+    name: name,
+  },
+};
+
+std.prune(k.core.v1.list.new(tfServing.components))
diff --git a/object_detection/ks-app/components/params.libsonnet b/object_detection/ks-app/components/params.libsonnet
@@ -0,0 +1,16 @@
+{
+  global: {},
+  components: {
+    // Component-level parameters, defined initially from 'ks prototype use ...'
+    // Each object below should correspond to a component in the components/ directory
+    model1: {
+      cloud: 'gcp',
+      deployHttpProxy: true,
+      gcpCredentialSecretName: 'user-sa',
+      modelPath: 'gs://kai-test2-models/object-detection',
+      modelServerImage: 'gcr.io/kubeflow-images-public/tensorflow-serving-1.8gpu:latest',
+      name: 'coco',
+      numGpus: 1,
+    },
+  },
+}
diff --git a/object_detection/ks-app/vendor/kubeflow/tf-serving/README.md b/object_detection/ks-app/vendor/kubeflow/tf-serving/README.md
@@ -0,0 +1,73 @@
+<!-- START doctoc generated TOC please keep comment here to allow auto update -->
+<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
+**Table of Contents**  *generated with [DocToc](https://github.com/thlorenz/doctoc)*
+
+- [tf-serving](#tf-serving)
+  - [Quickstart](#quickstart)
+  - [Using the library](#using-the-library)
+    - [io.ksonnet.pkg.tf-serving](#ioksonnetpkgtf-serving)
+      - [Example](#example)
+      - [Parameters](#parameters)
+
+<!-- END doctoc generated TOC please keep comment here to allow auto update -->
+
+# tf-serving
+
+> TensorFlow serving is a server for TensorFlow models.
+
+
+* [Quickstart](#quickstart)
+* [Using Prototypes](#using-prototypes)
+  * [io.ksonnet.pkg.tf-serving](#io.ksonnet.pkg.tf-serving)
+
+## Quickstart
+
+*The following commands use the `io.ksonnet.pkg.tf-serving` prototype to generate Kubernetes YAML for tf-serving, and then deploys it to your Kubernetes cluster.*
+
+First, create a cluster and install the ksonnet CLI (see root-level [README.md](rootReadme)).
+
+If you haven't yet created a [ksonnet application](linkToSomewhere), do so using `ks init <app-name>`.
+
+Finally, in the ksonnet application directory, run the following:
+
+```shell
+# Expand prototype as a Jsonnet file, place in a file in the
+# `components/` directory. (YAML and JSON are also available.)
+$ ks prototype use io.ksonnet.pkg.tf-serving tf-serving \
+  --name tf-serving \
+  --namespace default
+
+# Apply to server.
+$ ks apply -f tf-serving.jsonnet
+```
+
+## Using the library
+
+The library files for tf-serving define a set of relevant *parts* (_e.g._, deployments, services, secrets, and so on) that can be combined to configure tf-serving for a wide variety of scenarios. For example, a database like Redis may need a secret to hold the user password, or it may have no password if it's acting as a cache.
+
+This library provides a set of pre-fabricated "flavors" (or "distributions") of tf-serving, each of which is configured for a different use case. These are captured as ksonnet *prototypes*, which allow users to interactively customize these distributions for their specific needs.
+
+These prototypes, as well as how to use them, are enumerated below.
+
+### io.ksonnet.pkg.tf-serving
+
+TensorFlow serving
+#### Example
+
+```shell
+# Expand prototype as a Jsonnet file, place in a file in the
+# `components/` directory. (YAML and JSON are also available.)
+$ ks prototype use io.ksonnet.pkg.tf-serving tf-serving \
+  --name YOUR_NAME_HERE \
+  --model_path YOUR_MODEL_PATH_HERE
+```
+
+#### Parameters
+
+The available options to pass prototype are:
+
+* `--name=<name>`: Name to give to each of the components [string]
+* `--model_path=<model_path>`: Path to the model. This can be a GCS path. [string]
+
+
+[rootReadme]: https://github.com/ksonnet/mixins
diff --git a/object_detection/ks-app/vendor/kubeflow/tf-serving/parts.yaml b/object_detection/ks-app/vendor/kubeflow/tf-serving/parts.yaml
@@ -0,0 +1,35 @@
+{
+   "name": "tf-serving",
+   "apiVersion": "0.0.1",
+   "kind": "ksonnet.io/parts",
+   "description": "TensorFlow serving is a server for TensorFlow models.\n",
+   "author": "kubeflow team <kubeflow-team@google.com>",
+   "contributors": [
+      {
+         "name": "Jeremy Lewi",
+         "email": "jlewi@google.com"
+      }
+   ],
+   "repository": {
+      "type": "git",
+      "url": "https://github.com/kubeflow/kubeflow"
+   },
+   "bugs": {
+      "url": "https://github.com/kubeflow/kubeflow/issues"
+   },
+   "keywords": [
+      "kubeflow",
+      "tensorflow",
+      "database"
+   ],
+   "quickStart": {
+      "prototype": "io.ksonnet.pkg.tf-serving",
+      "componentName": "tf-serving",
+      "flags": {
+         "name": "tf-serving",
+         "namespace": "default"
+      },
+      "comment": "Run TensorFlow Serving"
+   },
+   "license": "Apache 2.0"
+}
diff --git a/...ct_detection/ks-app/vendor/kubeflow/tf-serving/prototypes/tf-serving-all-features.jsonnet b/...ct_detection/ks-app/vendor/kubeflow/tf-serving/prototypes/tf-serving-all-features.jsonnet
@@ -0,0 +1,24 @@
+// @apiVersion 0.1
+// @name io.ksonnet.pkg.tf-serving
+// @description TensorFlow serving
+// @shortDescription A TensorFlow serving deployment
+// @param name string Name to give to each of the components
+
+local k = import "k.libsonnet";
+
+// ksonnet appears to require name be a parameter of the prototype which is why we handle it differently.
+local name = import "param://name";
+
+// updatedParams includes the namespace from env by default.
+// We can override namespace in params if needed
+local updatedParams = env + params;
+
+local tfServingBase = import "kubeflow/tf-serving/tf-serving.libsonnet";
+local tfServing = tfServingBase {
+  // Override parameters with user supplied parameters.
+  params+: updatedParams {
+    name: name,
+  },
+};
+
+std.prune(k.core.v1.list.new(tfServing.components))