The project provides a solution that helps to classify images by TensorFlow models.
It is an easy way to run web application locally on any device or to deploy it to cloud providers like Google Cloud in a few steps.
The solution includes docker images/containers that serve TensorFlow server and web application to quickly load data into the system, recognize data and provide information about that data based on output from machine learning models.
It is required to make next steps to start using ImageCLF:
- Clone or download an app repository
- Go into the folder where the app is cloned/downloaded
- Build a docker image by running the command ->
docker build . -t image_clf - Do something useful while the build is running
- After build is finished start a container to serve solution by running the command ->
docker run -p <your port for example 5000 or any free port on your system>:5000 -it --name image_clf_container image_clf - In opened container terminal run command ->
start_services - Go to
localhost:<your port for example 5000 or any free port on your system>in your browser - Enjoy usage
Define properties in properties file
Run setup.sh to deploy Spinnaker on GKE:
$ ~/setup.sh
We need Spinnaker’s API running on an endpoint that is publicly reachable. This is required to allow GitHub’s webhooks to reach Spinnaker.
First, we’ll start by creating LoadBalancer Services which will expose the API (Gate) and the UI (Deck) via a Load Balancer. We’ll do this by running the commands below and creating the spin-gate-public and spin-deck-public Services.
$ export NAMESPACE={namespace}
$ kubectl expose service -n ${NAMESPACE} spin-gate --type LoadBalancer \
--port 8084 \
--target-port 8084 \
--name spin-gate-public
$ kubectl expose service -n ${NAMESPACE} spin-deck --type LoadBalancer \
--port 9000 \
--target-port 9000 \
--name spin-deck-public
Once these Services have been created, we’ll need to update our Spinnaker deployment so that the UI understands where the API is located. To do this, we’ll use Halyard to override the base URL for both the API and the UI and then redeploy Spinnaker.
$ export NAMESPACE={namespace}
$ hal config security api edit --override-base-url http://<API_URL>:8084
$ hal config security ui edit --override-base-url http://<UI_URL>:9000
$ hal deploy apply
Use spin to create an app in Spinnaker.
spin application save --application-name test \
--owner-email [email protected] \
--cloud-providers kubernetes \
--gate-endpoint <your-gate-endpoint>
Run the following commands to upload an example pipeline to your Spinnaker instance
PROPERTIES_FILE="~/image_clf/properties"
source "$PROPERTIES_FILE"
sed -i s/PROJECT/$PROJECT/g sample.yaml
sed -i s/PROJECT/$PROJECT/g cloudbuild.yaml
sed s/PROJECT/$PROJECT/g pipeline.json > pipeline-final.json
sed s/SECRET/$SECRET/g pipeline.json > pipeline-final.json
spin pipeline save --gate-endpoint <your-gate-endpoint> -f pipeline-final.json