txt2img¶

Helm chart for deploying txt2img service.

Installing the Chart¶

To install the chart, run the following:

cd GenAIInfra/helm-charts/
./update_dependency.sh
helm dependency update txt2img
export HFTOKEN="insert-your-huggingface-token-here"
export MODELDIR="/mnt/opea-models"
# To run on Xeon
helm install txt2img txt2img --set global.HUGGINGFACEHUB_API_TOKEN=${HFTOKEN} --set global.modelUseHostPath=${MODELDIR}
# To run on Gaudi
#helm install txt2img txt2img --set global.HUGGINGFACEHUB_API_TOKEN=${HFTOKEN} --set global.modelUseHostPath=${MODELDIR} -f txt2img/gaudi-values.yaml

IMPORTANT NOTE¶

Make sure your MODELDIR exists on the node where your workload is schedueled so you can cache the downloaded model for next time use. Otherwise, set global.modelUseHostPath to ‘null’ if you don’t want to cache the model.

Verify¶

To verify the installation, run the command kubectl get pod to make sure all pods are running.

Curl command and UI are the two options that can be leveraged to verify the result.

Verify the workload through curl command¶

Then run the command kubectl port-forward svc/txt2img-text2image 9379:9379 to expose the service for access.

Open another terminal and run the following command to verify the service if working:

curl http://localhost:9379/v1/text2image \
    -XPOST \
    -d '{"prompt":"An astronaut riding a green horse", "num_images_per_prompt":1}' \
    -H 'Content-Type: application/json'

Verify the workload through UI¶

The UI has already been installed via the Helm chart. To access it, use the external IP of one your Kubernetes node along with the NGINX port. You can find the NGINX port using the following command:

export port=$(kubectl get service txt2img-nginx --output='jsonpath={.spec.ports[0].nodePort}')
echo $port

Open a browser to access http://<k8s-node-ip-address>:${port} to play with the txt2img workload.

Values¶

Key	Type	Default	Description
text2image.MODEL	string	`"stable-diffusion-v1-5/stable-diffusion-v1-5"`	Models id from https://huggingface.co/, or predownloaded model directory