Deploying Image-to-Image Service

This document provides a comprehensive guide to deploying the Image-to-Image microservice pipeline on Intel platforms.

This guide covers two deployment methods:

🚀 1. Quick Start with Docker Compose

This method uses Docker Compose to start the service with a single command. It is the fastest and easiest way to get the service running.

1.1. Access the Code

Clone the repository and navigate to the deployment directory:

git clone https://github.com/opea-project/GenAIComps.git
cd GenAIComps/comps/image2image/deployment/docker_compose

1.2. Set Hugging Face Token

Set your Hugging Face token as an environment variable. This is required to download the Stable Diffusion model.

export HF_TOKEN=<your huggingface token>

1.3. Deploy the Service

Choose the command corresponding to your target platform.

  • For Intel® Xeon® CPU:

    docker compose -f compose.yaml up image2image -d

  • For Intel® Gaudi® 2 HPU:

    docker compose -f compose.yaml up image2image-gaudi -d

1.4. Validate the Service

Once the container is running, you can validate the service. Note: Run this command from the root of the GenAIComps repository.

export ip_address=$(hostname -I | awk '{print $1}')
curl http://${ip_address}:9389/v1/image2image -XPOST \
-d '{"image": "https://huggingface.co/datasets/patrickvonplaten/images/resolve/main/aa_xl/000000009.png", "prompt":"a photo of an astronaut riding a horse on mars", "num_images_per_prompt":1}' \
-H 'Content-Type: application/json'

The expected output will be a JSON object containing the generated image data.

1.5. Clean Up the Deployment

To stop and remove the containers, run the following command from the comps/image2image/deployment/docker_compose directory:

docker compose down

🚀 2. Manual Step-by-Step Deployment (Advanced)

This section provides detailed instructions for building the Docker image and running the microservice container individually.

2.1. Clone the Repository

If you haven’t already, clone the repository and navigate to the root directory:

git clone https://github.com/opea-project/GenAIComps.git
cd GenAIComps

2.2. Build the Docker Image

  • For Intel® Xeon® CPU:

    docker build -t opea/image2image:latest --build-arg https_proxy=$https_proxy --build-arg http_proxy=$http_proxy -f comps/image2image/src/Dockerfile .

  • For Intel® Gaudi® 2 HPU:

    docker build -t opea/image2image-gaudi:latest --build-arg https_proxy=$https_proxy --build-arg http_proxy=$http_proxy -f comps/image2image/src/Dockerfile.intel_hpu .

2.3. Configure Environment Variables

Set the necessary environment variables for the container.

  • For both Intel® Xeon® and Intel® Gaudi® 2 HPU:

    export ip_address=$(hostname -I | awk '{print $1}')
export HF_TOKEN=<your huggingface token>
# SDXL is an example, you can use other compatible models.
export MODEL=stabilityai/stable-diffusion-xl-refiner-1.0

2.4. Run the Microservice Container

2.4.1. Run Image-to-Image Microservice on Xeon

docker run -d --name "image2image-service" --ipc=host -p 9389:9389 -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e HF_TOKEN=$HF_TOKEN -e MODEL=$MODEL opea/image2image:latest

2.4.2. Run Image-to-Image Microservice on Gaudi

docker run -d --name "image2image-gaudi-service" -p 9389:9389 --runtime=habana -e HABANA_VISIBLE_DEVICES=all -e OMPI_MCA_btl_vader_single_copy_mechanism=none --cap-add=sys_nice --ipc=host -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e HF_TOKEN=$HF_TOKEN -e MODEL=$MODEL opea/image2image-gaudi:latest

2.5. Validate the Service

After starting the container, test the service endpoint. Make sure you are in the root directory of the GenAIComps repository.

# The ip_address variable should be set from step 2.3
curl http://${ip_address}:9389/v1/image2image -XPOST \
-d '{"image": "https://huggingface.co/datasets/patrickvonplaten/images/resolve/main/aa_xl/000000009.png", "prompt":"a photo of an astronaut riding a horse on mars", "num_images_per_prompt":1}' \
-H 'Content-Type: application/json'

You should see a successful response containing the generated image data.

2.6. Clean Up the Deployment

To stop and remove the container you started manually, use the docker stop and docker rm commands.

  • For Intel® Xeon® CPU:

    docker stop image2image-service
docker rm image2image-service

  • For Intel® Gaudi® 2 HPU:

    docker stop image2image-gaudi-service
docker rm image2image-gaudi-service