Example Edge Craft Retrieval-Augmented Generation Deployment on Intel® Arc® Platform

This document outlines the deployment process for Edge Craft Retrieval-Augmented Generation service on Intel Arc server. This example includes the following sections:

• EdgeCraftRAG Quick Start Deployment: Demonstrates how to quickly deploy a Edge Craft Retrieval-Augmented Generation service/pipeline on Intel® Arc® platform.

• EdgeCraftRAG Docker Compose Files: Describes some example deployments and their docker compose files.

• EdgeCraftRAG Service Configuration: Describes the service and possible configuration changes.

EdgeCraftRAG Quick Start Deployment

This section describes how to quickly deploy and test the EdgeCraftRAG service manually on Intel® Arc® platform. The basic steps are:

1. Prerequisites

2. Access the Code

3. Prepare models

4. Prepare env variables and configurations

5. Configure the Deployment Environment

6. Deploy the Service Using Docker Compose

7. Access UI

8. Cleanup the Deployment

Prerequisites

EC-RAG supports vLLM deployment(default method) and local OpenVINO deployment for Intel Arc GPU. Prerequisites are shown as below:
Hardware: Intel Arc A770
OS: Ubuntu Server 22.04.1 or newer (at least 6.2 LTS kernel)
Driver & libraries: please to Installing GPUs Drivers for detailed driver & libraries setup

Below steps are based on vLLM as inference engine, if you want to choose OpenVINO, please refer to OpenVINO Local Inference

Access the Code

Clone the GenAIExample repository and access the EdgeCraftRAG Intel® Arc® platform Docker Compose files and supporting scripts:

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

Checkout a released version, such as v1.3:

git checkout v1.3

Prepare models

# Prepare models for embedding, reranking:
export MODEL_PATH="${PWD}/models" # Your model path for embedding, reranking and LLM models
mkdir -p $MODEL_PATH
pip install --upgrade --upgrade-strategy eager "optimum[openvino]"
optimum-cli export openvino -m BAAI/bge-small-en-v1.5 ${MODEL_PATH}/BAAI/bge-small-en-v1.5 --task sentence-similarity
optimum-cli export openvino -m BAAI/bge-reranker-large ${MODEL_PATH}/BAAI/bge-reranker-large --task text-classification

# Prepare LLM model
export LLM_MODEL="Qwen/Qwen3-8B" # Your model id
pip install modelscope
modelscope download --model $LLM_MODEL --local_dir "${MODEL_PATH}/${LLM_MODEL}"
# Optionally, you can also download models with huggingface:
# pip install -U huggingface_hub
# huggingface-cli download $LLM_MODEL --local-dir "${MODEL_PATH}/${LLM_MODEL}"

Prepare env variables and configurations

Below steps are for single Intel Arc GPU inference, if you want to setup multi Intel Arc GPUs inference, please refer to Multi-ARC Setup

Prepare env variables for vLLM deployment

ip_address=$(hostname -I | awk '{print $1}')
# Use `ip a` to check your active ip
export HOST_IP=$ip_address # Your host ip

# Check group id of video and render
export VIDEOGROUPID=$(getent group video | cut -d: -f3)
export RENDERGROUPID=$(getent group render | cut -d: -f3)

# If you have a proxy configured, uncomment below line
# export no_proxy=${no_proxy},${HOST_IP},edgecraftrag,edgecraftrag-server
# export NO_PROXY=${NO_PROXY},${HOST_IP},edgecraftrag,edgecraftrag-server
# If you have a HF mirror configured, it will be imported to the container
# export HF_ENDPOINT=https://hf-mirror.com # your HF mirror endpoint"

# Make sure all 3 folders have 1000:1000 permission, otherwise
# chown 1000:1000 ${MODEL_PATH} ${PWD} # the default value of DOC_PATH and TMPFILE_PATH is PWD ,so here we give permission to ${PWD}
# In addition, also make sure the .cache folder has 1000:1000 permission, otherwise
# chown 1000:1000 -R $HOME/.cache

For more advanced env variables and configurations, please refer to Prepare env variables for vLLM deployment

Generate nginx config file

export VLLM_SERVICE_PORT_0=8100 # You can set your own port for vllm service
# Generate your nginx config file
# nginx-conf-generator.sh requires 2 parameters: DP_NUM and output filepath
bash nginx/nginx-conf-generator.sh 1 nginx/nginx.conf
# set NGINX_CONFIG_PATH
export NGINX_CONFIG_PATH="${PWD}/nginx/nginx.conf"

Deploy the Service Using Docker Compose

# EC-RAG support Milvus as persistent database, by default milvus is disabled, you can choose to set MILVUS_ENABLED=1 to enable it
export MILVUS_ENABLED=0
# If you enable Milvus, the default storage path is PWD, uncomment if you want to change:
# export DOCKER_VOLUME_DIRECTORY= # change to your preference

# EC-RAG support chat history round setting, by default chat history is disabled, you can set CHAT_HISTORY_ROUND to control it
# export CHAT_HISTORY_ROUND= # change to your preference

# Launch EC-RAG service with compose
docker compose -f docker_compose/intel/gpu/arc/compose_vllm.yaml up -d

Access UI

Open your browser, access http://${HOST_IP}:8082

Your browser should be running on the same host of your console, otherwise you will need to access UI with your host domain name instead of ${HOST_IP}.

Below is the UI front page, for detailed operations on UI and EC-RAG settings, please refer to Explore_Edge_Craft_RAG

Cleanup the Deployment

To stop the containers associated with the deployment, execute the following command:

docker compose -f docker_compose/intel/gpu/arc/compose_vllm.yaml down

All the EdgeCraftRAG containers will be stopped and then removed on completion of the “down” command.

EdgeCraftRAG Docker Compose Files

The compose.yaml is default compose file using tgi as serving framework

Service Name	Image Name
etcd	quay.io/coreos/etcd:v3.5.5
minio	minio/minio:RELEASE.2023-03-20T20-16-18Z
milvus-standalone	milvusdb/milvus:v2.4.6
edgecraftrag-server	opea/edgecraftrag-server:latest
edgecraftrag-ui	opea/edgecraftrag-ui:latest
ecrag	opea/edgecraftrag:latest

EdgeCraftRAG Service Configuration

The table provides a comprehensive overview of the EdgeCraftRAG service utilized across various deployments as illustrated in the example Docker Compose files. Each row in the table represents a distinct service, detailing its possible images used to enable it and a concise description of its function within the deployment architecture.

Service Name	Possible Image Names	Optional	Description
etcd	quay.io/coreos/etcd:v3.5.5	No	Provides distributed key-value storage for service discovery and configuration management.
minio	minio/minio:RELEASE.2023-03-20T20-16-18Z	No	Provides object storage services for storing documents and model files.
milvus-standalone	milvusdb/milvus:v2.4.6	No	Provides vector database capabilities for managing embeddings and similarity search.
edgecraftrag-server	opea/edgecraftrag-server:latest	No	Serves as the backend for the EdgeCraftRAG service, with variations depending on the deployment.
edgecraftrag-ui	opea/edgecraftrag-ui:latest	No	Provides the user interface for the EdgeCraftRAG service.
ecrag	opea/edgecraftrag:latest	No	Acts as a reverse proxy, managing traffic between the UI and backend services.