Merge pull request #6 from eleiton/add-stable-diffusion

Add stable diffusion with SD.Next

README.md

@@ -1,9 +1,8 @@
-# Run Ollama using your Intel Arc GPU
+# Run Ollama and Stable Diffusion with your Intel Arc GPU
 
-A Docker-based setup for running Ollama as a backend and Open WebUI as a frontend, leveraging Intel Arc Series GPUs on Linux systems.
+Effortlessly deploy a Docker-based solution that uses [Open WebUI](https://github.com/open-webui/open-webui) as your user-friendly 
-
+AI Interface, [Ollama](https://github.com/ollama/ollama) for integrating Large Language Models (LLM), and [SD.Next](https://github.com/vladmandic/sdnext) to 
-## Overview
+streamline Stable Diffusion capabilities, all while tapping into the power of Intel Arc Series GPUs on Linux systems by using [Intel® Extension for PyTorch](https://github.com/intel/intel-extension-for-pytorch).
-This repository provides a convenient way to run Ollama as a backend and Open WebUI as a frontend, allowing you to interact with Large Language Models (LLM) using an Intel Arc Series GPU on your Linux system.  
 
 ![screenshot](resources/open-webui.png)
 
@@ -18,10 +17,15 @@ This repository provides a convenient way to run Ollama as a backend and Open We
 2. Open WebUI  
    * The official distribution of Open WebUI.  
    * `WEBUI_AUTH` is turned off for authentication-free usage.  
-   * `ENABLE_OPENAI_API` and `ENABLE_OLLAMA_API` flags are set to off and on, respectively, allowing interactions via Ollama only.  
+   * `ENABLE_OPENAI_API` and `ENABLE_OLLAMA_API` flags are set to off and on, respectively, allowing interactions via Ollama only.
+   * `ENABLE_IMAGE_GENERATION` is set to true, allowing you to generate images from the UI.
+   * `IMAGE_GENERATION_ENGINE` is set to automatic1111 (SD.Next is compatible).
+3. SD.Next
+   * Uses as the base container the official [Intel® Extension for PyTorch](https://pytorch-extension.intel.com/installation?platform=gpu&version=v2.6.10%2Bxpu&os=linux%2Fwsl2&package=docker)
+   * Uses a customized version of the SD.Next [docker file](https://github.com/vladmandic/sdnext/blob/dev/configs/Dockerfile.ipex), making it compatible with the Intel Extension for Pytorch image.
 
 ## Setup
-Run the following commands to start your Ollama instance
+Run the following commands to start your AI instance
 ```bash
 $ git clone https://github.com/eleiton/ollama-intel-arc.git
 $ cd ollama-intel-arc
@@ -45,8 +49,23 @@ When using Open WebUI, you should see this partial output in your console, indic
 ```
 
 ## Usage
-* Open your web browser to http://localhost:3000 to access the Open WebUI web page.  
+* Open your web browser to http://localhost:7860 to access the SD.Next web page.
-* For more information on using Open WebUI, refer to the official documentation at https://docs.openwebui.com/ .
+* For the purposes of this demonstration, we'll use the [DreamShaper](https://civitai.com/models/4384/dreamshaper) model.
+* Follow these steps:
+* Download the  `dreamshaper_8` model by clicking on its image (1).
+* Wait for it to download (~2GB in size) and then select it in the dropbox (2).
+* (Optional) If you want to stay in the SD.Next UI, feel free to explore (3).
+![screenshot](resources/sd.next.png)
+* For more information on using SD.Next, refer to the official [documentation](https://vladmandic.github.io/sdnext-docs/).
+* Open your web browser to http://localhost:3000 to access the Open WebUI web page.
+* Go to the administrator [settings](http://localhost:3000/admin/settings) page.
+* Go to the Image section (1)
+* Make sure all settings look good, and validate them pressing the refresh button (2)
+* (Optional) Save any changes if you made them. (3)
+![screenshot](resources/open-webui-settings.png)
+* For more information on using Open WebUI, refer to the official [documentation](https://docs.openwebui.com/)
+* That's it, go back to Open WebUI main page and start chatting.  Make sure to select the `Image` button to indicate you want to generate Images.
+![screenshot](resources/open-webui-chat.png)
 
 ## Updating the containers
 If there are new updates in the [ipex-llm-inference-cpp-xpu](https://hub.docker.com/r/intelanalytics/ipex-llm-inference-cpp-xpu) docker Image or in the Open WebUI docker Image, you may want to update your containers, to stay up to date.

docker-compose.yml

@@ -32,9 +32,38 @@ services:
       - WEBUI_AUTH=False
       - ENABLE_OPENAI_API=False
       - ENABLE_OLLAMA_API=True
+      - ENABLE_IMAGE_GENERATION=True
+      - IMAGE_GENERATION_ENGINE=automatic1111
+      - IMAGE_GENERATION_MODEL=dreamshaper_8
+      - IMAGE_SIZE=400x400
+      - IMAGE_STEPS=30
+      - AUTOMATIC1111_BASE_URL=http://sdnext-ipex:7860/
+      - AUTOMATIC1111_CFG_SCALE=9
+      - AUTOMATIC1111_SAMPLER=DPM++ SDE
+      - AUTOMATIC1111_SCHEDULER=Karras
     extra_hosts:
       - host.docker.internal:host-gateway
     restart: unless-stopped
+
+  sdnext-ipex:
+    build:
+      context: sdnext
+      dockerfile: Dockerfile
+    image: sdnext-ipex:latest
+    container_name: sdnext-ipex
+    restart: unless-stopped
+    devices:
+      - /dev/dri:/dev/dri
+    ports:
+      - 7860:7860
+    volumes:
+      - sdnext-app-volume:/app
+      - sdnext-mnt-volume:/mnt
+      - sdnext-huggingface-volume:/root/.cache/huggingface
+
 volumes:
   ollama-volume: {}
   open-webui-volume: {}
+  sdnext-app-volume: {}
+  sdnext-mnt-volume: {}
+  sdnext-huggingface-volume: {}

sdnext/Dockerfile

@@ -0,0 +1,46 @@
+FROM intel/intel-extension-for-pytorch:2.6.10-xpu
+
+# essentials
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends --fix-missing \
+    software-properties-common \
+    build-essential \
+    ca-certificates \
+    wget \
+    gpg \
+    git
+
+# python3.10
+RUN apt-get install -y --no-install-recommends --fix-missing python3.10-venv
+
+# jemalloc is not required but it is highly recommended (also used with optional ipexrun)
+RUN apt-get install -y --no-install-recommends --fix-missing libjemalloc-dev
+ENV LD_PRELOAD=libjemalloc.so.2
+
+# cleanup
+RUN /usr/sbin/ldconfig
+RUN apt-get clean && rm -rf /var/lib/apt/lists/*
+
+# stop pip and uv from caching
+ENV PIP_NO_CACHE_DIR=true
+ENV UV_NO_CACHE=true
+
+# set paths to use with sdnext
+ENV SD_DOCKER=true
+ENV SD_DATADIR="/mnt/data"
+ENV SD_MODELSDIR="/mnt/models"
+ENV venv_dir="/mnt/python/venv"
+
+# git clone and start sdnext
+RUN echo '#!/bin/bash\ngit status || git clone https://github.com/vladmandic/sdnext.git .\n/app/webui.sh "$@"' | tee /bin/startup.sh
+RUN chmod 755 /bin/startup.sh
+
+# run sdnext
+WORKDIR /app
+ENTRYPOINT [ "startup.sh", "-f", "--use-ipex", "--uv", "--listen", "--debug", "--api-log", "--log", "sdnext.log" ]
+
+# expose port
+EXPOSE 7860
+
+# stop signal
+STOPSIGNAL SIGINT