feat/stablediffusion-webui-modelcache

.github/workflows/stable-diffusion-web-ui-production.yml

@@ -0,0 +1,45 @@
+name: Build and Push Stable Diffusion WebUI Docker Image
+
+on:
+  workflow_dispatch:
+  push:
+    paths:
+      - "containers/official-templates/stable-diffusion-webui/**"
+    branches-ignore:
+      - main
+
+jobs:
+  build:
+    name: Build and Push Docker Image
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+
+      - name: Clear space to remove unused folders
+        run: |
+          rm -rf /usr/share/dotnet
+          rm -rf /opt/ghc
+          rm -rf "/usr/local/share/boost"
+          rm -rf "$AGENT_TOOLSDIRECTORY"
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      # Set up Docker Buildx
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      # Log in to Docker Hub
+      - name: Log in to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      - name: Build and Push Docker Image
+        working-directory: containers/official-templates/stable-diffusion-webui
+        uses: docker/bake-action@v2
+        with:
+          push: true

official-templates/stable-diffusion-webui/Dockerfile

@@ -1,9 +1,5 @@
 FROM nvidia/cuda:11.8.0-devel-ubuntu22.04
 
-ARG WEBUI_VERSION
-ARG TORCH_VERSION
-ARG XFORMERS_VERSION
-
 ENV DEBIAN_FRONTEND noninteractive
 ENV SHELL=/bin/bash
 ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/lib/x86_64-linux-gnu
@@ -15,13 +11,18 @@ WORKDIR /workspace
 SHELL ["/bin/bash", "-o", "pipefail", "-c"]
 
 # Install system dependencies
-RUN apt update --yes && \
-    apt upgrade --yes && \
-    apt install --yes --no-install-recommends \
+RUN rm -f /etc/apt/sources.list.d/*.list && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/* && \
+    apt-get update --yes && \
+    apt-get upgrade --yes && \
+    apt-get install --yes --no-install-recommends \
     git openssh-server libglib2.0-0 libsm6 libgl1 libxrender1 libxext6 ffmpeg wget curl psmisc rsync vim bc nginx \
     pkg-config libffi-dev libcairo2 libcairo2-dev libgoogle-perftools4 libtcmalloc-minimal4 apt-transport-https \
     software-properties-common ca-certificates && \
-    update-ca-certificates
+    update-ca-certificates && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
 
 # Install Python 3.10
 RUN add-apt-repository ppa:deadsnakes/ppa && \
@@ -32,51 +33,47 @@ RUN add-apt-repository ppa:deadsnakes/ppa && \
     curl https://bootstrap.pypa.io/get-pip.py -o get-pip.py && \python get-pip.py && \
     pip install -U --no-cache-dir pip
 
-# Cache Models
-RUN mkdir /sd-models && mkdir /cn-models && \
-    wget https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned.ckpt -O /sd-models/v1-5-pruned.ckpt && \
-    wget https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/resolve/main/sd_xl_base_1.0.safetensors -O /sd-models/sd_xl_base_1.0.safetensors && \
-    wget https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/resolve/main/sd_xl_refiner_1.0.safetensors -O /sd-models/sd_xl_refiner_1.0.safetensors && \
-    wget https://huggingface.co/lllyasviel/ControlNet-v1-1/resolve/main/control_v11p_sd15_canny.pth -O /cn-models/control_v11p_sd15_canny.pth
-
 # Create a virtual environment
 RUN python -m venv /workspace/venv && \
     pip install -U --no-cache-dir jupyterlab jupyterlab_widgets ipykernel ipywidgets
 
+WORKDIR /
+
+ARG WEBUI_VERSION
+
 # Install Automatic1111's WebUI
 RUN git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git && \
-    cd stable-diffusion-webui && \
+    cd /stable-diffusion-webui && \
     git checkout tags/${WEBUI_VERSION} && \
-    pip install -r requirements_versions.txt && \
+    pip install httpx>=0.25.0 xformers && \
+    pip install --no-cache-dir -r requirements_versions.txt && \
     python -c "from launch import prepare_environment; prepare_environment()" --skip-torch-cuda-test
 
-# Install torch and xformers
-RUN cd /workspace/stable-diffusion-webui && \
-    pip install torch===${TORCH_VERSION} torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 && \
-    pip install xformers==${XFORMERS_VERSION} --index-url https://download.pytorch.org/whl/cu118
+WORKDIR /stable-diffusion-webui
+
+# Apply hotfix to replace `pytorch_lightning.utilities.distributed` with `pytorch_lightning.utilities.rank_zero`
+# see https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/11458#issuecomment-1609900319
+# RUN sed -i 's/pytorch_lightning.utilities.distributed/pytorch_lightning.utilities.rank_zero/g' \
+#     repositories/stable-diffusion-stability-ai/ldm/models/diffusion/ddpm.py && \
+#     sed -i 's/pytorch_lightning.utilities.distributed/pytorch_lightning.utilities.rank_zero/g' \
+#     extensions-builtin/LDSR/sd_hijack_ddpm_v1.py
 
 # Install DeForum
-RUN cd /workspace/stable-diffusion-webui && \
-    git clone https://github.com/deforum-art/sd-webui-deforum extensions/deforum && \
+RUN git clone https://github.com/deforum-art/sd-webui-deforum extensions/deforum && \
     cd extensions/deforum && \
     pip install basicsr && \
     pip install -r requirements.txt
 
 # Install ControlNet
-RUN cd /workspace/stable-diffusion-webui && \
-    git clone https://github.com/Mikubill/sd-webui-controlnet.git extensions/sd-webui-controlnet && \
+RUN git clone https://github.com/Mikubill/sd-webui-controlnet.git extensions/sd-webui-controlnet && \
     cd extensions/sd-webui-controlnet && \
+    pip install opencv-python-headless>=4.9.0 && \
     pip install -r requirements.txt
 
-COPY cache-sd-model.py /workspace/stable-diffusion-webui/
-RUN cd /workspace/stable-diffusion-webui/ && \
-    python cache-sd-model.py --use-cpu=all --ckpt /sd-models/SDv1-5.ckpt
+WORKDIR / 
 
 RUN mv /workspace/venv /venv && \
-    mv /workspace/stable-diffusion-webui /stable-diffusion-webui && \
-    mkdir /workspace/downloader && git clone https://github.com/jjangga0214/sd-models-downloader.git /workspace/downloader
-
-COPY relauncher.py webui-user.sh /stable-diffusion-webui/
+    mkdir -p /workspace/downloader && git clone https://github.com/jjangga0214/sd-models-downloader.git /workspace/downloader
 
 # NGINX Proxy
 COPY --from=proxy nginx.conf /etc/nginx/nginx.conf
@@ -85,6 +82,12 @@ COPY --from=proxy readme.html /usr/share/nginx/html/readme.html
 # Copy the README.md
 COPY README.md /usr/share/nginx/html/README.md
 
+WORKDIR /stable-diffusion-webui
+
+COPY relauncher.py webui-user.sh ./
+
+WORKDIR /workspace
+
 # Start Scripts
 COPY pre_start.sh /pre_start.sh
 COPY --from=scripts start.sh /

official-templates/stable-diffusion-webui/README.md

@@ -1,29 +1,75 @@
-## 🚀 RunPod Automatic1111 Stable Diffusion Template
+# Automatic1111 Stable Diffusion WebUI Template
 
-### 📝 General
+A ready-to-use template for running Stable Diffusion WebUI (AUTOMATIC1111) on RunPod.
 
-⚠️ **Please note, this doesn't work out-of-the-box with encrypted volumes!**
+## Quick Start
 
-This is a RunPod packaged template for stable diffusion using the Automatic1111 repository. RunPod does not maintain the code for this repo, we merely package it for easier use.
+1. Wait for the pod to fully initialize (GPU Utilization should be 0%)
+2. Access the WebUI through port 3000
+3. Start creating!
 
-If you need help with settings, etc., feel free to ask us, but remember we're not stable diffusion experts! 😅 We'll do our best to assist, but the RunPod community or automatic/stable diffusion communities might be more effective in helping you.
+⚠️ **Note**: You may encounter a 502 error if you try to connect before initialization is complete.
 
-🔵 **Please wait until the GPU Utilization % is 0 before attempting to connect. You'll likely encounter a 502 error if the pod is still getting ready for use.**
+## Pre-installed Models
 
-### ⚙️ Changing Launch Parameters
+The following models are automatically mounted and ready to use:
 
-You might be used to altering a different file for your launch parameters. In our case, we use `relauncher.py` located in the `webui` directory to manage the launch flags like `--xformers`. Feel free to edit this file, and then restart your pod via the hamburger menu for the changes to take effect. `--xformers` and `--api` are commonly inquired about.
+### Stable Diffusion Models
 
-### 📥 Using Your Own Models
+- [Stable Diffusion XL Base 1.0](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/blob/main/sd_xl_base_1.0.safetensors)
+- [Stable Diffusion XL Refiner 1.0](https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/blob/main/sd_xl_refiner_1.0.safetensors)
+- [Stable Diffusion 1.5](https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/blob/main/v1-5-pruned-emaonly.safetensors)
 
-The best ways to introduce your models to your pod is by using [runpodctl](https://github.com/runpod/runpodctl/blob/main/README.md) or by uploading them to Google Drive or another cloud storage provider and downloading them to your pod from there.
+### ControlNet Models
 
-### 🚚 Uploading to Google Drive
+- Canny (control_v11p_sd15_canny.pth)
 
-If you're finished with the pod and want to transfer things to Google Drive, [this colab](https://colab.research.google.com/drive/1ot8pODgystx1D6_zvsALDSvjACBF1cj6) can assist you using `runpodctl`. You can run `runpodctl` in a web terminal (found in the pod connect menu), or in a terminal on the desktop.
+## Network Ports
 
-## 🔌 Template Ports
+| Application            | Port | Protocol | Description                         |
+| ---------------------- | ---- | -------- | ----------------------------------- |
+| Stable Diffusion WebUI | 3000 | HTTP     | Main interface for Stable Diffusion |
+| Jupyter Lab            | 8888 | HTTP     | Python notebook interface           |
+| SSH                    | 22   | TCP      | Secure shell access                 |
 
-- **3001** | HTTP - This is the WebUI port that gets proxied to the internal 3000 port.
-- **8888** | HTTP - This is the JupyterLab port that gets proxied to the internal 8888 port.
-- **22** | TCP  - This is the SSH port that gets proxied to the internal 22 port.
+## Customization
+
+### Modifying Launch Parameters
+
+Launch parameters are configured in `webui-user.sh`. To modify them:
+
+1. Edit the file in the workspace
+2. Restart the pod to apply changes
+
+Current default parameters include:
+
+- `--xformers` for optimized memory usage
+- `--listen` for network access
+- `--enable-insecure-extension-access` for extension support
+
+## Adding Custom Models
+
+You have two options for adding your own models:
+
+1. Using [runpodctl](https://github.com/runpod/runpodctl/blob/main/README.md)
+2. Downloading from cloud storage (Google Drive, etc.)
+
+## Backing Up Your Work
+
+To save your work to Google Drive:
+
+1. Use [this Google Colab notebook](https://colab.research.google.com/drive/1ot8pODgystx1D6_zvsALDSvjACBF1cj6)
+2. Follow the instructions to transfer files using `runpodctl`
+
+You can run `runpodctl` either through:
+
+- The web terminal (in pod connect menu)
+- The desktop terminal
+
+## Important Notes
+
+- This template doesn't support encrypted volumes
+- For technical support, consider:
+  - [RunPod Community on Discord](https://discord.gg/cUpRmau42V)
+  - [Automatic1111 GitHub Issues](https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues)
+  - [Stable Diffusion on Reddit](https://www.reddit.com/r/StableDiffusion/)

official-templates/stable-diffusion-webui/docker-bake.hcl

@@ -1,5 +1,5 @@
 variable "RELEASE" {
-    default = "12.0.0"
+    default = "13.0.0"
 }
 
 target "default" {
@@ -10,8 +10,7 @@ target "default" {
     proxy = "../../container-template/proxy"
   }
   args = {
-    WEBUI_VERSION = "v1.8.0"
-    TORCH_VERSION = "2.1.2+cu118"
-    XFORMERS_VERSION = "0.0.23.post1+cu118"
+    WEBUI_VERSION = "v1.10.0"
   }
+  platforms = ["linux/amd64"]
 }

official-templates/stable-diffusion-webui/pre_start.sh

@@ -3,13 +3,26 @@ rsync --remove-source-files -rlptDu --ignore-existing /venv/ /workspace/venv/
 
 echo "**** syncing stable diffusion to workspace, please wait ****"
 rsync --remove-source-files -rlptDu --ignore-existing /stable-diffusion-webui/ /workspace/stable-diffusion-webui/
-ln -s /sd-models/* /workspace/stable-diffusion-webui/models/Stable-diffusion/
-ln -s /cn-models/* /workspace/stable-diffusion-webui/extensions/sd-webui-controlnet/models/
 
-if [[ $RUNPOD_STOP_AUTO ]]
-then
-    echo "Skipping auto-start of webui"
-else
+# Create symbolic links for the models
+mkdir -p /workspace/stable-diffusion-webui/models/Stable-diffusion/
+ln -sf /runpod/cache/model/Comfy-Org/stable-diffusion-v1-5-archive/main/v1-5-pruned-emaonly.safetensors /workspace/stable-diffusion-webui/models/Stable-diffusion/
+ln -sf /runpod/cache/model/stabilityai/stable-diffusion-xl-base-1.0/main/sd_xl_base_1.0.safetensors /workspace/stable-diffusion-webui/models/Stable-diffusion/
+ln -sf /runpod/cache/model/stabilityai/stable-diffusion-xl-refiner-1.0/main/sd_xl_refiner_1.0.safetensors /workspace/stable-diffusion-webui/models/Stable-diffusion/
+ln -sf /runpod/cache/model/stabilityai/stable-diffusion-3-medium/main/sd3_medium.safetensors /workspace/stable-diffusion-webui/models/Stable-diffusion/
+ln -sf /runpod/cache/model/stabilityai/stable-diffusion-3-medium/main/sd3_medium_incl_clips.safetensors /workspace/stable-diffusion-webui/models/Stable-diffusion/
+
+# Create symbolic links for clips
+mkdir -p /workspace/stable-diffusion-webui/models/Clip/
+ln -sf /runpod/cache/model/stabilityai/stable-diffusion-3-medium/main/text_encoders/clip_g.safetensors /workspace/stable-diffusion-webui/models/Clip/
+ln -sf /runpod/cache/model/stabilityai/stable-diffusion-3-medium/main/text_encoders/clip_l.safetensors /workspace/stable-diffusion-webui/models/Clip/
+ln -sf /runpod/cache/model/stabilityai/stable-diffusion-3-medium/main/text_encoders/t5xxl_fp8_e4m3fn.safetensors /workspace/stable-diffusion-webui/models/Clip/
+
+# Create symbolic link for ControlNet model
+mkdir -p /workspace/stable-diffusion-webui/extensions/sd-webui-controlnet/models/
+ln -sf /runpod/cache/model/lllyasviel/ControlNet-v1-1/main/control_v11p_sd15_canny.pth /workspace/stable-diffusion-webui/extensions/sd-webui-controlnet/models/
+
+if [[ ! $RUNPOD_STOP_AUTO ]]; then
     echo "Started webui through relauncher script"
     cd /workspace/stable-diffusion-webui
     python relauncher.py &

official-templates/stable-diffusion-webui/relauncher.py

@@ -1,25 +1,18 @@
-import os
 import time
+import subprocess
 
 
-def relaunch_process(launch_counter=0):
-    '''
-
-    '''
+def relaunch_process():
+    """Relaunches the web UI process in a loop"""
     while True:
-        print('Relauncher: Launching...')
-        if launch_counter > 0:
-            print(f'\tRelaunch count: {launch_counter}')
-
+        print("Relauncher: Launching...")
         try:
-            launch_string = "/workspace/stable-diffusion-webui/webui.sh -f"
-            os.system(launch_string)
+            webui_path = "/workspace/stable-diffusion-webui/webui.sh"
+            subprocess.run(f"bash {webui_path} -f", shell=True, check=True)
         except Exception as err:
             print(f"An error occurred: {err}")
-        finally:
-            print('Relauncher: Process is ending. Relaunching in 2s...')
-            launch_counter += 1
-            time.sleep(2)
+        print("Relauncher: Process is ending. Relaunching in 2s...")
+        time.sleep(2)
 
 
 if __name__ == "__main__":

official-templates/stable-diffusion-webui/webui-user.sh

@@ -1,4 +1,4 @@
-# #!/bin/bash
+#!/bin/bash
 #########################################################
 # Uncomment and change the variables below to your need:#
 #########################################################
@@ -10,7 +10,7 @@ install_dir="/workspace"
 #clone_dir="stable-diffusion-webui"
 
 # Commandline arguments for webui.py, for example: export COMMANDLINE_ARGS="--medvram --opt-split-attention"
-export COMMANDLINE_ARGS="--port 3000 --xformers --skip-install --listen --enable-insecure-extension-access"
+export COMMANDLINE_ARGS="--port 3000 --xformers --skip-install --listen --enable-insecure-extension-access --skip-torch-cuda-test"
 #export XFORMERS_PACKAGE="xformers==0.0.17.dev447"
 
 # python3 executable