|

Docs

Image Generation WebUI Forge

This guide explains how to install, configure, and run Stable Diffusion WebUI Forge for use with our image generation application.

System Requirements

- Operating System: Windows 10/11 (64-bit), Linux, or macOS

- GPU: NVIDIA (8GB+ VRAM recommended) or AMD (experimental support)

- RAM: Minimum 16GB (32GB recommended)

- Disk Space: Minimum 20GB free (more space for additional models)

- Software: Python 3.10, Git

Step-by-Step Installation
1. Download Stable Diffusion WebUI Forge
  1. Open Command Prompt (CMD) or PowerShell.
  2. Navigate to the folder where you want to install Forge.
  3. Run the following commands:
git clone https://github.com/lllyasviel/stable-diffusion-webui-forge.gitcd stable-diffusion-webui-forge
2. Configure the webui-user.bat file
  1. Inside the folder, locate the `webui-user.bat` file. It may be inside the `webui` folder.
  2. Open it with a text editor (Notepad, VS Code, etc.).
  3. Replace all content with the following - note that all we're doing is updating `COMMANDLINE_ARGS`:
@echo off

set PYTHON=
set GIT=
set VENV_DIR=
set COMMANDLINE_ARGS=--api --cors-allow-origins=* --listen --server-name 0.0.0.0 --port 7860 --disable-safe-unpickle

call webui.bat

4. Save and close the file.

3. Download Image Generation Models

Forge does not come with pre-installed models. You need to download at least one model to start generating images.

Option 1: Download models manually

  1. Create a folder named `webui/models/Stable-diffusion` if it doesn't exist.
  2. Download models from trusted sites such as Civitai or Hugging Face.
  3. Place the downloaded `.safetensors` or `.ckpt` files in the `models/Stable-diffusion` folder.

Option 2: Use Forge's model manager

  1. Start Forge by following step 4 below.
  2. In the web interface, go to the "Models" tab.
  3. Click on "Download Models" and choose from the available models.
4. Run Forge
  1. In the Forge folder, double-click the `webui-user.bat` file.
  2. Wait for the dependencies to install (this may take a few minutes on first run).
  3. The browser will automatically open with the Forge interface.
  4. If the browser doesn't open, manually access: `http://127.0.0.1:7860`.
NOTE: Right now there is a bug in Forge - you can watch this issue to see how to fix it: https://github.com/lllyasviel/stable-diffusion-webui-forge/issues/3001
5. Verify that the API is working

To verify that the API is working correctly:

  1. With Forge running, access: `http://127.0.0.1:7860/docs`.
  2. You should see the Forge API documentation.
6. Using with HammerAI
  1. Select Forge under Image Models when chatting with a character or on the `Create Images` page.
  2. In the Image Mode settings, set the Forge address as: `http://127.0.0.1:7860`.
  3. Now you can use our application to generate images through Forge.
Troubleshooting Common Issues

Forge doesn't start

  • Check if you have Python 3.10 installed.
  • Try running `webui.bat` directly to see error messages.
  • Verify that you have enough disk space.

"CUDA out of memory" error

  • Reduce the image resolution (512x512 is a good starting point).
  • Use the `--medvram` or `--lowvram` option by adding it to `COMMANDLINE_ARGS` in the `webui-user.bat` file.

API is not working

  • Check if Forge was started with the `--api` flag (already included in the provided webui-user.bat).
  • Verify that the address and port are correct in the application.

Low quality images

  • Download better quality models.
  • Increase the number of steps (20-30 is a good value).
  • Adjust the CFG Scale (7-9 usually produces good results).
Important Notes
  • Forge must be running whenever you want to use our application to generate images.
  • The first generation after starting Forge may be slower (model loading).
  • - The quality and style of the images depend on the model you downloaded and are using.
Minimum Hardware Requirements

For a basic experience:

  • NVIDIA GPU with 4GB+ VRAM
  • 8GB RAM
  • Quad-core CPU

For an optimal experience:

  • NVIDIA GPU with 8GB+ VRAM (RTX 2070 or higher)
  • 16GB+ RAM
  • Recent multi-core CPU
  • SSD with at least 20GB free space