Stable Diffusion WebUI Forge - Classic
[ Classic | Neo ]
Stable Diffusion WebUI Forge is a platform on top of the original Stable Diffusion WebUI by AUTOMATIC1111, to make development easier, optimize resource management, speed up inference, and study experimental features.
The name "Forge" is inspired by "Minecraft Forge". This project aims to become the Forge of Stable Diffusion WebUI.
- lllyasviel
(paraphrased)
"**Classic**" mainly serves as an archive for the "`previous`" version of Forge, which was built on [Gradio](https://github.com/gradio-app/gradio) `3.41.2` before the major changes *(see the original [announcement](https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/801))* were introduced. Additionally, this fork is focused exclusively on **SD1** and **SDXL** checkpoints, having various optimizations implemented, with the main goal of being the lightest WebUI without any bloatwares.
> [!Tip]
> [How to Install](#installation)
## Features [Aug. 13]
> Most base features of the original [Automatic1111 Webui](https://github.com/AUTOMATIC1111/stable-diffusion-webui) should still function
#### New Features
- [X] Support [uv](https://github.com/astral-sh/uv) package manager
- requires **manually** installing [uv](https://github.com/astral-sh/uv/releases)
- drastically speed up installation
- see [Commandline](#by-classic)
- [X] Support [SageAttention](https://github.com/thu-ml/SageAttention)
- requires **manually** installing [triton](https://github.com/triton-lang/triton)
- [how to install](#install-triton)
- requires RTX **30** +
- ~10% speed up for SDXL
- see [Commandline](#by-classic)
- [X] Support [FlashAttention](https://arxiv.org/abs/2205.14135)
- requires **manually** installing [flash-attn](https://github.com/Dao-AILab/flash-attention)
- [how to install](#install-flash-attn)
- ~10% speed up
- [X] Support fast `fp16_accumulation`
- requires PyTorch **2.7.0** +
- ~25% speed up
- see [Commandline](#by-classic)
- [X] Support fast `cublas` operation *(`CublasLinear`)*
- requires **manually** installing [cublas_ops](https://github.com/aredden/torch-cublas-hgemm)
- [how to install](#install-cublas)
- ~25% speed up
- enable in **Settings/Optimizations**
- [X] Support fast `fp8` operation *(`torch._scaled_mm`)*
- requires RTX **40** +
- requires **UNet Weights in fp8** option
- ~10% speed up; reduce quality
- enable in **Settings/Optimizations**
> [!Note]
> - Both `fp16_accumulation` and `cublas_ops` achieve the same speed up; if you already install/update to PyTorch **2.7.0**, you do not need to go for `cublas_ops`
> - The `fp16_accumulation` and `cublas_ops` require `fp16` precision, thus is not compatible with the `fp8` operation
- [X] Persistent LoRA Patching
- speed up LoRA loading in subsequent generations
- see [Commandline](#by-classic)
- [X] Patch LoRA in-place
- reduce VRAM usage when loading LoRA
- enable in **Settings/Extra Networks**
- [X] Implement new Samplers
- *(ported from reForge Webui)*
- [X] Implement Scheduler dropdown
- *(backported from Automatic1111 Webui upstream)*
- enable in **Settings/UI Alternatives**
- [X] Add `CFG` slider to the `Hires. fix` section
- [X] Implement RescaleCFG
- reduce burnt colors; mainly for `v-pred` checkpoints
- enable in **Settings/UI Alternatives**
- [X] Implement MaHiRo
- alternative CFG calculation; improve prompt adherence
- enable in **Settings/UI Alternatives**
- [X] Implement full precision calculation for `Mask blur` blending
- enable in **Settings/img2img**
- [X] Support loading upscalers in `half` precision
- speed up; reduce quality
- enable in **Settings/Upscaling**
- [X] Support running tile composition on GPU
- enable in **Settings/Upscaling**
- [X] Allow `newline` in LoRA metadata
- *(backported from Automatic1111 Webui upstream)*
- [X] Implement sending parameters from generation result rather than from UI
- **e.g.** send the prompts instead of `Wildcard` syntax
- enable in **Settings/Infotext**
- [X] Implement tiling optimization for VAE
- reduce memory usage; reduce speed
- enable in **Settings/VAE**
- [X] Implement `diskcache` for hashes
- *(backported from Automatic1111 Webui upstream)*
- [X] Implement `skip_early_cond`
- *(backported from Automatic1111 Webui upstream)*
- enable in **Settings/Optimizations**
- [X] Allow inserting the upscaled image to the Gallery instead of overriding the input image
- *(backported from upstream [PR](https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/16405))*
- [X] Support `v-pred` **SDXL** checkpoints *(**e.g.** [NoobAI](https://civitai.com/models/833294?modelVersionId=1190596))*
- [X] Support new LoRA architectures
- [X] Update `spandrel`
- support new Upscaler architectures
- [X] Add `pillow-heif` package
- support `.avif` and `.heif` images
- [X] Automatically determine the optimal row count for `X/Y/Z Plot`
- [X] Support new LoRA architectures
- [X] `DepthAnything v2` Preprocessor
- [X] Support [NoobAI Inpaint](https://civitai.com/models/1376234/noobai-inpainting-controlnet) ControlNet
- [X] Support [Union](https://huggingface.co/xinsir/controlnet-union-sdxl-1.0) / [ProMax](https://huggingface.co/brad-twinkl/controlnet-union-sdxl-1.0-promax) ControlNet
- they simply always show up in the dropdown
#### Removed Features
- [X] SD2
- [X] Alt-Diffusion
- [X] Instruct-Pix2Pix
- [X] Hypernetworks
- [X] SVD
- [X] Z123
- [X] CLIP Interrogator
- [X] Deepbooru Interrogator
- [X] Textual Inversion Training
- [X] Checkpoint Merging
- [X] LDSR
- [X] Most built-in Extensions
- [X] Some built-in Scripts
- [X] Some Samplers
- [X] Sampler in RadioGroup
- [X] `test` scripts
- [X] Some Preprocessors *(ControlNet)*
- [X] `Photopea` and `openpose_editor` *(ControlNet)*
- [X] Unix `.sh` launch scripts
- You can still use this WebUI by simply copying a launch script from other working WebUI
#### Optimizations
- [X] **[Freedom]** Natively integrate the `SD1` and `SDXL` logics
- no longer `git` `clone` any repository on fresh install
- no more random hacks and monkey patches
- [X] Fix `canvas-zoom-and-pan` built-in extension
- no more infinite-resizing bug when using `Send to` buttons
- [X] Fix memory leak when switching checkpoints
- [X] Clean up the `ldm_patched` *(**i.e.** `comfy`)* folder
- [X] Remove unused `cmd_args`
- [X] Remove unused `args_parser`
- [X] Remove unused `shared_options`
- [X] Remove legacy codes
- [X] Fix some typos
- [X] Remove redundant upscaler codes
- put every upscaler inside the `ESRGAN` folder
- [X] Optimize upscaler logics
- [X] Optimize certain operations in `Spandrel`
- [X] Optimize the creation of Extra Networks pages
- *(backported from Automatic1111 Webui upstream)*
- [X] Improve color correction
- [X] Improve hash caching
- [X] Improve error logs
- no longer print `TypeError: 'NoneType' object is not iterable`
- [X] Revamp settings
- improve formatting
- update descriptions
- [X] Check for Extension updates in parallel
- [X] Move `embeddings` folder into `models` folder
- [X] ControlNet Rewrite
- change Units to `gr.Tab`
- remove multi-inputs, as they are "[misleading](https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/932)"
- change `visible` toggle to `interactive` toggle; now the UI will no longer jump around
- improve `Presets` application
- fix `Inpaint not masked` mode
- [X] Disable Refiner by default
- enable again in **Settings/UI Alternatives**
- [X] Disable Tree View by default
- enable again in **Settings/Extra Networks**
- [X] Hide Sampler Parameters by default
- enable again by adding **--adv-samplers** flag
- [X] Hide some X/Y/Z Plot options by default
- enable again by adding **--adv-xyz** flag
- [X] Run `text encoder` on CPU by default
- [X] Fix `pydantic` Errors
- [X] Fix `Soft Inpainting`
- [X] Fix `Controllllite`
- [X] Fix `MultiDiffusion`
- [X] Lint & Format
- [X] Update `Pillow`
- faster image processing
- [X] Update `protobuf`
- faster `insightface` loading
- [X] Update to latest PyTorch
- `torch==2.7.1+cu128`
- `xformers==0.0.30`
> [!Note]
> If your GPU does not support the latest PyTorch, manually [install](#install-older-pytorch) older version of PyTorch
- [X] No longer install `open-clip` twice
- [X] Update some packages to newer versions
- [X] Update recommended Python to `3.11.9`
- [X] many more... :tm:
## Commandline
> These flags can be added after the `set COMMANDLINE_ARGS=` line in the `webui-user.bat` *(separate each flag with space)*
#### A1111 built-in
- `--no-download-sd-model`: Do not download a default checkpoint
- can be removed after you download some checkpoints of your choice
- `--xformers`: Install the `xformers` package to speed up generation
- `--port`: Specify a server port to use
- defaults to `7860`
- `--api`: Enable [API](https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/API) access
- Once you have successfully launched the WebUI, you can add the following flags to bypass some validation steps in order to improve the Startup time
- `--skip-prepare-environment`
- `--skip-install`
- `--skip-python-version-check`
- `--skip-torch-cuda-test`
- `--skip-version-check`
> [!Important]
> Remove them if you are installing an Extension, as those also block Extension from installing requirements
#### by. Forge
- For RTX **30** and above, you can add the following flags to slightly increase the performance; but in rare occurrences, they may cause `OutOfMemory` errors or even crash the WebUI; and in certain configurations, they may even lower the speed instead
- `--cuda-malloc`
- `--cuda-stream`
- `--pin-shared-memory`
#### by. Classic
- `--uv`: Replace the `python -m pip` calls with `uv pip` to massively speed up package installation
- requires **uv** to be installed first *(see [Installation](#installation))*
- `--uv-symlink`: Same as above; but additionally pass `--link-mode symlink` to the commands
- significantly reduces installation size (`~7 GB` to `~100 MB`)
> [!Important]
> Using `symlink` means it will directly access the packages from the cache folders; refrain from clearing the cache when setting this option
- `--model-ref`: Points to a central `models` folder that contains all your models
- said folder should contain subfolders like `Stable-diffusion`, `Lora`, `VAE`, `ESRGAN`, etc.
> [!Important]
> This simply **replaces** the `models` folder, rather than adding on top of it
- `--persistent-patches`: Enable the persistent LoRA patching
- no longer apply LoRA every single generation, if the weight is unchanged
- save around 1 second per generation when using LoRA
- `--fast-fp16`: Enable the `allow_fp16_accumulation` option
- requires PyTorch **2.7.0** +
- `--sage`: Install the `sageattention` package to speed up generation
- requires **triton**
- requires RTX **30** +
- only affects **SDXL**
> [!Note]
> For RTX **50** users, you may need to manually [install](#install-sageattention-2) `sageattention 2` instead
with SageAttention 2
- `--sageattn2-api`: Select the function used by **SageAttention 2**
- **options:**
- `auto` (default)
- `triton-fp16`
- `cuda-fp16`
- `cuda-fp8`
- try the `fp16` options if you get `NaN` *(black images)* on `auto`
## Installation
0. Install **[git](https://git-scm.com/downloads)**
1. Clone the Repo
```bash
git clone https://github.com/Haoming02/sd-webui-forge-classic
```
2. Setup Python
Recommended Method
- Install **[uv](https://github.com/astral-sh/uv#installation)**
- Set up **venv**
```bash
cd sd-webui-forge-classic
uv venv venv --python 3.11 --seed
```
- Add the `--uv` flag to `webui-user.bat`
Standard Method
- Install **[Python 3.11.9](https://www.python.org/downloads/release/python-3119/)**
- Remember to enable `Add Python to PATH`
3. **(Optional)** Configure [Commandline](#commandline)
4. Launch the WebUI via `webui-user.bat`
5. During the first launch, it will automatically install all the requirements
6. Once the installation is finished, the WebUI will start in a browser automatically
### Install cublas
Expand
0. Ensure the WebUI can properly launch already, by following the [installation](#installation) steps first
1. Open the console in the WebUI directory
```bash
cd sd-webui-forge-classic
```
2. Start the virtual environment
```bash
venv\scripts\activate
```
3. Create a new folder
```bash
mkdir repo
cd repo
```
4. Clone the repo
```bash
git clone https://github.com/aredden/torch-cublas-hgemm
cd torch-cublas-hgemm
```
5. Install the library
```
pip install -e . --no-build-isolation
```
- If you installed `uv`, use `uv pip install` instead
- The installation takes a few minutes
### Install triton
Expand
0. Ensure the WebUI can properly launch already, by following the [installation](#installation) steps first
1. Open the console in the WebUI directory
```bash
cd sd-webui-forge-classic
```
2. Start the virtual environment
```bash
venv\scripts\activate
```
3. Install the library
- **Windows**
```bash
pip install triton-windows
```
- **Linux**
```bash
pip install triton
```
- If you installed `uv`, use `uv pip install` instead
### Install flash-attn
Expand
0. Ensure the WebUI can properly launch already, by following the [installation](#installation) steps first
1. Open the console in the WebUI directory
```bash
cd sd-webui-forge-classic
```
2. Start the virtual environment
```bash
venv\scripts\activate
```
3. Install the library
- **Windows**
- Download the pre-built `.whl` package from https://github.com/kingbri1/flash-attention/releases
```bash
pip install flash_attn...win...whl
```
- **Linux**
- Download the pre-built `.whl` package from https://github.com/Dao-AILab/flash-attention/releases
```bash
pip install flash_attn...linux...whl
```
- If you installed `uv`, use `uv pip install` instead
- **Important:** Download the correct `.whl` for your Python and PyTorch version
### Install sageattention 2
Expand
0. Ensure the WebUI can properly launch already, by following the [installation](#installation) steps first
1. Open the console in the WebUI directory
```bash
cd sd-webui-forge-classic
```
2. Start the virtual environment
```bash
venv\scripts\activate
```
3. Create a new folder
```bash
mkdir repo
cd repo
```
4. Clone the repo
```bash
git clone https://github.com/thu-ml/SageAttention
cd SageAttention
```
5. Install the library
```
pip install -e . --no-build-isolation
```
- If you installed `uv`, use `uv pip install` instead
- The installation takes a few minutes
### Install older PyTorch
Expand
0. Navigate to the WebUI directory
1. Edit the `webui-user.bat` file
2. Add a new line to specify an older version:
```bash
set TORCH_COMMAND=pip install torch==2.1.2 torchvision==0.16.2 --extra-index-url https://download.pytorch.org/whl/cu121
```
## Attention
> [!Important]
> The `--xformers` and `--sage` args are only responsible for installing the packages, **not** whether its respective attention is used *(this also means you can remove them once the packages are successfully installed)*
**Forge Classic** tries to import the packages and automatically choose the first available attention function in the following order:
1. `SageAttention`
2. `FlashAttention`
3. `xformers`
4. `PyTorch`
5. `Basic`
> [!Tip]
> To skip a specific attention, add the respective disable arg such as `--disable-sage`
> [!Note]
> The **VAE** only checks for `xformers`, so `--xformers` is still recommended even if you already have `--sage`
In my experience, the speed of each attention function for SDXL is ranked in the following order:
- `SageAttention` ≥ `FlashAttention` > `xformers` > `PyTorch` >> `Basic`
> [!Note]
> `SageAttention` is based on quantization, so its quality might be slightly worse than others
> [!Important]
> When using `SageAttention 2`, both positive prompts and negative prompts are required; omitting negative prompts can cause `NaN` issues
## Issues & Requests
- **Issues** about removed features will simply be ignored
- **Issues** regarding installation will be ignored if it's obviously user-error
- **Feature Request** not related to performance or optimization will simply be ignored
- For cutting edge features, check out [reForge](https://github.com/Panchovix/stable-diffusion-webui-reForge) instead
- Non-Windows platforms will not be supported, as I cannot verify nor maintain them
Special thanks to AUTOMATIC1111, lllyasviel, and comfyanonymous, kijai,
along with the rest of the contributors,
for their invaluable efforts in the open-source image generation community