mirror of
https://github.com/vladmandic/automatic
synced 2026-09-19 09:14:35 +02:00
wiki fix all links
Co-authored-by: Copilot <copilot@github.com> Signed-off-by: vladmandic <mandic00@live.com>
vladmandic
+38
-19
@@ -1,16 +1,16 @@
|
||||
# AMD ROCm
|
||||
|
||||
To use **AMD ROCm** with **SD.Next** you need to
|
||||
To use **AMD ROCm** with **SD.Next** you need to
|
||||
1. Install ROCm libraries first.
|
||||
2. Run SD.Next with `--use-rocm` flag to force it to install appropriate version of `torch`.
|
||||
|
||||
> [!IMPORTANT]
|
||||
> AMD ROCm is officially supported for specific AMD GPUs.
|
||||
|
||||
>
|
||||
> [!IMPORTANT]
|
||||
> Currently, PyTorch support on Windows is not officially maintained by PyTorch team.
|
||||
> See [AMD's announcement](https://www.amd.com/en/resources/support-articles/release-notes/RN-AMDGPU-WINDOWS-PYTORCH-PREVIEW.html) for more information.
|
||||
|
||||
>
|
||||
> [!WARNING]
|
||||
> Unofficial support for other platforms is provided by the community and **SD.Next** does not guarantee it will work.
|
||||
> Use of any third-party libraries is at your own risk.
|
||||
@@ -23,7 +23,8 @@ To use **AMD ROCm** with **SD.Next** you need to
|
||||
### Install Guide for Ubuntu 24.04
|
||||
|
||||
Install ROCm:
|
||||
```
|
||||
|
||||
```shell
|
||||
sudo apt update
|
||||
wget https://repo.radeon.com/amdgpu-install/6.4.3/ubuntu/noble/amdgpu-install_6.4.60403-1_all.deb
|
||||
sudo apt install ./amdgpu-install_6.4.60403-1_all.deb
|
||||
@@ -32,6 +33,7 @@ sudo usermod -a -G render,video $LOGNAME
|
||||
```
|
||||
|
||||
Install git and python:
|
||||
|
||||
```shell
|
||||
sudo apt install git python3 python3-dev python3-venv python3-pip
|
||||
```
|
||||
@@ -43,17 +45,20 @@ Simply change the wget line from "noble" to "jammy" if using Ubuntu 22.04.
|
||||
### Install Guide for openSUSE Tumbleweed
|
||||
|
||||
Install prerequisites:
|
||||
|
||||
```shell
|
||||
sudo zypper in python312-devel python312-virtualenv python312-pip patterns-devel-base-devel_basis
|
||||
```
|
||||
|
||||
Add the [ROCm repository](https://build.opensuse.org/project/show/science:GPU:ROCm) (not official, but maintained by AMD employees):
|
||||
```shell
|
||||
|
||||
```shell
|
||||
sudo zypper ar obs://science:GPU:ROCm/openSUSE_Factory ROCm
|
||||
sudo zypper ref # Answer "ultimately trust"
|
||||
```
|
||||
```
|
||||
|
||||
Install the relevant packages (there is no pattern so you must install them all manually):
|
||||
|
||||
```shell
|
||||
sudo zypper in rocm-runtime \
|
||||
miopen rccl rocblas amdsmi \
|
||||
@@ -70,16 +75,18 @@ sudo zypper in rocm-runtime \
|
||||
|
||||
This procedure should also work for Leap based distributions and Slowroll (change the relevant lines for the distribution, check the repository page for details), but they haven't been tested.
|
||||
|
||||
**NOTE**: This installs also the build dependencies for flash-attention.
|
||||
**NOTE**: This installs also the build dependencies for flash-attention.
|
||||
|
||||
### Install Guide for Arch Linux
|
||||
|
||||
Install ROCm and git:
|
||||
|
||||
```shell
|
||||
sudo pacman -S rocm-hip-runtime git
|
||||
```
|
||||
|
||||
Install Python 3.12 (or anything between 3.10 and 3.13):
|
||||
|
||||
```shell
|
||||
sudo pacman -S base-devel python-pip python-virtualenv
|
||||
git clone https://aur.archlinux.org/python312.git
|
||||
@@ -96,23 +103,27 @@ Install ROCm SDK:
|
||||
> [!NOTE]
|
||||
> ROCm SDK is optional. Only required for building flash atten or similar custom kernels.
|
||||
> ROCm SDK uses 26 GB of disk space.
|
||||
|
||||
```shell
|
||||
sudo pacman -S rocm-hip-sdk libxml2-legacy gcc14 gcc14-libs
|
||||
```
|
||||
|
||||
### Running SD.Next with ROCm locally
|
||||
## Running SD.Next with ROCm locally
|
||||
|
||||
Open the terminal in a folder you want to install SD.Next and install SD.Next from Github with this command:
|
||||
|
||||
```shell
|
||||
git clone https://github.com/vladmandic/sdnext
|
||||
```
|
||||
|
||||
Then enter into the sdnext folder:
|
||||
|
||||
```shell
|
||||
cd sdnext
|
||||
```
|
||||
|
||||
Then run SD.Next with this command:
|
||||
|
||||
```shell
|
||||
./webui.sh --use-rocm
|
||||
```
|
||||
@@ -129,7 +140,7 @@ Checkout the [Docker wiki](https://github.com/vladmandic/sdnext/wiki/Docker) if
|
||||
|
||||
Using Docker with a prebuilt image:
|
||||
|
||||
```
|
||||
```shell
|
||||
export SDNEXT_DOCKER_ROOT_FOLDER=~/sdnext
|
||||
sudo docker run -it \
|
||||
--name sdnext-rocm \
|
||||
@@ -155,19 +166,20 @@ For more details, see [AMD-MIOpen Guide](AMD-MIOpen.md) page.
|
||||
### MIOpen database tuning
|
||||
|
||||
On first use, when using a resolution for the first time, or when upgrading Pytorch versions, ROCm runs a series of benchmarks to select the most efficient approach. This can lead to slow (up to 5-8 minutes) startup times, in particular if you use a refine pass at high resolution, but it only happens once per resolution (subsequent runs are much faster). If for any case this is undesirable, you can set the environment variable `MIOPEN_FIND_MODE` to `FAST`. This will reduce a lot the startup time on first use at the price of worse performance when generating.
|
||||
On the other hand, for best performance during generation (but slower startup on first use), you can set the variable `MIOPEN_FIND_ENFORCE` to `SEARCH`.
|
||||
On the other hand, for best performance during generation (but slower startup on first use), you can set the variable `MIOPEN_FIND_ENFORCE` to `SEARCH`.
|
||||
|
||||
### Reduce VRAM consumption
|
||||
|
||||
If you use the `bf16` data type (*Settings > Compute Settings > Execution Precision > Device precision type*), which is autodetected on RDNA3 and newer cards, there is the chance that VRAM usage will be very high (16+ GB) when decoding the final image and when upscaling with non-latent upscalers. To workaround the problem, ensure to set `Device precision type` as `fp16`, and disable VAE upcasting in *Variational Auto Encoder > VAE upcasting*.
|
||||
Setting `fp16` has also a noticeable improvement on performance.
|
||||
If you use the `bf16` data type (*Settings > Compute Settings > Execution Precision > Device precision type*), which is autodetected on RDNA3 and newer cards, there is the chance that VRAM usage will be very high (16+ GB) when decoding the final image and when upscaling with non-latent upscalers. To workaround the problem, ensure to set `Device precision type` as `fp16`, and disable VAE upcasting in *Variational Auto Encoder > VAE upcasting*.
|
||||
Setting `fp16` has also a noticeable improvement on performance.
|
||||
|
||||
### Composable Kernel (CK) Flash attention
|
||||
|
||||
On RDNA3 hardware (RX 7000 series), you can set the option to use CK Flash Attention for improved performance in *Compute Settings > Cross Attention > SDP Options* by toggling `CK Flash attention` and restarting SD.Next. Notice that enabling this option requires `rocm-hip-sdk` installed as it will download and compile an additional Python package from source on startup.
|
||||
|
||||
On RDNA3 hardware (RX 7000 series), you can set the option to use CK Flash Attention for improved performance in *Compute Settings > Cross Attention > SDP Options* by toggling `CK Flash attention` and restarting SD.Next. Notice that enabling this option requires `rocm-hip-sdk` installed as it will download and compile an additional Python package from source on startup.
|
||||
|
||||
In case you want to install it manually, activate the virtual environment then run `pip`:
|
||||
```
|
||||
|
||||
```shell
|
||||
pip install --no-build-isolation git+https://github.com/Disty0/flash-attention@navi_rotary_fix
|
||||
```
|
||||
|
||||
@@ -175,18 +187,25 @@ pip install --no-build-isolation git+https://github.com/Disty0/flash-attention@n
|
||||
|
||||
1. Install [Git](https://git-scm.com/downloads/win) and [Python 3.12](https://www.python.org/downloads/release/python-31210/).
|
||||
2. Open the terminal in a folder you want to install SD.Next and install SD.Next from GitHub with this command:
|
||||
|
||||
```shell
|
||||
git clone https://github.com/vladmandic/sdnext
|
||||
```
|
||||
3. Enter into the sdnext folder:
|
||||
|
||||
1. Enter into the sdnext folder:
|
||||
|
||||
```shell
|
||||
cd sdnext
|
||||
```
|
||||
4. Make sure that you are up to date.
|
||||
|
||||
1. Make sure that you are up to date.
|
||||
|
||||
```shell
|
||||
git pull
|
||||
```
|
||||
5. Run SD.Next with this command:
|
||||
|
||||
1. Run SD.Next with this command:
|
||||
|
||||
```shell
|
||||
./webui.bat --use-rocm
|
||||
```
|
||||
```
|
||||
|
||||
+17
-11
@@ -2,24 +2,30 @@
|
||||
|
||||
## Start Scripts
|
||||
|
||||
Start scripts `webui.bat` or `webui.sh` are provided to create and activate VENV and immediately start launcher.
|
||||
No other work is performed in the shell scripts.
|
||||
Start scripts `webui.bat` and `webui.sh` create and activate `venv` (a Python virtual environment), then start the launcher.
|
||||
The scripts do not perform any other setup tasks.
|
||||
|
||||
Actual launcher is started using `python launch.py` command.
|
||||
The launcher itself is started with `python launch.py`.
|
||||
|
||||
If you start launcher manually without creating & activating VENV first, it will install packages system wide.
|
||||
This may be desired when running **SD.Next** in a dedicated container where there is no benefits of running additional isolation provided by **VEVN**.
|
||||
If you start the launcher manually without creating and activating `venv` first, dependencies are installed system-wide.
|
||||
This can be intentional when running **SD.Next** in a dedicated container where extra virtual-environment isolation is not needed.
|
||||
|
||||
## VENV
|
||||
|
||||
SD.Next by default uses `venv` to install all dependencies
|
||||
Usage of `venv` is not required, but it is recommended to avoid library version conflicts with other applications
|
||||
By default, SD.Next uses `venv` to install dependencies.
|
||||
Using `venv` is optional, but recommended to avoid dependency conflicts with other applications.
|
||||
|
||||
You can also pre-create `venv` to use specific settings, for example:
|
||||
> python -m venv venv --system-site-packages
|
||||
You can pre-create `venv` with specific options. For example:
|
||||
|
||||
This will instruct **VENV** to use system site packages where available and only install missing/incorrect packages inside **VENV**
|
||||
```bash
|
||||
python -m venv venv --system-site-packages
|
||||
```
|
||||
|
||||
This setup uses system site packages where available and installs only missing or incompatible packages inside `venv`.
|
||||
|
||||
## Upgrades
|
||||
|
||||
**SD.Next** has built-in upgrade mechanism when using `--upgrade` command line flag, but its fully supported to run manual upgrades using `git pull` as well.
|
||||
**SD.Next** includes a built-in upgrade path with the `--upgrade` flag, and manual upgrades with `git pull` are also supported.
|
||||
|
||||
> [!IMPORTANT]
|
||||
> If you use manual upgrades with `git pull`, also update all **submodules**.
|
||||
|
||||
+10
-10
@@ -1,43 +1,43 @@
|
||||
# FAQ: Frequently asked Questions
|
||||
|
||||
### Where is the "PNG Info" tab?
|
||||
## Where is the "PNG Info" tab?
|
||||
|
||||
*The functionality is integrated into the "Process Image" tabs, removing the need for a separate "PNG Info" tab.*
|
||||
|
||||
### Where are the command-line flags like --xformers?
|
||||
## Where are the command-line flags like --xformers?
|
||||
|
||||
*Most command-line flags have been moved to UI → Settings. For a complete list, start the web UI with the --help flag.*
|
||||
|
||||
### How can I get more information about what’s happening with my app?
|
||||
## How can I get more information about what’s happening with my app?
|
||||
|
||||
*Launch the app with the --debug flag and review the setup.log file for detailed insights.*
|
||||
|
||||
### How do I add command-line options to webui.bat in Windows?
|
||||
## How do I add command-line options to webui.bat in Windows?
|
||||
|
||||
*1. Right-click on webui.bat and select "Create Shortcut".
|
||||
2. Right-click the shortcut and open "Properties".
|
||||
3. Add the options at the end of the "Target" field after a space. Example:
|
||||
``"C:\path\to\webui.bat" --medvram --autolaunch``
|
||||
`"C:\path\to\webui.bat" --medvram --autolaunch`
|
||||
4. Click OK and use this shortcut to start the app.*
|
||||
|
||||
### How do I use an AMD GPU on Windows?
|
||||
## How do I use an AMD GPU on Windows?
|
||||
|
||||
*Add the --use-zluda command-line flag when starting the app.*
|
||||
Checkout the [Zluda wiki](ZLUDA) for more info.
|
||||
|
||||
### How can I create large images (e.g., 2048x2048) with limited VRAM?
|
||||
## How can I create large images (e.g., 2048x2048) with limited VRAM?
|
||||
|
||||
*Render a smaller image (e.g., 512x512) and upscale it using the "Process Image" tab with an upscaler like SwinIR_4x. This method tiles the image to minimize VRAM usage.*
|
||||
|
||||
### Why are my images dull or have incorrect colors?
|
||||
## Why are my images dull or have incorrect colors?
|
||||
|
||||
*Try a different VAE file. Add ,sd_vae to the Settings → User Interface → Quicksettings list to enable quick VAE selection. Place new VAE files in the models\VAE folder.*
|
||||
|
||||
### How do I update to the latest version?
|
||||
## How do I update to the latest version?
|
||||
|
||||
*Run webui.bat --upgrade. This pulls the latest updates and applies them to your installation.*
|
||||
|
||||
### Something broke. How do I restore my SD.Next installation?
|
||||
## Something broke. How do I restore my SD.Next installation?
|
||||
|
||||
*Use webui.bat --reinstall to reinstall required components or webui.bat --reset to reset to the latest version.*
|
||||
|
||||
|
||||
+5
-4
@@ -5,6 +5,7 @@ SD.Next localization system is based on the single JSON file: [html/locale_en.js
|
||||
Structure of the file is simple: For each visible UI label, there is a corresponding hint that is displayed when user hovers over the label.
|
||||
|
||||
*Example*: generate button:
|
||||
|
||||
```json
|
||||
{"id":"","label":"Generate","localized":"","hint":"Start processing"}
|
||||
```
|
||||
@@ -12,16 +13,16 @@ Structure of the file is simple: For each visible UI label, there is a correspon
|
||||
> [!TIP]
|
||||
> See [Hints](Hints) document for more details
|
||||
|
||||
This file which is then auto-translated using LLM model to other languages and saved as separate files in a format `hmtl/locale_{locale}.json`
|
||||
This file is then auto-translated using an LLM model to other languages and saved as separate files in the format `html/locale_{locale}.json`.
|
||||
Those files should not be edited directly as they are auto-generated and will be overwritten on the next update, but if you have suggestions for improvements, please contact us
|
||||
|
||||
> [!NOTE]
|
||||
> Translation is performed using `cli/localize.js` script
|
||||
|
||||
### Editing Localization
|
||||
## Editing Localization
|
||||
|
||||
In case of bad ML translation or you just want to provide a better alternernative,
|
||||
you can create/edit `hmtl/override_{locale}.json` which only needs to include specific items you want to override
|
||||
In case of a bad ML translation, or if you want to provide a better alternative,
|
||||
you can create/edit `html/override_{locale}.json`, which only needs to include specific items you want to override.
|
||||
|
||||
### Supported Languages
|
||||
|
||||
|
||||
+3
-3
@@ -92,7 +92,7 @@ One more thing to watch out for, is file paths. For the normal search the Window
|
||||
|
||||
Considering the RegEx search, the field which is searched is structured in the following way:
|
||||
|
||||
```JS
|
||||
```js
|
||||
`filename: ${filename}|name: ${name}|tags: ${tags}`
|
||||
```
|
||||
|
||||
@@ -108,7 +108,7 @@ For example a negative lookahead can be used to exclude certain tags from the se
|
||||
|
||||
An example RegEx would be the following for this use case:
|
||||
|
||||
```RegEx
|
||||
```regex
|
||||
r#^(?!.*\|tags:.*(1girl|woman|girl)).*$
|
||||
```
|
||||
|
||||
@@ -117,7 +117,7 @@ This search excludes every network which has tags which contain either `1girl`,
|
||||
Another usecase would be searching in just the filename, but not in the entire path.
|
||||
For example requiring the network to contain `SDXL`, with it not mattering whether it is in a folder called `SDXL` or not:
|
||||
|
||||
```RegEx
|
||||
```regex
|
||||
r#^.*\|name:[^|]*([^\/\\]*SDXL[^\/\\]*)$(?<![\/\\]).*$
|
||||
```
|
||||
|
||||
|
||||
+1
-1
@@ -46,7 +46,7 @@ SD.Next is a powerful, open-source WebUI app for AI image and video generation,
|
||||
## Features & Capabilities
|
||||
|
||||
SD.Next is feature-rich with a focus on performance, flexibility, and user experience. Key features include:
|
||||
- [Multi-platform](#platform-support!
|
||||
- [Multi-platform](#supported-platforms--hardware)
|
||||
- Many [diffusion models](https://vladmandic.github.io/sdnext-docs/Model-Support/)!
|
||||
- Fully localized to ~15 languages and with support for many [UI themes](https://vladmandic.github.io/sdnext-docs/Themes/)!
|
||||
- [Desktop](#screenshot-desktop-interface) and [Mobile](#screenshot-mobile-interface) support!
|
||||
|
||||
+6
-3
@@ -1,12 +1,12 @@
|
||||
# NudeNet
|
||||
|
||||
### Censorship... with style! ™
|
||||
## Censorship... with style! ™
|
||||
|
||||
## Features
|
||||
|
||||
### Main NudeNet
|
||||
|
||||
- Detect geneder and any body part
|
||||
- Detect gender and any body part
|
||||
e.g. *female face, belly, feet*
|
||||
- Exposed or unexposed:
|
||||
e.g. *breast or breast-bare*
|
||||
@@ -27,7 +27,8 @@
|
||||
### Language Check
|
||||
|
||||
If **Check language** is enabled, NudeNet will use `facebook/fasttext-language-identification` to check if prompt is in allowed language and using allowed alphabet.
|
||||
For example:
|
||||
For example:
|
||||
|
||||
```log
|
||||
LangDetect: ['eng_latn:0.83']
|
||||
```
|
||||
@@ -38,6 +39,7 @@ Means that language is English and uses Latin alphabet with 83% confidence.
|
||||
|
||||
If **Policy check** is enabled, NudeNet will use `LlavaGuard` VLM to check if image violates any policies and reasons why.
|
||||
For example:
|
||||
|
||||
```log
|
||||
'rating': 'Unsafe',
|
||||
'category': 'O3: Sexual Content',
|
||||
@@ -48,6 +50,7 @@ For example:
|
||||
|
||||
If **Banned words** check is enabled, NudeNet will use `regex` to check prompt variations if they contain any banned words or expressions.
|
||||
For example:
|
||||
|
||||
```log
|
||||
BannedWords: ['naked']
|
||||
```
|
||||
|
||||
Reference in New Issue
Block a user