10/03/2025, 19:49 Command Line Arguments and Settings · AUTOMATIC1111/stable-diffusion-webui Wiki · GitHub

All command line arguments

Argument
Value Default Description
Command

CONFIGURATION

-h, --help None False Show this help message and exit.

--exit Terminate after installation

--data-dir DATA_DIR ./ base path where all user data is stored

--models-dir MODELS None base path where models are stored; overrides --data-dir

configs/stable-diffusion/v1-
--config CONFIG Path to config which constructs model.
inference.yaml

Path to checkpoint of Stable Diffusion model; if specified,

--ckpt CKPT model.ckpt this checkpoint will be added to the list of checkpoints and
loaded.

--ckpt-dir CKPT_DIR None Path to directory with Stable Diffusion checkpoints.

--no-download-sd-
None False Don't download SD1.5 model even if no model is found.
model

--do-not-download- do not download CLIP model even if it's not included in the
None False
clip checkpoint

--vae-dir VAE_PATH None Path to Variational Autoencoders model

--vae-path VAE_PATH None Checkpoint to use as VAE; setting this argument

--gfpgan-dir GFPGAN_DIR GFPGAN/ GFPGAN directory.

--gfpgan-model GFPGAN_MODEL GFPGAN model file name.

--codeformer-
CODEFORMER_MODELS_PATH models/Codeformer/ Path to directory with codeformer model file(s).
models-path

--gfpgan-models-
GFPGAN_MODELS_PATH models/GFPGAN Path to directory with GFPGAN model file(s).
path

--esrgan-models-
ESRGAN_MODELS_PATH models/ESRGAN Path to directory with ESRGAN model file(s).
path

--bsrgan-models-
BSRGAN_MODELS_PATH models/BSRGAN Path to directory with BSRGAN model file(s).
path

--realesrgan-
REALESRGAN_MODELS_PATH models/RealESRGAN Path to directory with RealESRGAN model file(s).
models-path

--scunet-models-
SCUNET_MODELS_PATH models/ScuNET Path to directory with ScuNET model file(s).
path

--swinir-models-
SWINIR_MODELS_PATH models/SwinIR Path to directory with SwinIR and SwinIR v2 model file(s).
path

--ldsr-models-path LDSR_MODELS_PATH models/LDSR Path to directory with LDSR model file(s).

--dat-models-path DAT__MODELS_PATH models/DAT Path to directory with DAT model file(s).

--lora-dir LORA_DIR models/Lora Path to directory with Lora networks.

--clip-models-path CLIP_MODELS_PATH None Path to directory with CLIP model file(s).

Embeddings directory for textual inversion (default:

--embeddings-dir EMBEDDINGS_DIR embeddings/
embeddings).

--textual-inversion-
TEXTUAL_INVERSION_TEMPLATES_DIR textual_inversion_templates Directory with textual inversion templates.
templates-dir

--hypernetwork-dir HYPERNETWORK_DIR models/hypernetworks/ hypernetwork directory.

--localizations-dir LOCALIZATIONS_DIR localizations/ Localizations directory.

--styles-file STYLES_FILE styles.csv Path or wildcard path of styles files, allow multiple entries.

--ui-config-file UI_CONFIG_FILE ui-config.json Filename to use for UI configuration.

Do not hide progress bar in gradio UI (we hide it because it

--no-progressbar-
None False slows down ML if you have hardware acceleration in
hiding
browser).

--ui-settings-file UI_SETTINGS_FILE config.json Filename to use for UI settings.

--allow-code None False Allow custom script execution from web UI.

Use share=True for gradio and make the UI accessible

--share None False
through their site.

Launch gradio with 0.0.0.0 as server name, allowing to

--listen None False
respond to network requests.

Launch gradio with given server port, you need root/admin

--port PORT 7860
rights for ports < 1024; defaults to 7860 if available.

https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Command-Line-Arguments-and-Settings#all-command-line-arguments 1/4
10/03/2025, 19:49 Command Line Arguments and Settings · AUTOMATIC1111/stable-diffusion-webui Wiki · GitHub

Argument
Value Default Description
Command

--hide-ui-dir-config None False Hide directory configuration from web UI.

--freeze-settings None False disable editing of all settings globally

disable editing settings in specific sections of the settings

--freeze-settings-in- page by specifying a comma-delimited list such like
None False
sections "saving-images,upscaling". The list of setting names can be
found in the modules/shared_options.py file

disable editing of individual settings by specifying a

--freeze-specific-
None False comma-delimited list like "samples_save,samples_format".
settings
The list of setting names can be found in the config.json file

--enable-insecure-
None False Enable extensions tab regardless of other options.
extension-access

--gradio-debug None False Launch gradio with --debug option.

Set gradio authentication like username:password ; or

--gradio-auth GRADIO_AUTH None
comma-delimit multiple like u1:p1,u2:p2,u3:p3 .

Set gradio authentication file path

--gradio-auth-path GRADIO_AUTH_PATH None ex. /path/to/auth/file same auth format as --gradio-
auth .

--disable-console-
None False Do not output progress bars to console.
progressbars

--enable-console- Print prompts to console when generating with txt2img

None False
prompts and img2img.

--api None False Launch web UI with API.

Set authentication for API like username:password ; or

--api-auth API_AUTH None
comma-delimit multiple like u1:p1,u2:p2,u3:p3 .

--api-log None False Enable logging of all API requests.

--nowebui None False Only launch the API, without the UI.

--ui-debug-mode None False Don't load model to quickly launch UI.

Select the default CUDA device to use

--device-id DEVICE_ID None (export CUDA_VISIBLE_DEVICES=0,1 etc might be needed
before).

--administrator None False Administrator privileges.

Allowed CORS origin(s) in the form of a comma-separated

--cors-allow-origins CORS_ALLOW_ORIGINS None
list (no spaces).

--cors-allow-origins- Allowed CORS origin(s) in the form of a single regular

CORS_ALLOW_ORIGINS_REGEX None
regex expression.

Partially enables TLS, requires --tls-certfile to fully

--tls-keyfile TLS_KEYFILE None
function.

Partially enables TLS, requires --tls-keyfile to fully

--tls-certfile TLS_CERTFILE None
function.

--disable-tls-verify None False When passed, enables the use of self-signed certificates.

customize the subpath for

--subpath SERVER_SUB_PATH
gradio, use with reverse proxy

--server-name SERVER_NAME None Sets hostname of server.

Disables gradio queue; causes the webpage to use http

--no-gradio-queue None False requests instead of websockets; was the default in earlier
versions.

--gradio-allowed- Add path to Gradio's allowed_paths ; make it possible to

None None
path serve files from it.

Disable SHA-256 hashing of checkpoints to help loading

--no-hashing None False
performance.

--skip-version-check None False Do not check versions of torch and xformers.

--skip-python-
None False Do not check versions of Python.
version-check

--skip-torch-cuda-
None False Do not check if CUDA is able to work properly.
test

--skip-install None False Skip installation of packages.

log level; one of: CRITICAL, ERROR, WARNING, INFO,

--loglevel None None
DEBUG

launch.py argument: print a detailed log of what's

--log-startup None False
happening at startup

--api-server-stop None False enable server stop/restart/kill via api

https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Command-Line-Arguments-and-Settings#all-command-line-arguments 2/4
10/03/2025, 19:49 Command Line Arguments and Settings · AUTOMATIC1111/stable-diffusion-webui Wiki · GitHub

Argument
Value Default Description
Command

--timeout-keep-
int 30 set timeout_keep_alive for uvicorn
alive

PERFORMANCE

--xformers None False Enable xformers for cross attention layers.

Enable xformers for cross attention layers regardless of

--force-enable-
None False whether the checking code thinks you can run it; do not
xformers
make bug reports if this fails to work.

--xformers-flash- Enable xformers with Flash Attention to improve

None False
attention reproducibility (supported for SD2.x or variant only).

Enable scaled dot product cross-attention layer

--opt-sdp-attention None False
optimization; requires PyTorch 2.*

Enable scaled dot product cross-attention layer

--opt-sdp-no-mem-
False None optimization without memory efficient attention, makes
attention
image generation deterministic; requires PyTorch 2.*

Force-enables Doggettx's cross-attention layer

--opt-split-attention None False
optimization. By default, it's on for CUDA-enabled systems.

--opt-split- Force-enables InvokeAI's cross-attention layer optimization.

None False
attention-invokeai By default, it's on when CUDA is unavailable.

--opt-split- Enable older version of split attention optimization that

None False
attention-v1 does not consume all VRAM available.

--opt-sub-quad- Enable memory efficient sub-quadratic cross-attention

None False
attention layer optimization.

--sub-quad-q- Query chunk size for the sub-quadratic cross-attention

SUB_QUAD_Q_CHUNK_SIZE 1024
chunk-size layer optimization to use.

--sub-quad-kv- KV chunk size for the sub-quadratic cross-attention layer

SUB_QUAD_KV_CHUNK_SIZE None
chunk-size optimization to use.

--sub-quad-chunk- The percentage of VRAM threshold for the sub-quadratic

SUB_QUAD_CHUNK_THRESHOLD None
threshold cross-attention layer optimization to use chunking.

Enable alternative layout for 4d tensors, may result in faster

--opt-channelslast None False inference only on Nvidia cards with Tensor cores (16xx and
higher).

--disable-opt-split-
None False Force-disables cross-attention layer optimization.
attention

Do not check if produced images/latent spaces have nans;

--disable-nan-check None False
useful for running without a checkpoint in CI.

{all, sd, interrogate, gfpgan, bsrgan,

--use-cpu None Use CPU as torch device for specified modules.
esrgan, scunet, codeformer}

--use-ipex None False Use Intel XPU as torch device

--no-half None False Do not switch the model to 16-bit floats.

--precision {full, half, autocast} autocast Evaluate at this precision.

--no-half-vae None False Do not switch the VAE model to 16-bit floats.

Upcast sampling. No effect with --no-half . Usually

--upcast-sampling None False produces similar results to --no-half with better
performance while using less memory.

Enable Stable Diffusion model optimizations for sacrificing

--medvram None False
a some performance for low VRAM usage.

--medvram-sdxl None False enable --medvram optimization just for SDXL models

Enable Stable Diffusion model optimizations for sacrificing

--lowvram None False
a lot of speed for very low VRAM usage.

Load Stable Diffusion checkpoint weights to VRAM instead

--lowram None False
of RAM.

--disable-model-
disable an optimization that reduces RAM use when
loading-ram- None False
loading a model
optimization

FEATURES

Open the web UI URL in the system's default browser upon

--autolaunch None False
launch.

Open the web UI with the specified theme

--theme None Unset ( light or dark ). If not specified, uses the default browser
theme.

Use textbox for seeds in UI (no up/down, but possible to

--use-textbox-seed None False
input long seeds).

https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Command-Line-Arguments-and-Settings#all-command-line-arguments 3/4
10/03/2025, 19:49 Command Line Arguments and Settings · AUTOMATIC1111/stable-diffusion-webui Wiki · GitHub

Argument
Value Default Description
Command

--disable-safe-
None False Disable checking PyTorch models for malicious code.
unpickle

--ngrok NGROK None ngrok authtoken, alternative to gradio --share .

--ngrok-region NGROK_REGION us The region in which ngrok should start.

The options to pass to ngrok in JSON format,

e.g.: {"authtoken_from_env":true,
--ngrok-options NGROK_OPTIONS None "basic_auth":"user:password",
"oauth_provider":"google",
"oauth_allow_emails":"[email protected]"}

On startup, notifies whether or not your web UI version

--update-check None None
(commit) is up-to-date with the current master branch.

--update-all- On startup, it pulls the latest updates for all extensions you
None None
extensions have installed.

Force-reinstall xformers. Useful for upgrading - but remove

--reinstall-xformers None False
it after upgrading or you'll reinstall xformers perpetually.

Force-reinstall torch. Useful for upgrading - but remove it

--reinstall-torch None False
after upgrading or you'll reinstall torch perpetually.

Run test to validate web UI functionality, see wiki topic for

--tests TESTS False
more details.

--no-tests None False Do not run tests even if --tests option is specified.

launch.py argument: dump limited sysinfo file (without

--dump-sysinfo None False
information about extensions, options) to disk and quit

--disable-all-
None False disable all non-built-in extensions from running
extensions

--disable-extra-
None False disable all extensions from running
extensions

--skip-load-model- if load a model at web start, only take effect when --

None False
at-start nowebui

--unix-filenames- allow any symbols except '/' in filenames. May conflict with
None False
sanitization your browser and file system

maximal length of filenames of saved images, longer

--filenames-max-
int 128 filenames will be truncated. if overridden it can potentially
length
cause issues with the file system

disable read prompt from last generation feature;

--no-prompt-history None False
disables --data-path/params.txt

DEFUNCT OPTIONS

--show-negative-
None False No longer has an effect.
prompt

--deepdanbooru None False No longer has an effect.

--unload-gfpgan None False No longer has an effect.

--gradio-img2img-
GRADIO_IMG2IMG_TOOL None No longer has an effect.
tool

--gradio-inpaint-
GRADIO_INPAINT_TOOL None No longer has an effect.
tool

--gradio-queue None False No longer has an effect.

--add-stop-route None False No longer has an effect.

--always-batch- No longer has an effect, move into UI under Setting >

None False
cond-uncond Optimizations

No longer has an effect. moved to ui-

--max-batch-count MAX_BATCH_COUNT 16 config.json txt2img/Batch count/maximum img2img/Batch
count/maximum User-Interface-Customizations

https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Command-Line-Arguments-and-Settings#all-command-line-arguments 4/4