-
-
Notifications
You must be signed in to change notification settings - Fork 432
CLI Arguments
Tip
All command line arguments can also be set as environment flags, for example --debug
is equivalent to SD_DEBUG=True
All options listed here are available as arguments to use from the command line or as environment variables, there's no need to do both.
webui --help
Configuration:
--backend {original,diffusers} force model pipeline type
--config CONFIG Use specific server configuration file, default: config.json
--ui-config UI_CONFIG Use specific UI configuration file, default: ui-config.json
--medvram Split model stages and keep only active part in VRAM, default: False
--lowvram Split model components and keep only active part in VRAM, default: False
--freeze Disable editing settings
Paths:
--ckpt CKPT Path to model checkpoint to load immediately, default: None
--data-dir DATA_DIR Base path where all user data is stored, default:
--models-dir MODELS_DIR Base path where all models are stored, default: models
Diagnostics:
--no-hashing Disable hashing of checkpoints, default: False
--no-metadata Disable reading of metadata from models, default: False
--disable-queue Disable queues, default: False
--device-id DEVICE_ID Select the default CUDA device to use, default: None
HTTP:
--server-name SERVER_NAME Sets hostname of server, default: None
--tls-keyfile TLS_KEYFILE Enable TLS and specify key file, default: None
--tls-certfile TLS_CERTFILE Enable TLS and specify cert file, default: None
--tls-selfsign Enable TLS with self-signed certificates, default: False
--cors-origins CORS_ORIGINS Allowed CORS origins as comma-separated list, default: None
--cors-regex CORS_REGEX Allowed CORS origins as regular expression, default: None
--subpath SUBPATH Customize the URL subpath for usage with reverse proxy
--autolaunch Open the UI URL in the system's default browser upon launch
--auth AUTH Set access authentication like "user:pwd,user:pwd""
--auth-file AUTH_FILE Set access authentication using file, default: None
--api-only Run in API only mode without starting UI
--allowed-paths ALLOWED_PATHS [ALLOWED_PATHS ...] add additional paths to paths allowed for web access
--share Enable UI accessible through Gradio site, default: False
--insecure Enable extensions tab regardless of other options, default: False
--listen Launch web server using public IP address, default: False
--port PORT Launch web server with given server port, default: 7860
Setup:
--reset Reset main repository to latest version, default: False
--upgrade, --update Upgrade main repository to latest version, default: False
--requirements Force re-check of requirements, default: False
--reinstall Force reinstallation of all requirements, default: False
--uv Use uv instead of pip to install the packages
Startup:
--quick Bypass version checks, default: False
--skip-requirements Skips checking and installing requirements, default: False
--skip-extensions Skips running individual extension installers, default: False
--skip-git Skips running all GIT operations, default: False
--skip-torch Skips running Torch checks, default: False
--skip-all Skips running all checks, default: False
--skip-env Skips setting of env variables during startup, default: False
Compute Engine:
--use-directml Use DirectML if no compatible GPU is detected, default: False
--use-openvino Use Intel OpenVINO backend, default: False
--use-ipex Force use Intel OneAPI XPU backend, default: False
--use-cuda Force use nVidia CUDA backend, default: False
--use-rocm Force use AMD ROCm backend, default: False
--use-zluda Force use ZLUDA, AMD GPUs only, default: False
--use-xformers Force use xFormers cross-optimization, default: False
Diagnostics:
--safe Run in safe mode with no user extensions
--experimental Allow unsupported versions of libraries, default: False
--test Run test only and exit
--version Print version information
--ignore Ignore any errors and attempt to continue
Logging:
--log LOG Set log file, default: None
--debug Run installer with debug logging, default: False
--profile Run profiler, default: False
--docs Mount API docs, default: False
--api-log Enable logging of all API requests, default: False
-
--config
: Specify the server configuration file. This option allows you to use a specific server configuration file. The default is set to<path to data>/config.json
. You can customize this by providing a different file path or by setting the environment variableSD_CONFIG
. -
--ui-config
: Specify the UI configuration file. This option allows you to use a specific UI configuration file. The default is set to<path to data>/ui-config.json
. You can customize this by providing a different file path or by setting the environment variableSD_UICONFIG
. -
--autolaunch
: Open the UI URL in the system's default browser upon launch. Enabling this flag (True
) automatically opens the UI URL in the system's default browser upon launch. The default isFalse
, and you can set it toTrue
using the environment variableSD_AUTOLAUNCH
. -
--backend
: Force model pipeline type. Enabling this option allows you to force a specific model pipeline type. The choices are ['original', 'diffusers']. The default isOriginal
, and you can set a custom backend by providing the argument value or by setting the environment variableSD_BACKEND
. -
--upgrade
: Upgrade the main repository to the latest version. Use this option when you want to ensure that you are using the most recent version of the application. Enabling this flag (True
) upgrades the main repository to the latest version. The default isFalse
, and you can set it toTrue
using the environment variableSD_UPGRADE
.
This is relatively safe to use with master branch. Dev branch is another story, be wary. -
--debug
: Run SDNext with debug logging to the console. Enabling this flag (True
) runs SDNext with debug logging. The default isFalse
, and you can set it toTrue
using the environment variableSD_DEBUG
.
The sdnext.log files always show debug output for troubleshooting purposes. -
--test
: Run the tests only and exit. This is useful for checking the integrity of the application without starting the actual service. Also useful for doing installation or reinstallation activities. Enabling this flag (True
) runs the application in test mode, executing tests and exiting without launching the full application. The default isFalse
, and you can set it toTrue
using the environment variableSD_TEST
.
Be careful setting this option as an environment variable as it will only test on every startup, you will never get inside SDNext.
-
--quick
: Run with startup sequence only, does not check requirements, extensions, git, or torch tests. Enabling this flag (True
) runs with the startup sequence only. The default isFalse
, and you can set it toTrue
using the environment variableSD_QUICK
. -
--skip-requirements
: Skips checking and installing requirements. Enabling this flag (True
) during setup skips checking and installing requirements. The default isFalse
, and you can set it toTrue
using the environment variableSD_SKIPREQUIREMENTS
. -
--skip-extensions
: Skips running individual extension installers. Enabling this flag (True
) during setup skips running individual extension installers. The default isFalse
, and you can set it toTrue
using the environment variableSD_SKIPEXTENSION
. -
--skip-git
: Skips running all GIT operations. Enabling this flag (True
) during setup skips running all GIT operations. The default isFalse
, and you can set it toTrue
using the environment variableSD_SKIPGIT
. -
--skip-torch
: Skips running Torch checks. Enabling this flag (True
) during setup skips running Torch checks. The default isFalse
, and you can set it toTrue
using the environment variableSD_SKIPTORCH
. -
--skip-env
: Skips setting env variables. Enabling this flag (True
) during setup skips setting of any and all env variables used for tuning. The default isFalse
, and you can set it toTrue
using the environment variableSD_SKIPENV
.
-
--medvram
: Split model stages and keep only the active part in VRAM. Enabling this flag (True
) allows the application to split model stages, conserving GPU memory by keeping only the active part in VRAM. The default isFalse
, and you can set it toTrue
using the environment variableSD_MEDVRAM
. -
--lowvram
: Split model components and keep only the active part in VRAM. Enabling this flag (True
) allows the application to split model components. The default isFalse
, and you can set it toTrue
using the environment variableSD_LOWVRAM
.
-
--use-directml
: Use DirectML if no compatible GPU is detected. Enabling this flag (True
) allows the use of DirectML if no compatible GPU is detected. The default isFalse
, and you can set it toTrue
using the environment variableSD_USEDIRECTML
. -
--use-openvino
: Use Intel OpenVINO backend. Enabling this flag (True
) allows the use of the Intel OpenVINO backend. The default isFalse
, and you can set it toTrue
using the environment variableSD_USEOPENVINO
. -
--use-ipex
: Force use Intel OneAPI XPU backend. Enabling this flag (True
) forces the use of the Intel OneAPI XPU backend. The default isFalse
, and you can set it toTrue
using the environment variableSD_USEIPEX
. -
--use-cuda
: Force use NVIDIA CUDA backend. Enabling this flag (True
) forces the use of the NVIDIA CUDA backend. The default isFalse
, and you can set it toTrue
using the environment variableSD_USECUDA
. -
--use-rocm
: Force use AMD ROCm backend. Enabling this flag (True
) forces the use of the AMD ROCm backend. The default isFalse
, and you can set it toTrue
using the environment variableSD_USEROCM
. -
--use-zluda
: Force use ZLUDA backend. Enabling this flag (True
) forces the use of the AMD ROCm backend wrapped with ZLUDA. The default isFalse
and you can set it toTrue
using the environment variableSD_USEZLUDA
. You should not havetorch
installed before enabling this flag.
-
DISABLE_IPEXRUN
: Disable ipexrun for Intel ARC. If you encounter memory pointer or SIGSEGV-related errors on Linux, you can try setting this environment variable toTrue
to disable ipexrun. -
IPEX_SDPA_SLICE_TRIGGER_RATE
: Specify when dynamic attention slicing for Scaled Dot Product Attention should get triggered for Intel ARC. This environment variable allows you to set the trigger rate in gigabytes (GB). The default is6
GB. -
IPEX_ATTENTION_SLICE_RATE
: Specify the dynamic attention slicing rate for Intel ARC. This environment variable allows you to set the slicing rate in gigabytes (GB). The default is4
GB. -
IPEX_FORCE_ATTENTION_SLICE
: Force use dynamic attention slicing even if the GPU supports 64 bit. Useful with Intel Data Center GPU MAX series.
-
--log
: Set log file name and path. This argument allows you to set the log filename and location. The default issdnext.log
in the base directory, and you can set it with the environment variableSD_LOG
. -
--ckpt
: Path to the model checkpoint to load immediately. This option allows you to specify the path to a model checkpoint for immediate loading. The default isNone
, and you can set a custom path by providing the argument value or by setting the environment variableSD_MODEL
. -
--vae
: Path to the VAE checkpoint to load immediately. This option allows you to specify the path to a VAE checkpoint for immediate loading. The default isNone
, and you can set a custom path by providing the argument value or by setting the environment variableSD_VAE
. -
--data-dir
: Base path where all user data is stored. You can set the base path where all user data is stored using this option. The default is an empty string (''
). Customize this by providing a different path or by setting the environment variableSD_DATADIR
. -
--models-dir
: Base path where all models are stored. This option sets the base path where all models are stored. The default is'models'
. Customize this by providing a different path or by setting the environment variableSD_MODELSDIR
.
Also see Troubleshooting Wiki page
-
--safe
: Run in safe mode with no user extensions. Safe mode can be useful when troubleshooting or when you want to restrict the execution of potentially unsafe user-provided code. Enabling this flag (True
) runs the application in safe mode, disabling user extensions. The default isFalse
, and you can set it toTrue
using the environment variableSD_SAFE
.
Try this first. -
--requirements
: Force re-check of python (pip) package requirements and installs any that are not up-to-date. Enabling this flag (True
) forces a re-check of requirements. The default isFalse
, and you can set it toTrue
using the environment variableSD_REQUIREMENTS
.
Try this second. -
--reinstall
: Force reinstallation of all requirements. Use this option when you want to ensure that all dependencies are freshly installed, potentially resolving any issues related to outdated or corrupted installations. Also useful to change hardware backends, such as to OpenVINO or DML. Enabling this flag (True
) forces the reinstallation of all requirements. The default isFalse
, and you can set it toTrue
using the environment variableSD_REINSTALL
.
Try this third.
Be careful setting this option as an environment variable as it will reinstall on every startup. -
--reset
: Reset main repository to latest version. Enabling this flag (True
) resets the main repository to the latest version. The default isFalse
, and you can set it toTrue
using the environment variableSD_RESET
.
This is more serious with consequences, but you can try this fourth.
Be careful setting this option as an environment variable as it will reset on every startup. -
--experimental
: Allow unsupported versions of libraries. This is useful for testing or trying out features that may not be officially supported yet. Enabling this flag (True
) allows the application to use unsupported versions of libraries. The default isFalse
, and you can set it toTrue
using the environment variableSD_EXPERIMENTAL
.
Be careful setting this option as an environment variable as it can easily cause issues. -
--ignore
: Try to ignore any errors and attempt to continue. This can be useful in scenarios where certain errors are known and can be safely bypassed. Enabling this flag (True
) instructs the application to ignore any errors encountered during the setup process and attempt to continue. The default isFalse
, and you can set it toTrue
using the environment variableSD_IGNORE
.
-
--device-id
: Select the default CUDA device to use. This option allows you to select the default CUDA device for GPU operations. The default isNone
, and you can set a custom device ID by providing the argument value or by setting the environment variableSD_DEVICEID
. -
--use-xformers
: Forces the installation and use of xFormers cross-optimization. Enabling this flag (True
) forces the installation and use of xFormers. The default isFalse
, and you can set it toTrue
using the environment variableSD_USEXFORMERS
. -
--no-hashing
: Disable hashing of checkpoints. Enabling this flag (True
) disables the hashing of checkpoints. The default isFalse
, and you can set it toTrue
using the environment variableSD_NOHASHING
. -
--no-metadata
: Disable reading of metadata from models. Enabling this flag (True
) disables the reading of metadata from models. The default isFalse
, and you can set it toTrue
using the environment variableSD_NOMETADATA
. -
--profile
: Run profiler. Enabling this flag (True
) runs the profiler. The default isFalse
, and you can set it toTrue
using the environment variableSD_PROFILE
. -
--disable-queue
: Disable queues. Enabling this flag (True
) disables the use of queues. The default isFalse
, and you can set it toTrue
using the environment variableSD_DISABLEQUEUE
. -
--allow-code
: Allow custom script execution. This is useful for scenarios where users may want to run their own code. Enabling this flag (True
) allows the execution of custom scripts. The default isFalse
, and you can set it toTrue
using the environment variableSD_ALLOWCODE
. -
--use-cpu
: This option forces the use of CPU for specified modules. You can provide a list of modules as arguments. The default is an empty list ([]
). Customize this by providing module names or by setting the environment variableSD_USECPU
.
Not advised, legacy code with poor performance. -
--freeze
: Disable editing settings. This is useful to lock down configurations. Enabling this flag (True
) prevents editing of settings. The default isFalse
, and you can set it toTrue
using the environment variableSD_FREEZE
.
-
--docs
: Mount API docs at /docs i.e.,https://127.0.0.1/docs
. Enabling this flag (True
) mounts API documentation at the/docs
endpoint. The default isFalse
, and you can set it toTrue
using the environment variableSD_DOCS
. -
--api-only
: Run in API only mode without starting UI. Enabling this flag (True
) runs the application in API-only mode without starting the UI. The default isFalse
, and you can set it toTrue
using the environment variableSD_APIONLY
. -
--api-log
: Enable logging of all API requests. Enabling this flag (True
) logs all API requests. The default isFalse
, and you can set it toTrue
using the environment variableSD_APILOG
.
-
--share
: Enable UI to be accessible through the Gradio site. This is useful for sharing your SDNext with others. Enabling this flag (True
) allows the UI to be accessible through the Gradio site. The default isFalse
, and you can set it toTrue
using the environment variableSD_SHARE
.
The shared URL will be in your console log. -
--insecure
: Enable extensions tab regardless of other options. This should only be used when you want local network or web accessible control of your extensions, potentially dangerous on the web. Enabling this flag (True
) allows the extensions tab to be enabled regardless of other specified options. The default isFalse
, and you can set it toTrue
using the environment variableSD_INSECURE
.
Use with care when using--share
. -
--listen
: Launch web server to be accessible from local network. Enabling this flag (True
) allows the web server to launch for use on your own network. The default isFalse
, and you can set it toTrue
using the environment variableSD_LISTEN
. -
--auth
: Set access authentication like "user:pwd,user:pwd". This option allows you to set access authentication with a specified username and password combination. The default isNone
, and you can set a custom authentication string by providing the argument value or by setting the environment variableSD_AUTH
. -
--auth-file
: Set access authentication using file. This option allows you to set access authentication using a file. The default isNone
, and you can set a custom file path by providing the argument value or by setting the environment variableSD_AUTHFILE
. -
--server-name
: Sets hostname of server. This option sets the hostname of the server. The default isNone
, and you can set a custom server name by providing the argument value or by setting the environment variableSD_SERVERNAME
. -
--subpath
: Customize the URL subpath for usage with reverse proxy. This option allows you to customize the URL subpath for usage with a reverse proxy. The default isNone
, and you can set a custom subpath by providing the argument value or by setting the environment variableSD_SUBPATH
. -
--cors-origins
: Allowed CORS origins as comma-separated list. This option sets the allowed CORS origins as a comma-separated list. The default isNone
, and you can set custom origins by providing the argument value or by setting the environment variableSD_CORSORIGINS
. -
--cors-regex
: Allowed CORS origins as regular expression. This option sets the allowed CORS origins as a regular expression. The default isNone
, and you can set a custom regular expression by providing the argument value or by setting the environment variableSD_CORSREGEX
. -
--tls-keyfile
: Enable TLS and specify key file. This option enables TLS (Transport Layer Security) and specifies the key file. The default isNone
, and you can set a custom key file path by providing the argument value or by setting the environment variableSD_TLSKEYFILE
. -
--tls-certfile
: Enable TLS and specify cert file. This option enables TLS (Transport Layer Security) and specifies the certificate file. The default isNone
, and you can set a custom certificate file path by providing the argument value or by setting the environment variableSD_TLSCERTFILE
. -
--tls-selfsign
: Enable TLS with self-signed certificates. Enabling this flag (True
) enables TLS with self-signed certificates. The default isFalse
, and you can set it toTrue
using the environment variableSD_TLSSELFSIGN
.
These CORS options are important for controlling which domains are permitted to access the resources of your server. It helps in enhancing the security of your web application by preventing unauthorized cross-origin requests. When configuring these options, ensure that you only allow origins that you trust to interact with your server, as allowing any origin (*
) can introduce security vulnerabilities. Always specify the origins explicitly or using a secure regular expression pattern.
-
--cors-origins
: Allowed CORS origins as a comma-separated list. CORS is a security feature implemented by web browsers that restricts web pages from making requests to a different domain than the one that served the web page. This option allows you to specify a list of origins (domains) that are allowed to access resources on your server. For example, if your application is hosted onhttp://example.com
, and you want to allow access fromhttp://client.example.com
, you would set--cors-origins http://client.example.com
.
For local network use, often with other tools, you can just set--cors-origins *
. -
--cors-regex
: Allowed CORS origins as a regular expression. This option provides more flexibility by allowing you to specify CORS origins using a regular expression. The regular expression should match the origin(s) you want to allow. This is useful when you have a dynamic set of origins that follow a certain pattern. For instance, if you want to allow any subdomain underexample.com
, you could set--cors-regex "^https?://[a-z0-9-]+\.example\.com$"
.
When setting up TLS, it's important to use valid certificates from a trusted CA in a production environment to ensure secure and encrypted communication. In a development or testing environment, self-signed certificates can be used, but users should be cautious when accessing the application as browsers may show security warnings due to the self-signed nature. Always ensure that your TLS setup meets security best practices.
-
--tls-keyfile
: Enable TLS and specify the key file. TLS is a cryptographic protocol that ensures the secure transmission of data over a network. Enabling TLS in your application secures the communication between the client and the server. The--tls-keyfile
option allows you to specify the path to the private key file used for encryption. This private key file should be in PEM format.Example:
--tls-keyfile /path/to/private-key.pem
-
--tls-certfile
: Enable TLS and specify the cert file. Along with the private key, you need to specify the TLS certificate file. The certificate file contains the public key and information about the server. It should be in PEM format.Example:
--tls-certfile /path/to/certificate.pem
-
--tls-selfsign
: Enable TLS with self-signed certificates. If you don't have a certificate issued by a Certificate Authority (CA), you can enable this option to create self-signed certificates. Self-signed certificates are useful for development and testing but should not be used in production environments where security is crucial.Example:
--tls-selfsign /path/to/self-signed-certificate.pem
© SD.Next