up

.dockerignore

@@ -0,0 +1,2 @@
+**/input
+**/output

.gitignore

@@ -0,0 +1,3 @@
+/input
+/output
+token.txt

CHANGELOG.md

@@ -0,0 +1,386 @@
+# Changelog
+
+## [v1.34.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.33.0...v1.34.0) - 2023-03-17
+
+### Added
+
+* Add test for custom model and vae tiling
+* Add vae tiling for high resolution images
+* Update diffusers, onnx, and safetensors
+
+## [v1.33.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.32.0...v1.33.0) - 2023-02-25
+
+### Added
+
+* Update diffusers to 0.13.1
+
+### Fixed
+
+* Ensure booleans do not consume other args
+
+## [v1.32.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.31.0...v1.32.0) - 2023-02-20
+
+### Added
+
+* Ensure onnx option is tested
+* Split onnx into separate option
+
+## [v1.31.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.30.1...v1.31.0) - 2023-02-17
+
+### Added
+
+* Add test for instruct pix2pix
+* Add instruct pix2pix pipeline
+* Update transformers to 4.26.1
+* Update onnxruntime to 1.14.0
+
+### Changed
+
+* Add full-sized image for testing
+
+## [v1.30.1](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.30.0...v1.30.1) - 2023-02-03
+
+### Fixed
+
+* Switch xformers to 0.0.16 stable
+
+## [v1.30.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.29.0...v1.30.0) - 2023-01-28
+
+### Added
+
+* Update diffusers to 0.12.1
+
+### Changed
+
+* Reorder arguments alphabetically
+* Rename option names in tests
+* Rename certain options for ease-of-use
+
+### Fixed
+
+* Download test image since img folder is gone
+* Rename option names in depth diffusion test
+
+## [v1.29.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.28.0...v1.29.0) - 2023-01-26
+
+### Added
+
+* Add test for depth-guided diffusion
+* Add depth-guided stable diffusion
+
+## [v1.28.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.27.0...v1.28.0) - 2023-01-26
+
+### Added
+
+* Upgrade diffusers to 0.12.0
+
+### Fixed
+
+* Suppress CLIPFeatureExtractor warning
+
+## [v1.27.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.26.0...v1.27.0) - 2023-01-16
+
+### Added
+
+* Automatically publish new versions
+
+### Fixed
+
+* Update unstable xformers to 0.0.16rc425
+
+## [v1.26.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.25.1...v1.26.0) - 2023-01-13
+
+### Added
+
+* Support pulling image from ghcr
+* Publish docker pipeline to ghcr
+
+### Changed
+
+* Simplify ghcr publish action
+* Move ghcr url into variable
+
+## [v1.25.1](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.25.0...v1.25.1) - 2023-01-10
+
+### Fixed
+
+* Update safetensors and unstable xformers
+* Add numpy 1.23.5 to fix float errors
+
+## [v1.25.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.24.0...v1.25.0) - 2022-12-26
+
+### Added
+
+* Add test for scheduler option
+* Allow different schedulers to be used
+* Update torch to 1.13.1 and unstable xformers
+
+## [v1.24.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.23.0...v1.24.0) - 2022-12-20
+
+### Added
+
+* Add tests for stable diffusion 2.0 and 2.1
+* Update diffusers to 0.11.1
+* Switch to python slim and halve image size
+
+## [v1.23.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.22.0...v1.23.0) - 2022-12-14
+
+### Added
+
+* Add memory efficient transformers
+* Add test for image upscaling
+
+### Fixed
+
+* Ensure default test produces output
+
+## [v1.22.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.21.0...v1.22.0) - 2022-12-12
+
+### Added
+
+* Add upscaler pipeline
+* Update diffusers to 0.10.2
+
+## [v1.21.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.20.0...v1.21.0) - 2022-12-09
+
+### Added
+
+* Update diffusers to 0.10.1
+* Update diffusers to 0.10.0
+* Update transformers to 4.25.1
+
+### Fixed
+
+* Remove use of autocast
+* Remove unused arguments from pipeline
+* Return after setting gpu arg
+* Removed unused imports
+
+## [v1.20.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.19.1...v1.20.0) - 2022-11-29
+
+### Added
+
+* Update diffusers to 0.9.0
+* Update tensorflow to 2.11.0
+
+### Fixed
+
+* Ensure skip option works on onnx pipeline
+
+## [v1.19.1](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.19.0...v1.19.1) - 2022-11-22
+
+### Added
+
+* Test image to image pipeline
+
+### Fixed
+
+* Do not pass gpu arg if using cpu pipeline
+
+## [v1.19.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.18.0...v1.19.0) - 2022-11-17
+
+### Added
+
+* Add test for cpu pipeline
+* Add cpu pipeline using onnx
+
+### Changed
+
+* Test image with different dimensions
+
+## [v1.18.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.17.0...v1.18.0) - 2022-11-11
+
+### Added
+
+* Add diffusion inpainting
+
+## [v1.17.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.16.0...v1.17.0) - 2022-11-08
+
+### Added
+
+* Add image-to-image diffusion
+* Add input folder for image-to-image diffusion
+
+### Changed
+
+* Call run command directly from tests
+
+### Fixed
+
+* Ensure output folder is always created
+
+## [v1.16.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.15.0...v1.16.0) - 2022-11-07
+
+### Added
+
+* Add standard tests for all options
+* Upgrade diffusers to 0.7.2
+
+## [v1.15.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.14.0...v1.15.0) - 2022-11-04
+
+### Added
+
+* Add negative prompts
+
+### Changed
+
+* Switch to using num_images_per_prompt
+
+## [v1.14.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.13.0...v1.14.0) - 2022-11-03
+
+### Added
+
+* Upgrade diffusers to 0.7.0
+
+## [v1.13.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.12.0...v1.13.0) - 2022-11-02
+
+### Added
+
+* Upgrade torch to 1.13.0 and cuda to 11.7
+
+## [v1.12.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.11.1...v1.12.0) - 2022-11-01
+
+### Added
+
+* Update transformers to v4.24.0
+
+### Fixed
+
+* Always use torch instead of tensorflow
+
+## [v1.11.1](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.11.0...v1.11.1) - 2022-10-26
+
+### Changed
+
+* Split pipeline and inference
+
+## [v1.11.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.10.0...v1.11.0) - 2022-10-21
+
+### Added
+
+* Make model configurable
+
+## [v1.10.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.9.0...v1.10.0) - 2022-10-19
+
+### Added
+
+* Upgrade to diffusers 0.6.0
+
+### Fixed
+
+* Switch output and image names
+
+## [v1.9.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.8.1...v1.9.0) - 2022-10-14
+
+### Added
+
+* Upgrade to latest huggingface releases
+
+## [v1.8.1](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.8.0...v1.8.1) - 2022-10-13
+
+### Fixed
+
+* Remove cuda stubs to prevent errors
+
+## [v1.8.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.7.0...v1.8.0) - 2022-10-12
+
+### Added
+
+* Upgrade to diffusers 0.4.2
+
+## [v1.7.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.6.0...v1.7.0) - 2022-10-07
+
+### Added
+
+* Upgrade to diffusers 0.4.1
+* Pin dependencies using requirements.txt
+
+## [v1.6.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.5.0...v1.6.0) - 2022-09-14
+
+### Added
+
+* Add attention slicing
+
+### Changed
+
+* Use tagged version of tensorflow
+
+## [v1.5.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.4.0...v1.5.0) - 2022-09-08
+
+### Added
+
+* Specify user access token at command line
+
+## [v1.4.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.3.0...v1.4.0) - 2022-09-05
+
+### Added
+
+* Add pipeline iteration
+
+### Changed
+
+* Move image name out of loop
+
+## [v1.3.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.2.2...v1.3.0) - 2022-09-02
+
+### Added
+
+* Support skipping safety checker
+
+## [v1.2.2](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.2.1...v1.2.2) - 2022-09-01
+
+### Changed
+
+* Rename iso date time function
+
+### Fixed
+
+* Prevent errors when file name is too long
+
+## [v1.2.1](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.2.0...v1.2.1) - 2022-08-29
+
+### Fixed
+
+* Allow full range of random seeds
+
+## [v1.2.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.1.1...v1.2.0) - 2022-08-26
+
+### Added
+
+* Support half-sized (float16) tensors
+
+### Changed
+
+* Update scale argument description
+
+### Fixed
+
+* Remove unused sys import
+
+## [v1.1.1](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.1.0...v1.1.1) - 2022-08-25
+
+### Fixed
+
+* Double quote to prevent globbing
+* Only output two digits of precision
+
+## [v1.1.0](https://github.com/fboulnois/stable-diffusion-docker/compare/v1.0.0...v1.1.0) - 2022-08-24
+
+### Added
+
+* Add a subset of txt2img.py options
+* Include model params in image filename
+
+### Changed
+
+* Use enumerate instead of index
+
+### Fixed
+
+* Remove unnecessary image library import
+
+## [v1.0.0](https://github.com/fboulnois/stable-diffusion-docker/releases/tag/v1.0.0) - 2022-08-22
+
+### Added
+
+* Initial release

Dockerfile

@@ -0,0 +1,25 @@
+FROM python:3.10-slim-bullseye
+
+RUN rm -rf /usr/local/cuda/lib64/stubs
+
+COPY requirements.txt /
+
+RUN pip install -r requirements.txt \
+  --extra-index-url https://download.pytorch.org/whl/cu117
+
+RUN useradd -m huggingface
+
+USER huggingface
+
+WORKDIR /home/huggingface
+
+ENV USE_TORCH=1
+
+RUN mkdir -p /home/huggingface/.cache/huggingface \
+  && mkdir -p /home/huggingface/input \
+  && mkdir -p /home/huggingface/output
+
+COPY docker-entrypoint.py /usr/local/bin
+COPY token.txt /home/huggingface
+
+ENTRYPOINT [ "docker-entrypoint.py" ]

LICENSE

@@ -0,0 +1,661 @@
+                    GNU AFFERO GENERAL PUBLIC LICENSE
+                       Version 3, 19 November 2007
+
+ Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The GNU Affero General Public License is a free, copyleft license for
+software and other kinds of works, specifically designed to ensure
+cooperation with the community in the case of network server software.
+
+  The licenses for most software and other practical works are designed
+to take away your freedom to share and change the works.  By contrast,
+our General Public Licenses are intended to guarantee your freedom to
+share and change all versions of a program--to make sure it remains free
+software for all its users.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+them if you wish), that you receive source code or can get it if you
+want it, that you can change the software or use pieces of it in new
+free programs, and that you know you can do these things.
+
+  Developers that use our General Public Licenses protect your rights
+with two steps: (1) assert copyright on the software, and (2) offer
+you this License which gives you legal permission to copy, distribute
+and/or modify the software.
+
+  A secondary benefit of defending all users' freedom is that
+improvements made in alternate versions of the program, if they
+receive widespread use, become available for other developers to
+incorporate.  Many developers of free software are heartened and
+encouraged by the resulting cooperation.  However, in the case of
+software used on network servers, this result may fail to come about.
+The GNU General Public License permits making a modified version and
+letting the public access it on a server without ever releasing its
+source code to the public.
+
+  The GNU Affero General Public License is designed specifically to
+ensure that, in such cases, the modified source code becomes available
+to the community.  It requires the operator of a network server to
+provide the source code of the modified version running there to the
+users of that server.  Therefore, public use of a modified version, on
+a publicly accessible server, gives the public access to the source
+code of the modified version.
+
+  An older license, called the Affero General Public License and
+published by Affero, was designed to accomplish similar goals.  This is
+a different license, not a version of the Affero GPL, but Affero has
+released a new version of the Affero GPL which permits relicensing under
+this license.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                       TERMS AND CONDITIONS
+
+  0. Definitions.
+
+  "This License" refers to version 3 of the GNU Affero General Public License.
+
+  "Copyright" also means copyright-like laws that apply to other kinds of
+works, such as semiconductor masks.
+
+  "The Program" refers to any copyrightable work licensed under this
+License.  Each licensee is addressed as "you".  "Licensees" and
+"recipients" may be individuals or organizations.
+
+  To "modify" a work means to copy from or adapt all or part of the work
+in a fashion requiring copyright permission, other than the making of an
+exact copy.  The resulting work is called a "modified version" of the
+earlier work or a work "based on" the earlier work.
+
+  A "covered work" means either the unmodified Program or a work based
+on the Program.
+
+  To "propagate" a work means to do anything with it that, without
+permission, would make you directly or secondarily liable for
+infringement under applicable copyright law, except executing it on a
+computer or modifying a private copy.  Propagation includes copying,
+distribution (with or without modification), making available to the
+public, and in some countries other activities as well.
+
+  To "convey" a work means any kind of propagation that enables other
+parties to make or receive copies.  Mere interaction with a user through
+a computer network, with no transfer of a copy, is not conveying.
+
+  An interactive user interface displays "Appropriate Legal Notices"
+to the extent that it includes a convenient and prominently visible
+feature that (1) displays an appropriate copyright notice, and (2)
+tells the user that there is no warranty for the work (except to the
+extent that warranties are provided), that licensees may convey the
+work under this License, and how to view a copy of this License.  If
+the interface presents a list of user commands or options, such as a
+menu, a prominent item in the list meets this criterion.
+
+  1. Source Code.
+
+  The "source code" for a work means the preferred form of the work
+for making modifications to it.  "Object code" means any non-source
+form of a work.
+
+  A "Standard Interface" means an interface that either is an official
+standard defined by a recognized standards body, or, in the case of
+interfaces specified for a particular programming language, one that
+is widely used among developers working in that language.
+
+  The "System Libraries" of an executable work include anything, other
+than the work as a whole, that (a) is included in the normal form of
+packaging a Major Component, but which is not part of that Major
+Component, and (b) serves only to enable use of the work with that
+Major Component, or to implement a Standard Interface for which an
+implementation is available to the public in source code form.  A
+"Major Component", in this context, means a major essential component
+(kernel, window system, and so on) of the specific operating system
+(if any) on which the executable work runs, or a compiler used to
+produce the work, or an object code interpreter used to run it.
+
+  The "Corresponding Source" for a work in object code form means all
+the source code needed to generate, install, and (for an executable
+work) run the object code and to modify the work, including scripts to
+control those activities.  However, it does not include the work's
+System Libraries, or general-purpose tools or generally available free
+programs which are used unmodified in performing those activities but
+which are not part of the work.  For example, Corresponding Source
+includes interface definition files associated with source files for
+the work, and the source code for shared libraries and dynamically
+linked subprograms that the work is specifically designed to require,
+such as by intimate data communication or control flow between those
+subprograms and other parts of the work.
+
+  The Corresponding Source need not include anything that users
+can regenerate automatically from other parts of the Corresponding
+Source.
+
+  The Corresponding Source for a work in source code form is that
+same work.
+
+  2. Basic Permissions.
+
+  All rights granted under this License are granted for the term of
+copyright on the Program, and are irrevocable provided the stated
+conditions are met.  This License explicitly affirms your unlimited
+permission to run the unmodified Program.  The output from running a
+covered work is covered by this License only if the output, given its
+content, constitutes a covered work.  This License acknowledges your
+rights of fair use or other equivalent, as provided by copyright law.
+
+  You may make, run and propagate covered works that you do not
+convey, without conditions so long as your license otherwise remains
+in force.  You may convey covered works to others for the sole purpose
+of having them make modifications exclusively for you, or provide you
+with facilities for running those works, provided that you comply with
+the terms of this License in conveying all material for which you do
+not control copyright.  Those thus making or running the covered works
+for you must do so exclusively on your behalf, under your direction
+and control, on terms that prohibit them from making any copies of
+your copyrighted material outside their relationship with you.
+
+  Conveying under any other circumstances is permitted solely under
+the conditions stated below.  Sublicensing is not allowed; section 10
+makes it unnecessary.
+
+  3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+
+  No covered work shall be deemed part of an effective technological
+measure under any applicable law fulfilling obligations under article
+11 of the WIPO copyright treaty adopted on 20 December 1996, or
+similar laws prohibiting or restricting circumvention of such
+measures.
+
+  When you convey a covered work, you waive any legal power to forbid
+circumvention of technological measures to the extent such circumvention
+is effected by exercising rights under this License with respect to
+the covered work, and you disclaim any intention to limit operation or
+modification of the work as a means of enforcing, against the work's
+users, your or third parties' legal rights to forbid circumvention of
+technological measures.
+
+  4. Conveying Verbatim Copies.
+
+  You may convey verbatim copies of the Program's source code as you
+receive it, in any medium, provided that you conspicuously and
+appropriately publish on each copy an appropriate copyright notice;
+keep intact all notices stating that this License and any
+non-permissive terms added in accord with section 7 apply to the code;
+keep intact all notices of the absence of any warranty; and give all
+recipients a copy of this License along with the Program.
+
+  You may charge any price or no price for each copy that you convey,
+and you may offer support or warranty protection for a fee.
+
+  5. Conveying Modified Source Versions.
+
+  You may convey a work based on the Program, or the modifications to
+produce it from the Program, in the form of source code under the
+terms of section 4, provided that you also meet all of these conditions:
+
+    a) The work must carry prominent notices stating that you modified
+    it, and giving a relevant date.
+
+    b) The work must carry prominent notices stating that it is
+    released under this License and any conditions added under section
+    7.  This requirement modifies the requirement in section 4 to
+    "keep intact all notices".
+
+    c) You must license the entire work, as a whole, under this
+    License to anyone who comes into possession of a copy.  This
+    License will therefore apply, along with any applicable section 7
+    additional terms, to the whole of the work, and all its parts,
+    regardless of how they are packaged.  This License gives no
+    permission to license the work in any other way, but it does not
+    invalidate such permission if you have separately received it.
+
+    d) If the work has interactive user interfaces, each must display
+    Appropriate Legal Notices; however, if the Program has interactive
+    interfaces that do not display Appropriate Legal Notices, your
+    work need not make them do so.
+
+  A compilation of a covered work with other separate and independent
+works, which are not by their nature extensions of the covered work,
+and which are not combined with it such as to form a larger program,
+in or on a volume of a storage or distribution medium, is called an
+"aggregate" if the compilation and its resulting copyright are not
+used to limit the access or legal rights of the compilation's users
+beyond what the individual works permit.  Inclusion of a covered work
+in an aggregate does not cause this License to apply to the other
+parts of the aggregate.
+
+  6. Conveying Non-Source Forms.
+
+  You may convey a covered work in object code form under the terms
+of sections 4 and 5, provided that you also convey the
+machine-readable Corresponding Source under the terms of this License,
+in one of these ways:
+
+    a) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by the
+    Corresponding Source fixed on a durable physical medium
+    customarily used for software interchange.
+
+    b) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by a
+    written offer, valid for at least three years and valid for as
+    long as you offer spare parts or customer support for that product
+    model, to give anyone who possesses the object code either (1) a
+    copy of the Corresponding Source for all the software in the
+    product that is covered by this License, on a durable physical
+    medium customarily used for software interchange, for a price no
+    more than your reasonable cost of physically performing this
+    conveying of source, or (2) access to copy the
+    Corresponding Source from a network server at no charge.
+
+    c) Convey individual copies of the object code with a copy of the
+    written offer to provide the Corresponding Source.  This
+    alternative is allowed only occasionally and noncommercially, and
+    only if you received the object code with such an offer, in accord
+    with subsection 6b.
+
+    d) Convey the object code by offering access from a designated
+    place (gratis or for a charge), and offer equivalent access to the
+    Corresponding Source in the same way through the same place at no
+    further charge.  You need not require recipients to copy the
+    Corresponding Source along with the object code.  If the place to
+    copy the object code is a network server, the Corresponding Source
+    may be on a different server (operated by you or a third party)
+    that supports equivalent copying facilities, provided you maintain
+    clear directions next to the object code saying where to find the
+    Corresponding Source.  Regardless of what server hosts the
+    Corresponding Source, you remain obligated to ensure that it is
+    available for as long as needed to satisfy these requirements.
+
+    e) Convey the object code using peer-to-peer transmission, provided
+    you inform other peers where the object code and Corresponding
+    Source of the work are being offered to the general public at no
+    charge under subsection 6d.
+
+  A separable portion of the object code, whose source code is excluded
+from the Corresponding Source as a System Library, need not be
+included in conveying the object code work.
+
+  A "User Product" is either (1) a "consumer product", which means any
+tangible personal property which is normally used for personal, family,
+or household purposes, or (2) anything designed or sold for incorporation
+into a dwelling.  In determining whether a product is a consumer product,
+doubtful cases shall be resolved in favor of coverage.  For a particular
+product received by a particular user, "normally used" refers to a
+typical or common use of that class of product, regardless of the status
+of the particular user or of the way in which the particular user
+actually uses, or expects or is expected to use, the product.  A product
+is a consumer product regardless of whether the product has substantial
+commercial, industrial or non-consumer uses, unless such uses represent
+the only significant mode of use of the product.
+
+  "Installation Information" for a User Product means any methods,
+procedures, authorization keys, or other information required to install
+and execute modified versions of a covered work in that User Product from
+a modified version of its Corresponding Source.  The information must
+suffice to ensure that the continued functioning of the modified object
+code is in no case prevented or interfered with solely because
+modification has been made.
+
+  If you convey an object code work under this section in, or with, or
+specifically for use in, a User Product, and the conveying occurs as
+part of a transaction in which the right of possession and use of the
+User Product is transferred to the recipient in perpetuity or for a
+fixed term (regardless of how the transaction is characterized), the
+Corresponding Source conveyed under this section must be accompanied
+by the Installation Information.  But this requirement does not apply
+if neither you nor any third party retains the ability to install
+modified object code on the User Product (for example, the work has
+been installed in ROM).
+
+  The requirement to provide Installation Information does not include a
+requirement to continue to provide support service, warranty, or updates
+for a work that has been modified or installed by the recipient, or for
+the User Product in which it has been modified or installed.  Access to a
+network may be denied when the modification itself materially and
+adversely affects the operation of the network or violates the rules and
+protocols for communication across the network.
+
+  Corresponding Source conveyed, and Installation Information provided,
+in accord with this section must be in a format that is publicly
+documented (and with an implementation available to the public in
+source code form), and must require no special password or key for
+unpacking, reading or copying.
+
+  7. Additional Terms.
+
+  "Additional permissions" are terms that supplement the terms of this
+License by making exceptions from one or more of its conditions.
+Additional permissions that are applicable to the entire Program shall
+be treated as though they were included in this License, to the extent
+that they are valid under applicable law.  If additional permissions
+apply only to part of the Program, that part may be used separately
+under those permissions, but the entire Program remains governed by
+this License without regard to the additional permissions.
+
+  When you convey a copy of a covered work, you may at your option
+remove any additional permissions from that copy, or from any part of
+it.  (Additional permissions may be written to require their own
+removal in certain cases when you modify the work.)  You may place
+additional permissions on material, added by you to a covered work,
+for which you have or can give appropriate copyright permission.
+
+  Notwithstanding any other provision of this License, for material you
+add to a covered work, you may (if authorized by the copyright holders of
+that material) supplement the terms of this License with terms:
+
+    a) Disclaiming warranty or limiting liability differently from the
+    terms of sections 15 and 16 of this License; or
+
+    b) Requiring preservation of specified reasonable legal notices or
+    author attributions in that material or in the Appropriate Legal
+    Notices displayed by works containing it; or
+
+    c) Prohibiting misrepresentation of the origin of that material, or
+    requiring that modified versions of such material be marked in
+    reasonable ways as different from the original version; or
+
+    d) Limiting the use for publicity purposes of names of licensors or
+    authors of the material; or
+
+    e) Declining to grant rights under trademark law for use of some
+    trade names, trademarks, or service marks; or
+
+    f) Requiring indemnification of licensors and authors of that
+    material by anyone who conveys the material (or modified versions of
+    it) with contractual assumptions of liability to the recipient, for
+    any liability that these contractual assumptions directly impose on
+    those licensors and authors.
+
+  All other non-permissive additional terms are considered "further
+restrictions" within the meaning of section 10.  If the Program as you
+received it, or any part of it, contains a notice stating that it is
+governed by this License along with a term that is a further
+restriction, you may remove that term.  If a license document contains
+a further restriction but permits relicensing or conveying under this
+License, you may add to a covered work material governed by the terms
+of that license document, provided that the further restriction does
+not survive such relicensing or conveying.
+
+  If you add terms to a covered work in accord with this section, you
+must place, in the relevant source files, a statement of the
+additional terms that apply to those files, or a notice indicating
+where to find the applicable terms.
+
+  Additional terms, permissive or non-permissive, may be stated in the
+form of a separately written license, or stated as exceptions;
+the above requirements apply either way.
+
+  8. Termination.
+
+  You may not propagate or modify a covered work except as expressly
+provided under this License.  Any attempt otherwise to propagate or
+modify it is void, and will automatically terminate your rights under
+this License (including any patent licenses granted under the third
+paragraph of section 11).
+
+  However, if you cease all violation of this License, then your
+license from a particular copyright holder is reinstated (a)
+provisionally, unless and until the copyright holder explicitly and
+finally terminates your license, and (b) permanently, if the copyright
+holder fails to notify you of the violation by some reasonable means
+prior to 60 days after the cessation.
+
+  Moreover, your license from a particular copyright holder is
+reinstated permanently if the copyright holder notifies you of the
+violation by some reasonable means, this is the first time you have
+received notice of violation of this License (for any work) from that
+copyright holder, and you cure the violation prior to 30 days after
+your receipt of the notice.
+
+  Termination of your rights under this section does not terminate the
+licenses of parties who have received copies or rights from you under
+this License.  If your rights have been terminated and not permanently
+reinstated, you do not qualify to receive new licenses for the same
+material under section 10.
+
+  9. Acceptance Not Required for Having Copies.
+
+  You are not required to accept this License in order to receive or
+run a copy of the Program.  Ancillary propagation of a covered work
+occurring solely as a consequence of using peer-to-peer transmission
+to receive a copy likewise does not require acceptance.  However,
+nothing other than this License grants you permission to propagate or
+modify any covered work.  These actions infringe copyright if you do
+not accept this License.  Therefore, by modifying or propagating a
+covered work, you indicate your acceptance of this License to do so.
+
+  10. Automatic Licensing of Downstream Recipients.
+
+  Each time you convey a covered work, the recipient automatically
+receives a license from the original licensors, to run, modify and
+propagate that work, subject to this License.  You are not responsible
+for enforcing compliance by third parties with this License.
+
+  An "entity transaction" is a transaction transferring control of an
+organization, or substantially all assets of one, or subdividing an
+organization, or merging organizations.  If propagation of a covered
+work results from an entity transaction, each party to that
+transaction who receives a copy of the work also receives whatever
+licenses to the work the party's predecessor in interest had or could
+give under the previous paragraph, plus a right to possession of the
+Corresponding Source of the work from the predecessor in interest, if
+the predecessor has it or can get it with reasonable efforts.
+
+  You may not impose any further restrictions on the exercise of the
+rights granted or affirmed under this License.  For example, you may
+not impose a license fee, royalty, or other charge for exercise of
+rights granted under this License, and you may not initiate litigation
+(including a cross-claim or counterclaim in a lawsuit) alleging that
+any patent claim is infringed by making, using, selling, offering for
+sale, or importing the Program or any portion of it.
+
+  11. Patents.
+
+  A "contributor" is a copyright holder who authorizes use under this
+License of the Program or a work on which the Program is based.  The
+work thus licensed is called the contributor's "contributor version".
+
+  A contributor's "essential patent claims" are all patent claims
+owned or controlled by the contributor, whether already acquired or
+hereafter acquired, that would be infringed by some manner, permitted
+by this License, of making, using, or selling its contributor version,
+but do not include claims that would be infringed only as a
+consequence of further modification of the contributor version.  For
+purposes of this definition, "control" includes the right to grant
+patent sublicenses in a manner consistent with the requirements of
+this License.
+
+  Each contributor grants you a non-exclusive, worldwide, royalty-free
+patent license under the contributor's essential patent claims, to
+make, use, sell, offer for sale, import and otherwise run, modify and
+propagate the contents of its contributor version.
+
+  In the following three paragraphs, a "patent license" is any express
+agreement or commitment, however denominated, not to enforce a patent
+(such as an express permission to practice a patent or covenant not to
+sue for patent infringement).  To "grant" such a patent license to a
+party means to make such an agreement or commitment not to enforce a
+patent against the party.
+
+  If you convey a covered work, knowingly relying on a patent license,
+and the Corresponding Source of the work is not available for anyone
+to copy, free of charge and under the terms of this License, through a
+publicly available network server or other readily accessible means,
+then you must either (1) cause the Corresponding Source to be so
+available, or (2) arrange to deprive yourself of the benefit of the
+patent license for this particular work, or (3) arrange, in a manner
+consistent with the requirements of this License, to extend the patent
+license to downstream recipients.  "Knowingly relying" means you have
+actual knowledge that, but for the patent license, your conveying the
+covered work in a country, or your recipient's use of the covered work
+in a country, would infringe one or more identifiable patents in that
+country that you have reason to believe are valid.
+
+  If, pursuant to or in connection with a single transaction or
+arrangement, you convey, or propagate by procuring conveyance of, a
+covered work, and grant a patent license to some of the parties
+receiving the covered work authorizing them to use, propagate, modify
+or convey a specific copy of the covered work, then the patent license
+you grant is automatically extended to all recipients of the covered
+work and works based on it.
+
+  A patent license is "discriminatory" if it does not include within
+the scope of its coverage, prohibits the exercise of, or is
+conditioned on the non-exercise of one or more of the rights that are
+specifically granted under this License.  You may not convey a covered
+work if you are a party to an arrangement with a third party that is
+in the business of distributing software, under which you make payment
+to the third party based on the extent of your activity of conveying
+the work, and under which the third party grants, to any of the
+parties who would receive the covered work from you, a discriminatory
+patent license (a) in connection with copies of the covered work
+conveyed by you (or copies made from those copies), or (b) primarily
+for and in connection with specific products or compilations that
+contain the covered work, unless you entered into that arrangement,
+or that patent license was granted, prior to 28 March 2007.
+
+  Nothing in this License shall be construed as excluding or limiting
+any implied license or other defenses to infringement that may
+otherwise be available to you under applicable patent law.
+
+  12. No Surrender of Others' Freedom.
+
+  If conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot convey a
+covered work so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you may
+not convey it at all.  For example, if you agree to terms that obligate you
+to collect a royalty for further conveying from those to whom you convey
+the Program, the only way you could satisfy both those terms and this
+License would be to refrain entirely from conveying the Program.
+
+  13. Remote Network Interaction; Use with the GNU General Public License.
+
+  Notwithstanding any other provision of this License, if you modify the
+Program, your modified version must prominently offer all users
+interacting with it remotely through a computer network (if your version
+supports such interaction) an opportunity to receive the Corresponding
+Source of your version by providing access to the Corresponding Source
+from a network server at no charge, through some standard or customary
+means of facilitating copying of software.  This Corresponding Source
+shall include the Corresponding Source for any work covered by version 3
+of the GNU General Public License that is incorporated pursuant to the
+following paragraph.
+
+  Notwithstanding any other provision of this License, you have
+permission to link or combine any covered work with a work licensed
+under version 3 of the GNU General Public License into a single
+combined work, and to convey the resulting work.  The terms of this
+License will continue to apply to the part which is the covered work,
+but the work with which it is combined will remain governed by version
+3 of the GNU General Public License.
+
+  14. Revised Versions of this License.
+
+  The Free Software Foundation may publish revised and/or new versions of
+the GNU Affero General Public License from time to time.  Such new versions
+will be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+  Each version is given a distinguishing version number.  If the
+Program specifies that a certain numbered version of the GNU Affero General
+Public License "or any later version" applies to it, you have the
+option of following the terms and conditions either of that numbered
+version or of any later version published by the Free Software
+Foundation.  If the Program does not specify a version number of the
+GNU Affero General Public License, you may choose any version ever published
+by the Free Software Foundation.
+
+  If the Program specifies that a proxy can decide which future
+versions of the GNU Affero General Public License can be used, that proxy's
+public statement of acceptance of a version permanently authorizes you
+to choose that version for the Program.
+
+  Later license versions may give you additional or different
+permissions.  However, no additional obligations are imposed on any
+author or copyright holder as a result of your choosing to follow a
+later version.
+
+  15. Disclaimer of Warranty.
+
+  THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
+APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
+HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
+OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
+THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
+IS WITH YOU.  SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
+ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+  16. Limitation of Liability.
+
+  IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
+THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
+GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
+USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
+DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
+PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
+EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
+SUCH DAMAGES.
+
+  17. Interpretation of Sections 15 and 16.
+
+  If the disclaimer of warranty and limitation of liability provided
+above cannot be given local legal effect according to their terms,
+reviewing courts shall apply local law that most closely approximates
+an absolute waiver of all civil liability in connection with the
+Program, unless a warranty or assumption of liability accompanies a
+copy of the Program in return for a fee.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+state the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software: you can redistribute it and/or modify
+    it under the terms of the GNU Affero General Public License as published
+    by the Free Software Foundation, either version 3 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU Affero General Public License for more details.
+
+    You should have received a copy of the GNU Affero General Public License
+    along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+Also add information on how to contact you by electronic and paper mail.
+
+  If your software can interact with users remotely through a computer
+network, you should also make sure that it provides a way for users to
+get its source.  For example, if your program is a web application, its
+interface could display a "Source" link that leads users to an archive
+of the code.  There are many ways you could offer source, and different
+solutions will be better for different programs; see section 13 for the
+specific requirements.
+
+  You should also get your employer (if you work as a programmer) or school,
+if any, to sign a "copyright disclaimer" for the program, if necessary.
+For more information on this, and how to apply and follow the GNU AGPL, see
+<https://www.gnu.org/licenses/>.

README.md

@@ -1,10 +1,257 @@
----
-title: Sdui
-emoji: 🏢
-colorFrom: indigo
-colorTo: pink
-sdk: docker
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# Stable Diffusion in Docker
+
+Run the official [Stable Diffusion](https://huggingface.co/CompVis/stable-diffusion)
+releases on [Huggingface](https://huggingface.co/) in a GPU accelerated Docker
+container.
+
+```sh
+./build.sh run 'An impressionist painting of a parakeet eating spaghetti in the desert'
+```
+
+![An impressionist painting of a parakeet eating spaghetti in the desert 1](https://raw.githubusercontent.com/fboulnois/repository-assets/main/assets/stable-diffusion-docker/An_impressionist_painting_of_a_parakeet_eating_spaghetti_in_the_desert_s1.png)
+![An impressionist painting of a parakeet eating spaghetti in the desert 2](https://raw.githubusercontent.com/fboulnois/repository-assets/main/assets/stable-diffusion-docker/An_impressionist_painting_of_a_parakeet_eating_spaghetti_in_the_desert_s2.png)
+
+```sh
+./build.sh run --image parakeet_eating_spaghetti.png --strength 0.6 'Bouquet of roses'
+```
+
+![Bouquet of roses 1](https://raw.githubusercontent.com/fboulnois/repository-assets/main/assets/stable-diffusion-docker/Bouquet_of_roses_s1.png)
+![Bouquet of roses 2](https://raw.githubusercontent.com/fboulnois/repository-assets/main/assets/stable-diffusion-docker/Bouquet_of_roses_s2.png)
+
+## Before you start
+
+### Minimum requirements
+
+By default, the pipeline uses the full model and weights which requires a CUDA
+capable GPU with 8GB+ of VRAM. It should take a few seconds to create one image.
+On less powerful GPUs you may need to modify some of the options; see the
+[Examples](#examples) section for more details. If you lack a suitable GPU you
+can set the options `--device cpu` and `--onnx` instead.
+
+### Huggingface token
+
+Since it uses the official model, you will need to create a [user access token](https://huggingface.co/docs/hub/security-tokens)
+in your [Huggingface account](https://huggingface.co/settings/tokens). Save the
+user access token in a file called `token.txt` and make sure it is available
+when building the container. The token content should begin with `hf_...`
+
+## Quickstart
+
+The pipeline is managed using a single [`build.sh`](build.sh) script.
+
+Pull the latest version of `stable-diffusion-docker` using `./build.sh pull`.
+You will need to use the option `--token` to specify a valid [user access token](#huggingface-token)
+when using [`./build run`](#run).
+
+Alternately, build the image locally before running it.
+
+## Build
+
+Make sure your [user access token](#huggingface-token) is saved in a file called
+`token.txt`.
+
+To build:
+
+```sh
+./build.sh build  # or just ./build.sh
+```
+
+## Run
+
+### Text-to-Image (`txt2img`)
+
+Create an image from a text prompt.
+
+To run:
+
+```sh
+./build.sh run 'Andromeda galaxy in a bottle'
+```
+
+### Image-to-Image (`img2img`)
+
+Create an image from an existing image and a text prompt.
+
+First, copy an image to the `input` folder. Next, to run:
+
+```sh
+./build.sh run --image image.png 'Andromeda galaxy in a bottle'
+```
+
+### Depth-Guided Diffusion (`depth2img`)
+
+Modify an existing image with its depth map and a text prompt.
+
+First, copy an image to the `input` folder. Next, to run:
+
+```sh
+./build.sh run --model 'stabilityai/stable-diffusion-2-depth' \
+  --image image.png 'A detailed description of the objects to change'
+```
+
+### Instruct Pix2Pix (`pix2pix`)
+
+Modify an existing image with a text prompt.
+
+First, copy an image to the `input` folder. Next, to run:
+
+```sh
+./build.sh run --model 'timbrooks/instruct-pix2pix' \
+  --image image.png 'A detailed description of the objects to change'
+```
+
+### Image Upscaling (`upscale4x`)
+
+Create a high resolution image from an existing image with a text prompt.
+
+First, copy an image to the `input` folder. Next, to run:
+
+```sh
+./build.sh run --model 'stabilityai/stable-diffusion-x4-upscaler' \
+  --image image.png 'Andromeda galaxy in a bottle'
+```
+
+### Diffusion Inpainting (`inpaint`)
+
+Modify specific areas of an existing image with an image mask and a text prompt.
+
+First, copy an image and an image mask to the `input` folder. White areas of the
+mask will be diffused and black areas will be kept untouched. Next, to run:
+
+```sh
+./build.sh run --model 'runwayml/stable-diffusion-inpainting' \
+  --image image.png --mask mask.png 'Andromeda galaxy in a bottle'
+```
+
+## Options
+
+The following are the most common options:
+
+* `--prompt [PROMPT]`: the prompt to render into an image
+* `--model [MODEL]`: the model used to render images (default is
+`CompVis/stable-diffusion-v1-4`)
+* `--height [HEIGHT]`: image height in pixels (default 512, must be divisible by 64)
+* `--width [WIDTH]`: image width in pixels (default 512, must be divisible by 64)
+* `--iters [ITERS]`: number of times to run pipeline (default 1)
+* `--samples [SAMPLES]`: number of images to create per run (default 1)
+* `--scale [SCALE]`: how closely the image should follow the prompt (default 7.5)
+* `--scheduler [SCHEDULER]`: override the scheduler used to denoise the image
+(default `None`)
+* `--seed [SEED]`: RNG seed for repeatability (default is a random seed)
+* `--steps [STEPS]`: number of sampling steps (default 50)
+
+Other options:
+
+* `--attention-slicing`: use less memory but decrease inference speed (default
+is no attention slicing)
+* `--device [DEVICE]`: the cpu or cuda device to use to render images (default
+`cuda`)
+* `--half`: use float16 tensors instead of float32 (default `float32`)
+* `--image [IMAGE]`: the input image to use for image-to-image diffusion
+(default `None`)
+* `--image-scale [IMAGE_SCALE]`: how closely the image should follow the
+original image (default `None`)
+* `--mask [MASK]`: the input mask to use for diffusion inpainting (default
+`None`)
+* `--negative-prompt [NEGATIVE_PROMPT]`: the prompt to not render into an image
+(default `None`)
+* `--onnx`: use the onnx runtime for inference (default is off)
+* `--skip`: skip safety checker (default is the safety checker is on)
+* `--strength [STRENGTH]`: diffusion strength to apply to the input image
+(default 0.75)
+* `--token [TOKEN]`: specify a Huggingface user access token at the command line
+instead of reading it from a file (default is a file)
+* `--vae-tiling`: use less memory when generating ultra-high resolution images
+but massively decrease inference speed (default is no tiling)
+* `--xformers-memory-efficient-attention`: use less memory but require the
+xformers library (default is that xformers is not required)
+
+Some of the original `txt2img.py` options [have been renamed](https://github.com/fboulnois/stable-diffusion-docker/issues/49)
+for easy-of-use and compatibility with other pipelines:
+
+| txt2img | stable-diffusion-docker |
+|---------|-------------------------|
+| `--H` | `--height` |
+| `--W` | `--width` |
+| `--n_iter` | `--iters` |
+| `--n_samples` | `--samples` |
+| `--ddim_steps` | `--steps` |
+
+## Examples
+
+These commands are both identical:
+
+```sh
+./build.sh run 'abstract art'
+./build.sh run --prompt 'abstract art'
+```
+
+Set the seed to 42:
+
+```sh
+./build.sh run --seed 42 'abstract art'
+```
+
+Options can be combined:
+
+```sh
+./build.sh run --scale 7.0 --seed 42 'abstract art'
+```
+
+Many popular models are supported out-of-the-box:
+
+| Model Name | Option using `--model` |
+|------------|------------------------|
+| [Stable Diffusion 1.4](https://huggingface.co/CompVis/stable-diffusion-v1-4) | `'CompVis/stable-diffusion-v1-4'` |
+| [Stable Diffusion 1.5](https://huggingface.co/runwayml/stable-diffusion-v1-5) | `'runwayml/stable-diffusion-v1-5'` |
+| [Stable Diffusion 2.0](https://huggingface.co/stabilityai/stable-diffusion-2) | `'stabilityai/stable-diffusion-2'` |
+| [Stable Diffusion 2.1](https://huggingface.co/stabilityai/stable-diffusion-2-1) | `'stabilityai/stable-diffusion-2-1'` |
+| [OpenJourney 1.0](https://huggingface.co/prompthero/openjourney) | `'prompthero/openjourney'` |
+| [Dreamlike Diffusion 1.0](https://huggingface.co/dreamlike-art/dreamlike-diffusion-1.0) | `'dreamlike-art/dreamlike-diffusion-1.0'` |
+| [and more!](https://huggingface.co/models?other=stable-diffusion&sort=likes) | ... |
+
+```sh
+./build.sh run --model 'prompthero/openjourney' --prompt 'abstract art'
+```
+
+On systems without enough GPU VRAM, you can try mixing and matching options:
+
+* Give Docker Desktop more resources by increasing the CPU, memory, and swap in
+the Settings -> Resources section if the container is terminated
+* Make images smaller than 512x512 using `--height` and `--width` to decrease
+memory use and increase image creation speed
+* Use `--half` to decrease memory use but slightly decrease image quality
+* Use `--attention-slicing` to decrease memory use but also decrease image
+creation speed
+* Use `--xformers-memory-efficient-attention` to decrease memory use if the
+pipeline and the hardware supports the option
+* Decrease the number of samples and increase the number of iterations with
+`--samples` and `--iters` to decrease overall memory use
+* Skip the safety checker with `--skip` to run less code
+
+```sh
+./build.sh run --height 256 --width 256 --half \
+  --attention-slicing --xformers-memory-efficient-attention \
+  --samples 1 --iters 1 --skip --prompt 'abstract art'
+```
+
+On Windows, if you aren't using WSL2 and instead use MSYS, MinGW, or Git Bash,
+prefix your commands with `MSYS_NO_PATHCONV=1` (or export it beforehand):
+
+```sh
+MSYS_NO_PATHCONV=1 ./build.sh run --half --prompt 'abstract art'
+```
+
+## Outputs
+
+### Model
+
+The model and other files are cached in a volume called `huggingface`. The
+models are stored in `<volume>/diffusers/<model>/snapshots/<githash>/unet/<weights>`.
+Checkpoint files (`ckpt`s) are unofficial versions of the official models, and
+so these are not part of the official release.
+
+### Images
+
+The images are saved as PNGs in the `output` folder using the prompt text. The
+`build.sh` script creates and mounts this folder as a volume in the container.

build.sh

@@ -0,0 +1,95 @@
+#!/bin/sh
+
+set -eu
+
+CWD=$(basename "$PWD")
+
+set_gpu_arg() {
+    while [ "$#" -gt 0 ]; do
+        if [ "$1" = "--device" ] && [ "$2" = "cpu" ]; then
+            GPU_ARG=""
+            return
+        fi
+        shift
+    done
+    GPU_ARG="--gpus=all"
+}
+
+build() {
+    docker build . --tag "$CWD"
+}
+
+clean() {
+    docker system prune -f
+}
+
+dev() {
+    docker run --rm --gpus=all --entrypoint=sh \
+        -v huggingface:/home/huggingface/.cache/huggingface \
+        -v "$PWD"/input:/home/huggingface/input \
+        -v "$PWD"/output:/home/huggingface/output \
+        -it "$CWD"
+}
+
+pull() {
+    GHCR="ghcr.io/fboulnois/stable-diffusion-docker"
+    docker pull "$GHCR"
+    docker tag "$GHCR" "$CWD"
+}
+
+run() {
+    set_gpu_arg "$@"
+    docker run --rm ${GPU_ARG} \
+        -v huggingface:/home/huggingface/.cache/huggingface \
+        -v "$PWD"/input:/home/huggingface/input \
+        -v "$PWD"/output:/home/huggingface/output \
+        "$CWD" "$@"
+}
+
+tests() {
+    BASE_URL="https://raw.githubusercontent.com/fboulnois/repository-assets/main/assets/stable-diffusion-docker"
+    TEST_IMAGE="An_impressionist_painting_of_a_parakeet_eating_spaghetti_in_the_desert_full.png"
+    curl -sL "${BASE_URL}/${TEST_IMAGE}" > "$PWD/input/${TEST_IMAGE}"
+    run --skip --height 512 --width 640 "abstract art"
+    run --device cpu --onnx --image "${TEST_IMAGE}" --strength 0.6 "abstract art"
+    run --model "stabilityai/stable-diffusion-2" \
+        --skip --height 768 --width 768 "abstract art"
+    run --model "stabilityai/stable-diffusion-2-1" \
+        --skip --height 768 --width 768 "abstract art"
+    run --model "stabilityai/stable-diffusion-x4-upscaler" \
+        --image "${TEST_IMAGE}" --half --attention-slicing \
+        --xformers-memory-efficient-attention \
+        --prompt "An impressionist painting of a parakeet eating spaghetti in the desert"
+    run --model "stabilityai/stable-diffusion-2-depth" \
+        --height 768 --width 768 \
+        --image "${TEST_IMAGE}" --attention-slicing \
+        --xformers-memory-efficient-attention \
+        --negative-prompt "bad, ugly, deformed, malformed, mutated, bad anatomy" \
+        --prompt "a toucan"
+    run --model "timbrooks/instruct-pix2pix" \
+        --scale 7.0 --image-scale 2.0 \
+        --image "${TEST_IMAGE}" --attention-slicing \
+        --xformers-memory-efficient-attention \
+        --negative-prompt "bad, ugly, deformed, malformed, mutated, bad anatomy" \
+        --prompt "replace the sky with bricks"
+    run --model "dreamlike-art/dreamlike-diffusion-1.0" \
+        --skip --vae-tiling --xformers-memory-efficient-attention \
+        --height 1024 --width 1024 "abstract art"
+    run --model "runwayml/stable-diffusion-v1-5" \
+        --samples 2 --iters 2 --seed 42 \
+        --scheduler HeunDiscreteScheduler \
+        --scale 7.5 --steps 80 --attention-slicing \
+        --half --skip --negative-prompt "red roses" \
+        --prompt "bouquet of roses"
+}
+
+mkdir -p input output
+case ${1:-build} in
+    build) build ;;
+    clean) clean ;;
+    dev) dev "$@" ;;
+    pull) pull ;;
+    run) shift; run "$@" ;;
+    test) tests ;;
+    *) echo "$0: No command named '$1'" ;;
+esac

docker-entrypoint.py

@@ -0,0 +1,286 @@
+#!/usr/bin/env python
+import argparse, datetime, inspect, os, warnings
+import numpy as np
+import torch
+from PIL import Image
+from diffusers import (
+    OnnxStableDiffusionPipeline,
+    OnnxStableDiffusionInpaintPipeline,
+    OnnxStableDiffusionImg2ImgPipeline,
+    StableDiffusionDepth2ImgPipeline,
+    StableDiffusionPipeline,
+    StableDiffusionImg2ImgPipeline,
+    StableDiffusionInpaintPipeline,
+    StableDiffusionInstructPix2PixPipeline,
+    StableDiffusionUpscalePipeline,
+    schedulers,
+)
+
+
+def iso_date_time():
+    return datetime.datetime.now().isoformat()
+
+
+def load_image(path):
+    image = Image.open(os.path.join("input", path)).convert("RGB")
+    print(f"loaded image from {path}:", iso_date_time(), flush=True)
+    return image
+
+
+def remove_unused_args(p):
+    params = inspect.signature(p.pipeline).parameters.keys()
+    args = {
+        "prompt": p.prompt,
+        "negative_prompt": p.negative_prompt,
+        "image": p.image,
+        "mask_image": p.mask,
+        "height": p.height,
+        "width": p.width,
+        "num_images_per_prompt": p.samples,
+        "num_inference_steps": p.steps,
+        "guidance_scale": p.scale,
+        "image_guidance_scale": p.image_scale,
+        "strength": p.strength,
+        "generator": p.generator,
+    }
+    return {p: args[p] for p in params if p in args}
+
+
+def stable_diffusion_pipeline(p):
+    p.dtype = torch.float16 if p.half else torch.float32
+
+    if p.onnx:
+        p.diffuser = OnnxStableDiffusionPipeline
+        p.revision = "onnx"
+    else:
+        p.diffuser = StableDiffusionPipeline
+        p.revision = "fp16" if p.half else "main"
+
+    models = argparse.Namespace(
+        **{
+            "depth2img": ["stabilityai/stable-diffusion-2-depth"],
+            "pix2pix": ["timbrooks/instruct-pix2pix"],
+            "upscalers": ["stabilityai/stable-diffusion-x4-upscaler"],
+        }
+    )
+
+    if p.image is not None:
+        if p.revision == "onnx":
+            p.diffuser = OnnxStableDiffusionImg2ImgPipeline
+        elif p.model in models.depth2img:
+            p.diffuser = StableDiffusionDepth2ImgPipeline
+        elif p.model in models.pix2pix:
+            p.diffuser = StableDiffusionInstructPix2PixPipeline
+        elif p.model in models.upscalers:
+            p.diffuser = StableDiffusionUpscalePipeline
+        else:
+            p.diffuser = StableDiffusionImg2ImgPipeline
+        p.image = load_image(p.image)
+
+    if p.mask is not None:
+        if p.revision == "onnx":
+            p.diffuser = OnnxStableDiffusionInpaintPipeline
+        else:
+            p.diffuser = StableDiffusionInpaintPipeline
+        p.mask = load_image(p.mask)
+
+    if p.token is None:
+        with open("token.txt") as f:
+            p.token = f.read().replace("\n", "")
+
+    if p.seed == 0:
+        p.seed = torch.random.seed()
+
+    if p.revision == "onnx":
+        p.seed = p.seed >> 32 if p.seed > 2**32 - 1 else p.seed
+        p.generator = np.random.RandomState(p.seed)
+    else:
+        p.generator = torch.Generator(device=p.device).manual_seed(p.seed)
+
+    print("load pipeline start:", iso_date_time(), flush=True)
+
+    with warnings.catch_warnings():
+        warnings.filterwarnings("ignore", category=FutureWarning)
+        pipeline = p.diffuser.from_pretrained(
+            p.model,
+            torch_dtype=p.dtype,
+            revision=p.revision,
+            use_auth_token=p.token,
+        ).to(p.device)
+
+    if p.scheduler is not None:
+        scheduler = getattr(schedulers, p.scheduler)
+        pipeline.scheduler = scheduler.from_config(pipeline.scheduler.config)
+
+    if p.skip:
+        pipeline.safety_checker = None
+
+    if p.attention_slicing:
+        pipeline.enable_attention_slicing()
+
+    if p.xformers_memory_efficient_attention:
+        pipeline.enable_xformers_memory_efficient_attention()
+
+    if p.vae_tiling:
+        pipeline.vae.enable_tiling()
+
+    p.pipeline = pipeline
+
+    print("loaded models after:", iso_date_time(), flush=True)
+
+    return p
+
+
+def stable_diffusion_inference(p):
+    prefix = p.prompt.replace(" ", "_")[:170]
+    for j in range(p.iters):
+        result = p.pipeline(**remove_unused_args(p))
+
+        for i, img in enumerate(result.images):
+            idx = j * p.samples + i + 1
+            out = f"{prefix}__steps_{p.steps}__scale_{p.scale:.2f}__seed_{p.seed}__n_{idx}.png"
+            img.save(os.path.join("output", out))
+
+    print("completed pipeline:", iso_date_time(), flush=True)
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Create images from a text prompt.")
+    parser.add_argument(
+        "--attention-slicing",
+        action="store_true",
+        help="Use less memory at the expense of inference speed",
+    )
+    parser.add_argument(
+        "--device",
+        type=str,
+        nargs="?",
+        default="cuda",
+        help="The cpu or cuda device to use to render images",
+    )
+    parser.add_argument(
+        "--half",
+        action="store_true",
+        help="Use float16 (half-sized) tensors instead of float32",
+    )
+    parser.add_argument(
+        "--height", type=int, nargs="?", default=512, help="Image height in pixels"
+    )
+    parser.add_argument(
+        "--image",
+        type=str,
+        nargs="?",
+        help="The input image to use for image-to-image diffusion",
+    )
+    parser.add_argument(
+        "--image-scale",
+        type=float,
+        nargs="?",
+        help="How closely the image should follow the original image",
+    )
+    parser.add_argument(
+        "--iters",
+        type=int,
+        nargs="?",
+        default=1,
+        help="Number of times to run pipeline",
+    )
+    parser.add_argument(
+        "--mask",
+        type=str,
+        nargs="?",
+        help="The input mask to use for diffusion inpainting",
+    )
+    parser.add_argument(
+        "--model",
+        type=str,
+        nargs="?",
+        default="CompVis/stable-diffusion-v1-4",
+        help="The model used to render images",
+    )
+    parser.add_argument(
+        "--negative-prompt",
+        type=str,
+        nargs="?",
+        help="The prompt to not render into an image",
+    )
+    parser.add_argument(
+        "--onnx",
+        action="store_true",
+        help="Use the onnx runtime for inference",
+    )
+    parser.add_argument(
+        "--prompt", type=str, nargs="?", help="The prompt to render into an image"
+    )
+    parser.add_argument(
+        "--samples",
+        type=int,
+        nargs="?",
+        default=1,
+        help="Number of images to create per run",
+    )
+    parser.add_argument(
+        "--scale",
+        type=float,
+        nargs="?",
+        default=7.5,
+        help="How closely the image should follow the prompt",
+    )
+    parser.add_argument(
+        "--scheduler",
+        type=str,
+        nargs="?",
+        help="Override the scheduler used to denoise the image",
+    )
+    parser.add_argument(
+        "--seed", type=int, nargs="?", default=0, help="RNG seed for repeatability"
+    )
+    parser.add_argument(
+        "--skip",
+        action="store_true",
+        help="Skip the safety checker",
+    )
+    parser.add_argument(
+        "--steps", type=int, nargs="?", default=50, help="Number of sampling steps"
+    )
+    parser.add_argument(
+        "--strength",
+        type=float,
+        default=0.75,
+        help="Diffusion strength to apply to the input image",
+    )
+    parser.add_argument(
+        "--token", type=str, nargs="?", help="Huggingface user access token"
+    )
+    parser.add_argument(
+        "--vae-tiling",
+        action="store_true",
+        help="Use less memory when generating ultra-high resolution images",
+    )
+    parser.add_argument(
+        "--width", type=int, nargs="?", default=512, help="Image width in pixels"
+    )
+    parser.add_argument(
+        "--xformers-memory-efficient-attention",
+        action="store_true",
+        help="Use less memory but require the xformers library",
+    )
+    parser.add_argument(
+        "prompt0",
+        metavar="PROMPT",
+        type=str,
+        nargs="?",
+        help="The prompt to render into an image",
+    )
+
+    args = parser.parse_args()
+
+    if args.prompt0 is not None:
+        args.prompt = args.prompt0
+
+    pipeline = stable_diffusion_pipeline(args)
+    stable_diffusion_inference(pipeline)
+
+
+if __name__ == "__main__":
+    main()

requirements.txt

@@ -0,0 +1,6 @@
+diffusers[torch]==0.14.0
+onnxruntime==1.14.1
+safetensors==0.3.0
+torch==1.13.1+cu117
+transformers==4.26.1
+xformers==0.0.16