Upload folder using huggingface_hub

LICENSE

@@ -0,0 +1,674 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 3, 29 June 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 General Public License is a free, copyleft license for
+software and other kinds of works.
+
+  The licenses for most software and other practical works are designed
+to take away your freedom to share and change the works.  By contrast,
+the GNU General Public License is 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.  We, the Free Software Foundation, use the
+GNU General Public License for most of our software; it applies also to
+any other work released this way by its authors.  You can apply it to
+your programs, too.
+
+  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.
+
+  To protect your rights, we need to prevent others from denying you
+these rights or asking you to surrender the rights.  Therefore, you have
+certain responsibilities if you distribute copies of the software, or if
+you modify it: responsibilities to respect the freedom of others.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must pass on to the recipients the same
+freedoms that you received.  You must make sure that they, too, receive
+or can get the source code.  And you must show them these terms so they
+know their rights.
+
+  Developers that use the GNU GPL protect your rights with two steps:
+(1) assert copyright on the software, and (2) offer you this License
+giving you legal permission to copy, distribute and/or modify it.
+
+  For the developers' and authors' protection, the GPL clearly explains
+that there is no warranty for this free software.  For both users' and
+authors' sake, the GPL requires that modified versions be marked as
+changed, so that their problems will not be attributed erroneously to
+authors of previous versions.
+
+  Some devices are designed to deny users access to install or run
+modified versions of the software inside them, although the manufacturer
+can do so.  This is fundamentally incompatible with the aim of
+protecting users' freedom to change the software.  The systematic
+pattern of such abuse occurs in the area of products for individuals to
+use, which is precisely where it is most unacceptable.  Therefore, we
+have designed this version of the GPL to prohibit the practice for those
+products.  If such problems arise substantially in other domains, we
+stand ready to extend this provision to those domains in future versions
+of the GPL, as needed to protect the freedom of users.
+
+  Finally, every program is threatened constantly by software patents.
+States should not allow patents to restrict development and use of
+software on general-purpose computers, but in those that do, we wish to
+avoid the special danger that patents applied to a free program could
+make it effectively proprietary.  To prevent this, the GPL assures that
+patents cannot be used to render the program non-free.
+
+  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 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. Use with the GNU Affero General Public License.
+
+  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 Affero 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 special requirements of the GNU Affero General Public License,
+section 13, concerning interaction through a network will apply to the
+combination as such.
+
+  14. Revised Versions of this License.
+
+  The Free Software Foundation may publish revised and/or new versions of
+the GNU 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 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 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 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 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 General Public License for more details.
+
+    You should have received a copy of the GNU 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 the program does terminal interaction, make it output a short
+notice like this when it starts in an interactive mode:
+
+    <program>  Copyright (C) <year>  <name of author>
+    This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, your program's commands
+might be different; for a GUI interface, you would use an "about box".
+
+  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 GPL, see
+<https://www.gnu.org/licenses/>.
+
+  The GNU General Public License does not permit incorporating your program
+into proprietary programs.  If your program is a subroutine library, you
+may consider it more useful to permit linking proprietary applications with
+the library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.  But first, please read
+<https://www.gnu.org/licenses/why-not-lgpl.html>.

README.md

@@ -1,5 +1,5 @@
 ---
-title: ImageGen - Z-Image
+title: ImageGen3
 emoji: 🖼
 colorFrom: purple
 colorTo: red

app.py

@@ -1,7 +1,6 @@
 import spaces
 import os
 import sys
-import requests
 import site
 
 APP_DIR = os.path.dirname(os.path.abspath(__file__))
@@ -45,11 +44,11 @@ def dummy_gpu_for_startup():
     print("--- [GPU Startup] Startup check passed. ---")
     return "Startup check passed."
 
+
 def main():
-    from utils.app_utils import print_welcome_message
     from scripts import build_sage_attention
-
-    print_welcome_message()
+    from comfy_integration import setup as setup_comfyui
+    from utils.app_utils import load_ipadapter_presets
     
     print("--- [Setup] Attempting to build and install SageAttention... ---")
     try:
@@ -58,7 +57,9 @@ def main():
     except Exception as e:
         print(f"--- [Setup] ❌ SageAttention installation failed: {e}. Continuing with default attention. ---")
 
-
+    print("--- [Setup] Starting ComfyUI initialization ---")
+    setup_comfyui.initialize_comfyui()
+    
     print("--- [Setup] Reloading site-packages to detect newly installed packages... ---")
     try:
         site.main()
@@ -66,52 +67,18 @@ def main():
     except Exception as e:
         print(f"--- [Setup] ⚠️  Warning: Could not fully reload site-packages: {e} ---")
 
-    from comfy_integration import setup as setup_comfyui
-    from utils.app_utils import (
-        build_preprocessor_model_map,
-        build_preprocessor_parameter_map
-    )
-    from core import shared_state
-    from core.settings import ALL_MODEL_MAP, ALL_FILE_DOWNLOAD_MAP
-
-    def check_all_model_urls_on_startup():
-        print("--- [Setup] Checking all model URL validity (one-time check) ---")
-        for display_name, model_info in ALL_MODEL_MAP.items():
-            _, components, _, _ = model_info
-            if not components: continue
-            
-            for filename in components.values():
-                download_info = ALL_FILE_DOWNLOAD_MAP.get(filename, {})
-                repo_id = download_info.get('repo_id')
-                if not repo_id: continue
-                
-                repo_file_path = download_info.get('repository_file_path', filename)
-                url = f"https://huggingface.co/{repo_id}/resolve/main/{repo_file_path}"
-                
-                try:
-                    response = requests.head(url, timeout=5, allow_redirects=True)
-                    if response.status_code >= 400:
-                        print(f"❌ Invalid URL for '{display_name}' component '{filename}': {url} (Status: {response.status_code})")
-                        shared_state.INVALID_MODEL_URLS[display_name] = True
-                        break 
-                except requests.RequestException as e:
-                    print(f"❌ URL check failed for '{display_name}' component '{filename}': {e}")
-                    shared_state.INVALID_MODEL_URLS[display_name] = True
-                    break
-        print("--- [Setup] ✅ Finished checking model URLs. ---")
+    print("--- Initiating GPU Startup Check & SageAttention Patch ---")
+    try:
+        dummy_gpu_for_startup()
+    except Exception as e:
+        print(f"--- [GPU Startup] ⚠️ Warning: Startup check failed: {e} ---")
 
     print("--- Starting Application Setup ---")
     
-    setup_comfyui.initialize_comfyui()
+    print("--- Loading IPAdapter presets ---")
+    load_ipadapter_presets()
+    print("--- ✅ IPAdapter setup complete. ---")
 
-    check_all_model_urls_on_startup()
-    
-    print("--- Building ControlNet preprocessor maps ---")
-    from core.generation_logic import build_reverse_map
-    build_reverse_map()
-    build_preprocessor_model_map()
-    build_preprocessor_parameter_map()
-    print("--- ✅ ControlNet preprocessor setup complete. ---")
 
     print("--- Environment configured. Proceeding with module imports. ---")
     from ui.layout import build_ui

chain_injectors/anima_controlnet_lllite_injector.py

@@ -0,0 +1,48 @@
+def inject(assembler, chain_definition, chain_items):
+    if not chain_items:
+        return
+
+    ksampler_name = chain_definition.get('ksampler_node', 'ksampler')
+    if ksampler_name not in assembler.node_map:
+        print(f"Warning: KSampler node '{ksampler_name}' not found for Anima LLLite chain. Skipping.")
+        return
+        
+    ksampler_id = assembler.node_map[ksampler_name]
+
+    if 'model' not in assembler.workflow[ksampler_id]['inputs']:
+        print(f"Warning: KSampler node '{ksampler_name}' is missing 'model' input. Skipping.")
+        return
+        
+    current_model_connection = assembler.workflow[ksampler_id]['inputs']['model']
+    
+    for item_data in chain_items:
+        image_loader_id = assembler._get_unique_id()
+        image_loader_node = assembler._get_node_template("LoadImage")
+        image_loader_node['inputs']['image'] = item_data['image']
+        assembler.workflow[image_loader_id] = image_loader_node
+
+        image_scaler_id = assembler._get_unique_id()
+        image_scaler_node = assembler._get_node_template("ImageScaleToTotalPixels")
+        image_scaler_node['inputs']['image'] = [image_loader_id, 0]
+        image_scaler_node['inputs']['upscale_method'] = 'nearest-exact'
+        image_scaler_node['inputs']['megapixels'] = 1.0
+        assembler.workflow[image_scaler_id] = image_scaler_node
+
+        apply_cn_id = assembler._get_unique_id()
+        apply_cn_node = assembler._get_node_template("AnimaLLLiteApply")
+        
+        apply_cn_node['inputs']['lllite_name'] = item_data['control_net_name']
+        apply_cn_node['inputs']['strength'] = item_data['strength']
+        apply_cn_node['inputs']['start_percent'] = item_data.get('start_percent', 0.0)
+        apply_cn_node['inputs']['end_percent'] = item_data.get('end_percent', 1.0)
+
+        apply_cn_node['inputs']['model'] = current_model_connection
+        apply_cn_node['inputs']['image'] = [image_scaler_id, 0]
+        
+        assembler.workflow[apply_cn_id] = apply_cn_node
+
+        current_model_connection = [apply_cn_id, 0]
+
+    assembler.workflow[ksampler_id]['inputs']['model'] = current_model_connection
+    
+    print(f"Anima LLLite injector applied. KSampler model input re-routed through {len(chain_items)} LLLite(s).")

chain_injectors/controlnet_injector.py

@@ -13,6 +13,16 @@ def inject(assembler, chain_definition, chain_items):
        'negative' not in assembler.workflow[ksampler_id]['inputs']:
         print(f"Warning: KSampler node '{ksampler_name}' is missing 'positive' or 'negative' inputs. Skipping ControlNet chain.")
         return
+
+    vae_source_str = chain_definition.get('vae_source')
+    if not vae_source_str:
+        print("Warning: 'vae_source' definition missing in the recipe for the ControlNet chain. Skipping.")
+        return
+    vae_node_name, vae_idx_str = vae_source_str.split(':')
+    if vae_node_name not in assembler.node_map:
+        print(f"Warning: VAE source node '{vae_node_name}' for ControlNet chain not found. Skipping.")
+        return
+    vae_connection = [assembler.node_map[vae_node_name], int(vae_idx_str)]
         
     current_positive_connection = assembler.workflow[ksampler_id]['inputs']['positive']
     current_negative_connection = assembler.workflow[ksampler_id]['inputs']['negative']
@@ -37,6 +47,7 @@ def inject(assembler, chain_definition, chain_items):
         apply_cn_node['inputs']['negative'] = current_negative_connection
         apply_cn_node['inputs']['control_net'] = [cn_loader_id, 0]
         apply_cn_node['inputs']['image'] = [image_loader_id, 0]
+        apply_cn_node['inputs']['vae'] = vae_connection
         
         assembler.workflow[apply_cn_id] = apply_cn_node
 

chain_injectors/flux1_ipadapter_injector.py

@@ -0,0 +1,46 @@
+def inject(assembler, chain_definition, chain_items):
+    if not chain_items:
+        return
+
+    ksampler_name = chain_definition.get('ksampler_node', 'ksampler')
+    if ksampler_name not in assembler.node_map:
+        print(f"Warning: KSampler node '{ksampler_name}' not found for Flux1 IPAdapter chain. Skipping.")
+        return
+        
+    ksampler_id = assembler.node_map[ksampler_name]
+
+    if 'model' not in assembler.workflow[ksampler_id]['inputs']:
+        print(f"Warning: KSampler node '{ksampler_name}' is missing 'model' input. Skipping Flux1 IPAdapter chain.")
+        return
+        
+    current_model_connection = assembler.workflow[ksampler_id]['inputs']['model']
+    
+    for item_data in chain_items:
+        image_loader_id = assembler._get_unique_id()
+        image_loader_node = assembler._get_node_template("LoadImage")
+        image_loader_node['inputs']['image'] = item_data['image']
+        assembler.workflow[image_loader_id] = image_loader_node
+
+        ipadapter_loader_id = assembler._get_unique_id()
+        ipadapter_loader_node = assembler._get_node_template("IPAdapterFluxLoader")
+        ipadapter_loader_node['inputs']['ipadapter'] = "ip-adapter.bin"
+        ipadapter_loader_node['inputs']['clip_vision'] = "google/siglip-so400m-patch14-384"
+        ipadapter_loader_node['inputs']['provider'] = "cuda"
+        assembler.workflow[ipadapter_loader_id] = ipadapter_loader_node
+        
+        apply_ipa_id = assembler._get_unique_id()
+        apply_ipa_node = assembler._get_node_template("ApplyIPAdapterFlux")
+        
+        apply_ipa_node['inputs']['weight'] = item_data['weight']
+        apply_ipa_node['inputs']['start_percent'] = item_data.get('start_percent', 0.0)
+        apply_ipa_node['inputs']['end_percent'] = item_data.get('end_percent', 0.6)
+        
+        apply_ipa_node['inputs']['model'] = current_model_connection
+        apply_ipa_node['inputs']['ipadapter_flux'] = [ipadapter_loader_id, 0]
+        apply_ipa_node['inputs']['image'] = [image_loader_id, 0]
+        
+        assembler.workflow[apply_ipa_id] = apply_ipa_node
+        current_model_connection = [apply_ipa_id, 0]
+
+    assembler.workflow[ksampler_id]['inputs']['model'] = current_model_connection
+    print(f"Flux1 IPAdapter injector applied. KSampler model input re-routed through {len(chain_items)} IPAdapter(s).")

chain_injectors/ipadapter_injector.py

@@ -0,0 +1,106 @@
+def inject(assembler, chain_definition, chain_items):
+    if not chain_items:
+        return
+
+    final_settings = {}
+    if chain_items and isinstance(chain_items[-1], dict) and chain_items[-1].get('is_final_settings'):
+        final_settings = chain_items.pop()
+
+    if not chain_items:
+        return
+
+    end_node_name = chain_definition.get('end')
+    if not end_node_name or end_node_name not in assembler.node_map:
+        print(f"Warning: Target node '{end_node_name}' for IPAdapter chain not found. Skipping chain injection.")
+        return
+        
+    end_node_id = assembler.node_map[end_node_name]
+    
+    if 'model' not in assembler.workflow[end_node_id]['inputs']:
+        print(f"Warning: Target node '{end_node_name}' is missing 'model' input. Skipping IPAdapter chain.")
+        return
+    
+    current_model_connection = assembler.workflow[end_node_id]['inputs']['model']
+    
+    model_type = final_settings.get('model_type', 'sdxl')
+    megapixels = 1.05 if model_type == 'sdxl' else 0.39
+
+    pos_embed_outputs = []
+    neg_embed_outputs = []
+
+    for i, item_data in enumerate(chain_items):
+        loader_type = 'FaceID' if 'FACEID' in item_data.get('preset', '') else 'Unified'
+        
+        loader_template_name = "IPAdapterUnifiedLoader"
+        if loader_type == 'FaceID':
+            loader_template_name = "IPAdapterUnifiedLoaderFaceID"
+
+        image_loader_id = assembler._get_unique_id()
+        image_loader_node = assembler._get_node_template("LoadImage")
+        image_loader_node['inputs']['image'] = item_data['image']
+        assembler.workflow[image_loader_id] = image_loader_node
+        
+        image_scaler_id = assembler._get_unique_id()
+        image_scaler_node = assembler._get_node_template("ImageScaleToTotalPixels")
+        image_scaler_node['inputs']['image'] = [image_loader_id, 0]
+        image_scaler_node['inputs']['megapixels'] = megapixels
+        image_scaler_node['inputs']['upscale_method'] = "lanczos"
+        assembler.workflow[image_scaler_id] = image_scaler_node
+
+        ipadapter_loader_id = assembler._get_unique_id()
+        ipadapter_loader_node = assembler._get_node_template(loader_template_name)
+        ipadapter_loader_node['inputs']['model'] = current_model_connection
+        ipadapter_loader_node['inputs']['preset'] = item_data['preset']
+        if loader_type == 'FaceID':
+             ipadapter_loader_node['inputs']['lora_strength'] = item_data.get('lora_strength', 0.6)
+        assembler.workflow[ipadapter_loader_id] = ipadapter_loader_node
+        
+        encoder_id = assembler._get_unique_id()
+        encoder_node = assembler._get_node_template("IPAdapterEncoder")
+        encoder_node['inputs']['weight'] = item_data['weight']
+        encoder_node['inputs']['ipadapter'] = [ipadapter_loader_id, 1]
+        encoder_node['inputs']['image'] = [image_scaler_id, 0]
+        assembler.workflow[encoder_id] = encoder_node
+        
+        pos_embed_outputs.append([encoder_id, 0])
+        neg_embed_outputs.append([encoder_id, 1])
+
+    pos_combiner_id = assembler._get_unique_id()
+    pos_combiner_node = assembler._get_node_template("IPAdapterCombineEmbeds")
+    pos_combiner_node['inputs']['method'] = final_settings.get('final_combine_method', 'concat')
+    for i, conn in enumerate(pos_embed_outputs):
+        pos_combiner_node['inputs'][f'embed{i+1}'] = conn
+    assembler.workflow[pos_combiner_id] = pos_combiner_node
+
+    neg_combiner_id = assembler._get_unique_id()
+    neg_combiner_node = assembler._get_node_template("IPAdapterCombineEmbeds")
+    neg_combiner_node['inputs']['method'] = final_settings.get('final_combine_method', 'concat')
+    for i, conn in enumerate(neg_embed_outputs):
+        neg_combiner_node['inputs'][f'embed{i+1}'] = conn
+    assembler.workflow[neg_combiner_id] = neg_combiner_node
+    
+    final_loader_type = 'FaceID' if 'FACEID' in final_settings.get('final_preset', '') else 'Unified'
+    final_loader_template_name = "IPAdapterUnifiedLoader"
+    if final_loader_type == 'FaceID':
+        final_loader_template_name = "IPAdapterUnifiedLoaderFaceID"
+
+    final_loader_id = assembler._get_unique_id()
+    final_loader_node = assembler._get_node_template(final_loader_template_name)
+    final_loader_node['inputs']['model'] = current_model_connection
+    final_loader_node['inputs']['preset'] = final_settings.get('final_preset', 'STANDARD (medium strength)')
+    if final_loader_type == 'FaceID':
+        final_loader_node['inputs']['lora_strength'] = final_settings.get('final_lora_strength', 0.6)
+    assembler.workflow[final_loader_id] = final_loader_node
+
+    apply_embeds_id = assembler._get_unique_id()
+    apply_embeds_node = assembler._get_node_template("IPAdapterEmbeds")
+    apply_embeds_node['inputs']['weight'] = final_settings.get('final_weight', 1.0)
+    apply_embeds_node['inputs']['embeds_scaling'] = final_settings.get('final_embeds_scaling', 'V only')
+    apply_embeds_node['inputs']['model'] = [final_loader_id, 0]
+    apply_embeds_node['inputs']['ipadapter'] = [final_loader_id, 1]
+    apply_embeds_node['inputs']['pos_embed'] = [pos_combiner_id, 0]
+    apply_embeds_node['inputs']['neg_embed'] = [neg_combiner_id, 0]
+    assembler.workflow[apply_embeds_id] = apply_embeds_node
+
+    assembler.workflow[end_node_id]['inputs']['model'] = [apply_embeds_id, 0]
+    print(f"IPAdapter injector applied. Redirected '{end_node_name}' model input through {len(chain_items)} reference images.")

chain_injectors/newbie_lora_injector.py

@@ -0,0 +1,63 @@
+from copy import deepcopy
+
+def inject(assembler, chain_definition, chain_items):
+    if not chain_items:
+        return
+
+    output_map = chain_definition.get('output_map', {})
+    current_connections = {}
+    for key, type_name in output_map.items():
+        if ':' in str(key):
+            node_name, idx_str = key.split(':')
+            if node_name not in assembler.node_map:
+                print(f"Warning: [NewBie LoRA Injector] Node '{node_name}' in chain's output_map not found. Skipping.")
+                continue
+            node_id = assembler.node_map[node_name]
+            start_output_idx = int(idx_str)
+            current_connections[type_name] = [node_id, start_output_idx]
+        else:
+            print(f"Warning: [NewBie LoRA Injector] output_map key '{key}' is not in 'node:index' format. Skipping this connection.")
+
+    template_name = chain_definition.get('template')
+    if not template_name:
+        print(f"Warning: [NewBie LoRA Injector] No 'template' defined for chain. Skipping.")
+        return
+
+    for item_data in chain_items:
+        template = assembler._get_node_template(template_name)
+        node_data = deepcopy(template)
+        
+        node_data['inputs']['lora_name'] = item_data.get('lora_name')
+        node_data['inputs']['strength'] = item_data.get('strength_model', 1.0)
+        node_data['inputs']['enabled'] = True
+        
+        if 'model' in current_connections:
+            node_data['inputs']['model'] = current_connections['model']
+        if 'clip' in current_connections:
+            node_data['inputs']['clip'] = current_connections['clip']
+
+        new_node_id = assembler._get_unique_id()
+        assembler.workflow[new_node_id] = node_data
+        
+        current_connections['model'] = [new_node_id, 0]
+        current_connections['clip'] = [new_node_id, 1]
+
+    end_input_map = chain_definition.get('end_input_map', {})
+    for type_name, targets in end_input_map.items():
+        if type_name in current_connections:
+            if not isinstance(targets, list):
+                targets = [targets]
+            
+            for target_str in targets:
+                try:
+                    end_node_name, end_input_name = target_str.split(':')
+                    if end_node_name in assembler.node_map:
+                        end_node_id = assembler.node_map[end_node_name]
+                        assembler.workflow[end_node_id]['inputs'][end_input_name] = current_connections[type_name]
+                    else:
+                        print(f"Warning: [NewBie LoRA Injector] End node '{end_node_name}' for dynamic chain not found. Skipping connection.")
+                except ValueError:
+                    print(f"Warning: [NewBie LoRA Injector] Invalid target format '{target_str}' in end_input_map. Skipping.")
+
+    if chain_items:
+        print(f"NewBie LoRA injector applied. Re-routed model and clip through {len(chain_items)} LoRA(s).")

chain_injectors/reference_latent_injector.py

@@ -0,0 +1,157 @@
+def inject(assembler, chain_definition, chain_items):
+    if not chain_items:
+        return
+
+    guider_node_name = chain_definition.get('guider_node')
+    guider_target_inputs = chain_definition.get('guider_target_inputs', [])
+    start_connections_map = chain_definition.get('start_connections', {})
+    vae_node_name = chain_definition.get('vae_node', 'vae_loader')
+
+    if guider_node_name and guider_node_name in assembler.node_map and guider_target_inputs:
+        guider_id = assembler.node_map[guider_node_name]
+        if vae_node_name not in assembler.node_map:
+            print(f"Warning: VAE node '{vae_node_name}' not found for Guider chain. Skipping.")
+            return
+        vae_node_id = assembler.node_map[vae_node_name]
+        
+        print(f"ReferenceLatent injector targeting DualCFGGuider node '{guider_node_name}'.")
+
+        current_connections = {}
+        for target_input in guider_target_inputs:
+            conn_str = start_connections_map.get(target_input)
+            if not conn_str:
+                print(f"Warning: No start connection defined for '{target_input}' in Guider chain. Skipping this input.")
+                continue
+            try:
+                node_name, idx_str = conn_str.split(':')
+                node_id = assembler.node_map[node_name]
+                current_connections[target_input] = [node_id, int(idx_str)]
+            except (ValueError, KeyError):
+                print(f"Warning: Invalid start connection '{conn_str}' for '{target_input}'. Skipping.")
+
+        encoded_latents = []
+        for i, img_filename in enumerate(chain_items):
+            load_id = assembler._get_unique_id()
+            load_node = assembler._get_node_template("LoadImage")
+            load_node['inputs']['image'] = img_filename
+            assembler.workflow[load_id] = load_node
+            
+            scale_id = assembler._get_unique_id()
+            scale_node = assembler._get_node_template("ImageScaleToTotalPixels")
+            scale_node['inputs']['megapixels'] = 1.0
+            scale_node['inputs']['upscale_method'] = "lanczos"
+            scale_node['inputs']['image'] = [load_id, 0]
+            assembler.workflow[scale_id] = scale_node
+            
+            vae_encode_id = assembler._get_unique_id()
+            vae_encode_node = assembler._get_node_template("VAEEncode")
+            vae_encode_node['inputs']['pixels'] = [scale_id, 0]
+            vae_encode_node['inputs']['vae'] = [vae_node_id, 0]
+            assembler.workflow[vae_encode_id] = vae_encode_node
+            encoded_latents.append([vae_encode_id, 0])
+        
+        for target_input_name, start_connection in current_connections.items():
+            current_chain_head = start_connection
+            for i, latent_conn in enumerate(encoded_latents):
+                ref_latent_id = assembler._get_unique_id()
+                ref_latent_node = assembler._get_node_template("ReferenceLatent")
+                ref_latent_node['inputs']['conditioning'] = current_chain_head
+                ref_latent_node['inputs']['latent'] = latent_conn
+                ref_latent_node['_meta']['title'] = f"{target_input_name} RefLatent {i+1}"
+                assembler.workflow[ref_latent_id] = ref_latent_node
+                current_chain_head = [ref_latent_id, 0]
+            
+            assembler.workflow[guider_id]['inputs'][target_input_name] = current_chain_head
+            print(f"  - Input '{target_input_name}' of node '{guider_node_name}' re-routed through {len(chain_items)} reference images.")
+
+        return
+
+    flux_guidance_name = chain_definition.get('flux_guidance_node')
+    ksampler_name = chain_definition.get('ksampler_node', 'ksampler')
+
+    if ksampler_name not in assembler.node_map:
+        print(f"Warning: KSampler node '{ksampler_name}' not found for ReferenceLatent chain. Skipping.")
+        return
+    if vae_node_name not in assembler.node_map:
+        print(f"Warning: VAE loader node '{vae_node_name}' not found for ReferenceLatent chain. Skipping.")
+        return
+
+    ksampler_id = assembler.node_map[ksampler_name]
+    vae_node_id = assembler.node_map[vae_node_name]
+
+    pos_target_node_id = None
+    pos_target_input_name = None
+    if flux_guidance_name and flux_guidance_name in assembler.node_map:
+        flux_guidance_id = assembler.node_map[flux_guidance_name]
+        if 'conditioning' in assembler.workflow[flux_guidance_id]['inputs']:
+            pos_target_node_id = flux_guidance_id
+            pos_target_input_name = 'conditioning'
+            print(f"ReferenceLatent injector targeting FluxGuidance node '{flux_guidance_name}' for positive chain.")
+    
+    if not pos_target_node_id:
+        if 'positive' in assembler.workflow[ksampler_id]['inputs']:
+            pos_target_node_id = ksampler_id
+            pos_target_input_name = 'positive'
+            print(f"ReferenceLatent injector targeting KSampler node '{ksampler_name}' for positive chain.")
+        else:
+            print(f"Warning: Could not find a valid positive injection point for ReferenceLatent chain. Skipping.")
+            return
+
+    current_pos_conditioning = assembler.workflow[pos_target_node_id]['inputs'][pos_target_input_name]
+
+    neg_target_node_id = ksampler_id
+    neg_target_input_name = 'negative'
+    if 'negative' not in assembler.workflow[neg_target_node_id]['inputs']:
+        print(f"Warning: KSampler node '{ksampler_name}' has no 'negative' input. Skipping negative ReferenceLatent chain.")
+        neg_target_node_id = None
+    
+    current_neg_conditioning = None
+    if neg_target_node_id:
+        current_neg_conditioning = assembler.workflow[neg_target_node_id]['inputs'][neg_target_input_name]
+
+    for i, img_filename in enumerate(chain_items):
+        load_id = assembler._get_unique_id()
+        load_node = assembler._get_node_template("LoadImage")
+        load_node['inputs']['image'] = img_filename
+        load_node['_meta']['title'] = f"Load Reference Image {i+1}"
+        assembler.workflow[load_id] = load_node
+        
+        scale_id = assembler._get_unique_id()
+        scale_node = assembler._get_node_template("ImageScaleToTotalPixels")
+        scale_node['inputs']['megapixels'] = 1.0
+        scale_node['inputs']['upscale_method'] = "lanczos"
+        scale_node['inputs']['image'] = [load_id, 0]
+        scale_node['_meta']['title'] = f"Scale Reference {i+1}"
+        assembler.workflow[scale_id] = scale_node
+        
+        vae_encode_id = assembler._get_unique_id()
+        vae_encode_node = assembler._get_node_template("VAEEncode")
+        vae_encode_node['inputs']['pixels'] = [scale_id, 0]
+        vae_encode_node['inputs']['vae'] = [vae_node_id, 0]
+        vae_encode_node['_meta']['title'] = f"VAE Encode Reference {i+1}"
+        assembler.workflow[vae_encode_id] = vae_encode_node
+        
+        latent_conn = [vae_encode_id, 0]
+
+        pos_ref_latent_id = assembler._get_unique_id()
+        pos_ref_latent_node = assembler._get_node_template("ReferenceLatent")
+        pos_ref_latent_node['inputs']['conditioning'] = current_pos_conditioning
+        pos_ref_latent_node['inputs']['latent'] = latent_conn
+        pos_ref_latent_node['_meta']['title'] = f"Positive ReferenceLatent {i+1}"
+        assembler.workflow[pos_ref_latent_id] = pos_ref_latent_node
+        current_pos_conditioning = [pos_ref_latent_id, 0]
+
+        if neg_target_node_id:
+            neg_ref_latent_id = assembler._get_unique_id()
+            neg_ref_latent_node = assembler._get_node_template("ReferenceLatent")
+            neg_ref_latent_node['inputs']['conditioning'] = current_neg_conditioning
+            neg_ref_latent_node['inputs']['latent'] = latent_conn
+            neg_ref_latent_node['_meta']['title'] = f"Negative ReferenceLatent {i+1}"
+            assembler.workflow[neg_ref_latent_id] = neg_ref_latent_node
+            current_neg_conditioning = [neg_ref_latent_id, 0]
+
+    assembler.workflow[pos_target_node_id]['inputs'][pos_target_input_name] = current_pos_conditioning
+    if neg_target_node_id:
+        assembler.workflow[neg_target_node_id]['inputs'][neg_target_input_name] = current_neg_conditioning
+    
+    print(f"ReferenceLatent injector applied. Re-routed inputs through {len(chain_items)} reference images.")

chain_injectors/sd3_ipadapter_injector.py

@@ -0,0 +1,66 @@
+def inject(assembler, chain_definition, chain_items):
+    if not chain_items:
+        return
+
+    ksampler_name = chain_definition.get('ksampler_node', 'ksampler')
+    if ksampler_name not in assembler.node_map:
+        print(f"Warning: KSampler node '{ksampler_name}' not found for SD3 IPAdapter chain. Skipping.")
+        return
+        
+    ksampler_id = assembler.node_map[ksampler_name]
+
+    if 'model' not in assembler.workflow[ksampler_id]['inputs']:
+        print(f"Warning: KSampler node '{ksampler_name}' is missing 'model' input. Skipping SD3 IPAdapter chain.")
+        return
+        
+    current_model_connection = assembler.workflow[ksampler_id]['inputs']['model']
+    
+    clip_vision_loader_id = assembler._get_unique_id()
+    clip_vision_loader_node = assembler._get_node_template("CLIPVisionLoader")
+    clip_vision_loader_node['inputs']['clip_name'] = "sigclip_vision_patch14_384.safetensors"
+    assembler.workflow[clip_vision_loader_id] = clip_vision_loader_node
+
+    ipadapter_loader_id = assembler._get_unique_id()
+    ipadapter_loader_node = assembler._get_node_template("IPAdapterSD3Loader")
+    ipadapter_loader_node['inputs']['ipadapter'] = "ip-adapter_sd35l_instantx.bin"
+    ipadapter_loader_node['inputs']['provider'] = "cuda"
+    assembler.workflow[ipadapter_loader_id] = ipadapter_loader_node
+
+    for item_data in chain_items:
+        image_loader_id = assembler._get_unique_id()
+        image_loader_node = assembler._get_node_template("LoadImage")
+        image_loader_node['inputs']['image'] = item_data['image']
+        assembler.workflow[image_loader_id] = image_loader_node
+
+        image_scaler_id = assembler._get_unique_id()
+        image_scaler_node = assembler._get_node_template("ImageScaleToTotalPixels")
+        image_scaler_node['inputs']['image'] = [image_loader_id, 0]
+        image_scaler_node['inputs']['upscale_method'] = 'nearest-exact'
+        image_scaler_node['inputs']['megapixels'] = 1.0
+        assembler.workflow[image_scaler_id] = image_scaler_node
+
+        clip_vision_encode_id = assembler._get_unique_id()
+        clip_vision_encode_node = assembler._get_node_template("CLIPVisionEncode")
+        clip_vision_encode_node['inputs']['crop'] = "center"
+        clip_vision_encode_node['inputs']['clip_vision'] = [clip_vision_loader_id, 0]
+        clip_vision_encode_node['inputs']['image'] = [image_scaler_id, 0]
+        assembler.workflow[clip_vision_encode_id] = clip_vision_encode_node
+        
+        apply_ipa_id = assembler._get_unique_id()
+        apply_ipa_node = assembler._get_node_template("ApplyIPAdapterSD3")
+        
+        apply_ipa_node['inputs']['weight'] = item_data.get('weight', 1.0)
+        apply_ipa_node['inputs']['start_percent'] = item_data.get('start_percent', 0.0)
+        apply_ipa_node['inputs']['end_percent'] = item_data.get('end_percent', 1.0)
+        
+        apply_ipa_node['inputs']['model'] = current_model_connection
+        apply_ipa_node['inputs']['ipadapter'] = [ipadapter_loader_id, 0]
+        apply_ipa_node['inputs']['image_embed'] = [clip_vision_encode_id, 0]
+        
+        assembler.workflow[apply_ipa_id] = apply_ipa_node
+        
+        current_model_connection = [apply_ipa_id, 0]
+
+    assembler.workflow[ksampler_id]['inputs']['model'] = current_model_connection
+    
+    print(f"SD3 IPAdapter injector applied. KSampler model input re-routed through {len(chain_items)} IPAdapter(s).")

chain_injectors/style_injector.py

@@ -0,0 +1,71 @@
+def inject(assembler, chain_definition, chain_items):
+    if not chain_items:
+        return
+
+    flux_guidance_name = chain_definition.get('flux_guidance_node')
+    ksampler_name = chain_definition.get('ksampler_node', 'ksampler')
+
+    target_node_id = None
+    target_input_name = None
+    
+    if flux_guidance_name and flux_guidance_name in assembler.node_map:
+        flux_guidance_id = assembler.node_map[flux_guidance_name]
+        if 'conditioning' in assembler.workflow[flux_guidance_id]['inputs']:
+            target_node_id = flux_guidance_id
+            target_input_name = 'conditioning'
+
+    if not target_node_id:
+        if ksampler_name in assembler.node_map:
+            ksampler_id = assembler.node_map[ksampler_name]
+            if 'positive' in assembler.workflow[ksampler_id]['inputs']:
+                target_node_id = ksampler_id
+                target_input_name = 'positive'
+        else:
+            return
+            
+    if not target_node_id:
+        return
+
+    current_conditioning = assembler.workflow[target_node_id]['inputs'][target_input_name]
+
+    style_model_loader_id = assembler._get_unique_id()
+    style_model_loader_node = assembler._get_node_template("StyleModelLoader")
+    style_model_loader_node['inputs']['style_model_name'] = "flux1-redux-dev.safetensors"
+    assembler.workflow[style_model_loader_id] = style_model_loader_node
+
+    clip_vision_loader_id = assembler._get_unique_id()
+    clip_vision_loader_node = assembler._get_node_template("CLIPVisionLoader")
+    clip_vision_loader_node['inputs']['clip_name'] = "sigclip_vision_patch14_384.safetensors"
+    assembler.workflow[clip_vision_loader_id] = clip_vision_loader_node
+
+    for item_data in chain_items:
+        image = item_data.get('image')
+        strength = item_data.get('strength', 1.0)
+        if not image or strength is None:
+            continue
+
+        load_image_id = assembler._get_unique_id()
+        clip_vision_encode_id = assembler._get_unique_id()
+        style_apply_id = assembler._get_unique_id()
+
+        load_image_node = assembler._get_node_template("LoadImage")
+        clip_vision_encode_node = assembler._get_node_template("CLIPVisionEncode")
+        style_apply_node = assembler._get_node_template("StyleModelApply")
+        
+        load_image_node['inputs']['image'] = image
+        clip_vision_encode_node['inputs']['crop'] = "center"
+        clip_vision_encode_node['inputs']['clip_vision'] = [clip_vision_loader_id, 0]
+        clip_vision_encode_node['inputs']['image'] = [load_image_id, 0]
+
+        style_apply_node['inputs']['strength'] = strength
+        style_apply_node['inputs']['strength_type'] = "multiply"
+        style_apply_node['inputs']['conditioning'] = current_conditioning
+        style_apply_node['inputs']['style_model'] = [style_model_loader_id, 0]
+        style_apply_node['inputs']['clip_vision_output'] = [clip_vision_encode_id, 0]
+        
+        assembler.workflow[load_image_id] = load_image_node
+        assembler.workflow[clip_vision_encode_id] = clip_vision_encode_node
+        assembler.workflow[style_apply_id] = style_apply_node
+        current_conditioning = [style_apply_id, 0]
+
+    assembler.workflow[target_node_id]['inputs'][target_input_name] = current_conditioning

chain_injectors/vae_injector.py

@@ -0,0 +1,30 @@
+def inject(assembler, chain_definition, chain_items):
+    if not chain_items:
+        return
+
+    vae_name = chain_items[0] if isinstance(chain_items, list) else chain_items
+    if not vae_name or vae_name == "None":
+        return
+
+    targets = chain_definition.get('targets', [])
+    if not targets:
+        return
+
+    vae_loader_id = assembler._get_unique_id()
+    vae_loader_node = assembler._get_node_template("VAELoader")
+    vae_loader_node['inputs']['vae_name'] = vae_name
+    assembler.workflow[vae_loader_id] = vae_loader_node
+
+    injected_count = 0
+    for target_str in targets:
+        try:
+            node_name, input_name = target_str.split(':')
+            if node_name in assembler.node_map:
+                node_id = assembler.node_map[node_name]
+                assembler.workflow[node_id]['inputs'][input_name] = [vae_loader_id, 0]
+                injected_count += 1
+        except ValueError:
+            print(f"Warning: Invalid VAE injector target format '{target_str}'. Expected 'node_name:input_name'.")
+    
+    if injected_count > 0:
+        print(f"VAE injector applied. Rerouted {injected_count} connection(s) to new VAELoader ({vae_name}).")

comfy_integration/nodes.py

@@ -23,6 +23,11 @@ CLIPTextEncodeSDXL = NODE_CLASS_MAPPINGS['CLIPTextEncodeSDXL']
 LoraLoader = NODE_CLASS_MAPPINGS['LoraLoader']
 CLIPSetLastLayer = NODE_CLASS_MAPPINGS['CLIPSetLastLayer']
 
+if 'EmptyHunyuanImageLatent' in NODE_CLASS_MAPPINGS:
+    EmptyHunyuanImageLatent = NODE_CLASS_MAPPINGS['EmptyHunyuanImageLatent']
+else:
+    print("⚠️ Warning: 'EmptyHunyuanImageLatent' not found in NODE_CLASS_MAPPINGS. HunyuanImage txt2img may fail if this node is required.")
+
 try:
     KSamplerNode = NODE_CLASS_MAPPINGS['KSampler']
     SAMPLER_CHOICES = KSamplerNode.INPUT_TYPES()["required"]["sampler_name"][0]

comfy_integration/setup.py

@@ -39,14 +39,48 @@ def initialize_comfyui():
     except OSError as e:
         print(f"⚠️ Could not remove temporary directory '{COMFYUI_TEMP_DIR}': {e}")
 
+
     print("--- Cloning third-party extensions for ComfyUI ---")
-    controlnet_aux_path = os.path.join(APP_DIR, "custom_nodes", "comfyui_controlnet_aux")
-    if not os.path.exists(controlnet_aux_path):
-        os.system(f"git clone https://github.com/Fannovel16/comfyui_controlnet_aux.git {controlnet_aux_path}")
-        print("✅ comfyui_controlnet_aux extension cloned.")
+
+    # 1. ComfyUI_IPAdapter_plus
+    ipadapter_plus_path = os.path.join(APP_DIR, "custom_nodes", "ComfyUI_IPAdapter_plus")
+    if not os.path.exists(ipadapter_plus_path):
+        os.system(f"git clone https://github.com/cubiq/ComfyUI_IPAdapter_plus.git {ipadapter_plus_path}")
+        print("✅ ComfyUI_IPAdapter_plus extension cloned.")
+    else:
+        print("✅ ComfyUI_IPAdapter_plus extension already exists.")
+
+    # 2. ComfyUI-InstantX-IPAdapter-SD3
+    ipadapter_plus_path = os.path.join(APP_DIR, "custom_nodes", "ComfyUI-InstantX-IPAdapter-SD3")
+    if not os.path.exists(ipadapter_plus_path):
+        os.system(f"git clone https://github.com/Slickytail/ComfyUI-InstantX-IPAdapter-SD3.git {ipadapter_plus_path}")
+        print("✅ ComfyUI-InstantX-IPAdapter-SD3 extension cloned.")
     else:
-        print("✅ comfyui_controlnet_aux extension already exists.")
+        print("✅ ComfyUI-InstantX-IPAdapter-SD3 extension already exists.")
 
+    # 3. ComfyUI-IPAdapter-Flux
+    ipadapter_flux_path = os.path.join(APP_DIR, "custom_nodes", "ComfyUI-IPAdapter-Flux")
+    if not os.path.exists(ipadapter_flux_path):
+        os.system(f"git clone https://github.com/Shakker-Labs/ComfyUI-IPAdapter-Flux.git {ipadapter_flux_path}")
+        print("✅ ComfyUI-IPAdapter-Flux extension cloned.")
+    else:
+        print("✅ ComfyUI-IPAdapter-Flux extension already exists.")
+
+    # 4. ComfyUI-Newbie-Nodes
+    newbie_nodes_path = os.path.join(APP_DIR, "custom_nodes", "ComfyUI-Newbie-Nodes")
+    if not os.path.exists(newbie_nodes_path):
+        os.system(f"git clone https://github.com/NewBieAI-Lab/ComfyUI-Newbie-Nodes.git {newbie_nodes_path}")
+        print("✅ ComfyUI-Newbie-Nodes extension cloned.")
+    else:
+        print("✅ ComfyUI-Newbie-Nodes extension already exists.")
+
+    # 5. ComfyUI-Anima-LLLite
+    anima_controlnet_lllite_nodes_path = os.path.join(APP_DIR, "custom_nodes", "ComfyUI-Anima-LLLite")
+    if not os.path.exists(anima_controlnet_lllite_nodes_path):
+        os.system(f"git clone https://github.com/kohya-ss/ComfyUI-Anima-LLLite.git {anima_controlnet_lllite_nodes_path}")
+        print("✅ ComfyUI-Anima-LLLite extension cloned.")
+    else:
+        print("✅ ComfyUI-Anima-LLLite extension already exists.")
 
     print(f"✅ Current working directory is: {os.getcwd()}")
     
@@ -55,13 +89,10 @@ def initialize_comfyui():
 
     print("✅ ComfyUI initialized with default attention mechanism.")
 
-    os.makedirs(os.path.join(APP_DIR, CHECKPOINT_DIR), exist_ok=True)
-    os.makedirs(os.path.join(APP_DIR, LORA_DIR), exist_ok=True)
-    os.makedirs(os.path.join(APP_DIR, EMBEDDING_DIR), exist_ok=True)
-    os.makedirs(os.path.join(APP_DIR, CONTROLNET_DIR), exist_ok=True)
-    os.makedirs(os.path.join(APP_DIR, MODEL_PATCHES_DIR), exist_ok=True)
-    os.makedirs(os.path.join(APP_DIR, DIFFUSION_MODELS_DIR), exist_ok=True)
-    os.makedirs(os.path.join(APP_DIR, VAE_DIR), exist_ok=True)
-    os.makedirs(os.path.join(APP_DIR, TEXT_ENCODERS_DIR), exist_ok=True)
+    for dir_path in CATEGORY_TO_DIR_MAP.values():
+        os.makedirs(os.path.join(APP_DIR, dir_path), exist_ok=True)
+    
     os.makedirs(os.path.join(APP_DIR, INPUT_DIR), exist_ok=True)
+    os.makedirs(os.path.join(APP_DIR, OUTPUT_DIR), exist_ok=True)
+    
     print("✅ All required model directories are present.")

core/generation_logic.py

@@ -1,25 +1,10 @@
 from typing import Any, Dict
 import gradio as gr
 
-from core.pipelines.controlnet_preprocessor import ControlNetPreprocessorPipeline
 from core.pipelines.sd_image_pipeline import SdImagePipeline
 
-controlnet_preprocessor_pipeline = ControlNetPreprocessorPipeline()
 sd_image_pipeline = SdImagePipeline()
 
 
-def build_reverse_map():
-    from nodes import NODE_DISPLAY_NAME_MAPPINGS
-    import core.pipelines.controlnet_preprocessor as cn_module
-    
-    if cn_module.REVERSE_DISPLAY_NAME_MAP is None:
-        cn_module.REVERSE_DISPLAY_NAME_MAP = {v: k for k, v in NODE_DISPLAY_NAME_MAPPINGS.items()}
-        if "Semantic Segmentor (legacy, alias for UniFormer)" not in cn_module.REVERSE_DISPLAY_NAME_MAP:
-             cn_module.REVERSE_DISPLAY_NAME_MAP["Semantic Segmentor (legacy, alias for UniFormer)"] = "SemSegPreprocessor"
-
-
-def run_cn_preprocessor_entry(*args, **kwargs):
-    return controlnet_preprocessor_pipeline.run(*args, **kwargs)
-
 def generate_image_wrapper(ui_inputs: dict, progress=gr.Progress(track_tqdm=True)):
     return sd_image_pipeline.run(ui_inputs=ui_inputs, progress=progress)

core/model_manager.py

@@ -1,11 +1,8 @@
 import gc
 from typing import List
-
 import gradio as gr
-
-from core.settings import ALL_MODEL_MAP
 from utils.app_utils import _ensure_model_downloaded
-
+from core.settings import ALL_MODEL_MAP
 
 class ModelManager:
     _instance = None
@@ -23,26 +20,13 @@ class ModelManager:
 
     def ensure_models_downloaded(self, required_models: List[str], progress):
         print(f"--- [ModelManager] Ensuring models are downloaded: {required_models} ---")
-        
-        files_to_download = set()
-        for display_name in required_models:
-            if display_name in ALL_MODEL_MAP:
-                _, components, _, _ = ALL_MODEL_MAP[display_name]
-                for component_key, component_file in components.items():
-                    if component_key in ['unet', 'clip', 'vae', 'lora']:
-                        files_to_download.add(component_file)
-        
-        files_to_download = list(files_to_download)
-        total_files = len(files_to_download)
-        
-        for i, filename in enumerate(files_to_download):
+        for i, display_name in enumerate(required_models):
             if progress and hasattr(progress, '__call__'):
-                progress(i / total_files, desc=f"Checking file: {filename}")
+                progress(i / max(len(required_models), 1), desc=f"Checking file: {display_name}")
             try:
-                _ensure_model_downloaded(filename, progress)
+                _ensure_model_downloaded(display_name, progress)
             except Exception as e:
-                raise gr.Error(f"Failed to download model component '{filename}'. Reason: {e}")
-
+                raise gr.Error(f"Failed to download model '{display_name}'. Reason: {e}")
         print(f"--- [ModelManager] ✅ All required models are present on disk. ---")
     
 model_manager = ModelManager()

core/pipelines/sd_image_pipeline.py

@@ -11,12 +11,20 @@ import numpy as np
 from .base_pipeline import BasePipeline
 from core.settings import *
 from comfy_integration.nodes import *
-from utils.app_utils import get_value_at_index, sanitize_prompt, get_lora_path, get_embedding_path, ensure_controlnet_model_downloaded, sanitize_filename
+from utils.app_utils import get_value_at_index, sanitize_prompt, get_lora_path, get_embedding_path, ensure_controlnet_model_downloaded, ensure_ipadapter_models_downloaded, sanitize_filename
 from core.workflow_assembler import WorkflowAssembler
 
 class SdImagePipeline(BasePipeline):
     def get_required_models(self, model_display_name: str, **kwargs) -> List[str]:
-        return [model_display_name]
+        model_info = ALL_MODEL_MAP.get(model_display_name)
+        if not model_info:
+            return [model_display_name]
+
+        path_or_components = model_info[1]
+        if isinstance(path_or_components, dict):
+            return [v for v in path_or_components.values() if v and v != "pixel_space"]
+        else:
+            return [model_display_name]
 
     def _topological_sort(self, workflow: Dict[str, Any]) -> List[str]:
         graph = defaultdict(list)
@@ -47,7 +55,6 @@ class SdImagePipeline(BasePipeline):
             
         return sorted_nodes
 
-
     def _execute_workflow(self, workflow: Dict[str, Any], initial_objects: Dict[str, Any]):
         with torch.no_grad():
             computed_outputs = initial_objects
@@ -113,13 +120,13 @@ class SdImagePipeline(BasePipeline):
             
             return get_value_at_index(computed_outputs[image_source_node_id], image_source_index)
 
-    def _gpu_logic(self, ui_inputs: Dict, loras_string: str, required_models_for_gpu: List[str], workflow: Dict[str, Any], assembler: WorkflowAssembler, progress=gr.Progress(track_tqdm=True)):
+    def _gpu_logic(self, ui_inputs: Dict, loras_string: str, workflow: Dict[str, Any], assembler: WorkflowAssembler, progress=gr.Progress(track_tqdm=True)):
         model_display_name = ui_inputs['model_display_name']
         
         progress(0.4, desc="Executing workflow...")
         
         initial_objects = {}
-            
+
         decoded_images_tensor = self._execute_workflow(workflow, initial_objects=initial_objects)
         
         output_images = []
@@ -135,6 +142,7 @@ class SdImagePipeline(BasePipeline):
             params_string = f"{ui_inputs['positive_prompt']}\nNegative prompt: {ui_inputs['negative_prompt']}\n"
             params_string += f"Steps: {ui_inputs['num_inference_steps']}, Sampler: {ui_inputs['sampler']}, Scheduler: {ui_inputs['scheduler']}, CFG scale: {ui_inputs['guidance_scale']}, Seed: {current_seed}, Size: {width_for_meta}x{height_for_meta}, Base Model: {model_display_name}"
             if ui_inputs['task_type'] != 'txt2img': params_string += f", Denoise: {ui_inputs['denoise']}"
+            if ui_inputs.get('clip_skip') and ui_inputs['clip_skip'] != 1: params_string += f", Clip skip: {abs(ui_inputs['clip_skip'])}"
             if loras_string: params_string += f", {loras_string}"
             
             pil_image.info = {'parameters': params_string.strip()}
@@ -146,51 +154,58 @@ class SdImagePipeline(BasePipeline):
         progress(0, desc="Preparing models...")
         
         task_type = ui_inputs['task_type']
+        model_display_name = ui_inputs['model_display_name']
+        model_type = MODEL_TYPE_MAP.get(model_display_name, 'sdxl')
+        
+        architectures_dict = ARCHITECTURES_CONFIG.get('architectures', {})
+        workflow_model_type = architectures_dict.get(model_type, {}).get("model_type", "sdxl")
         
         ui_inputs['positive_prompt'] = sanitize_prompt(ui_inputs.get('positive_prompt', ''))
         ui_inputs['negative_prompt'] = sanitize_prompt(ui_inputs.get('negative_prompt', ''))
         
-        required_models = self.get_required_models(model_display_name=ui_inputs['model_display_name'])
-        
+        if 'clip_skip' in ui_inputs and ui_inputs['clip_skip'] is not None:
+             ui_inputs['clip_skip'] = -int(ui_inputs['clip_skip'])
+        else:
+             ui_inputs['clip_skip'] = -1
+
+        required_models = self.get_required_models(model_display_name=model_display_name)
         self.model_manager.ensure_models_downloaded(required_models, progress=progress)
         
         lora_data = ui_inputs.get('lora_data', [])
         active_loras_for_gpu, active_loras_for_meta = [], []
-        
         if lora_data:
             sources, ids, scales, files = lora_data[0::4], lora_data[1::4], lora_data[2::4], lora_data[3::4]
-            
             for i, (source, lora_id, scale, _) in enumerate(zip(sources, ids, scales, files)):
                 if scale > 0 and lora_id and lora_id.strip():
                     lora_filename = None
                     if source == "File":
                         lora_filename = sanitize_filename(lora_id)
                     elif source == "Civitai":
-                        local_path, status = get_lora_path(source, lora_id, ui_inputs['civitai_api_key'], progress)
+                        local_path, status = get_lora_path(source, lora_id, os.environ.get("CIVITAI_API_KEY", ""), progress)
                         if local_path: lora_filename = os.path.basename(local_path)
                         else: raise gr.Error(f"Failed to prepare LoRA {lora_id}: {status}")
                     
                     if lora_filename:
                         active_loras_for_gpu.append({"lora_name": lora_filename, "strength_model": scale, "strength_clip": scale})
                         active_loras_for_meta.append(f"{source} {lora_id}:{scale}")
-        
+
         ui_inputs['denoise'] = 1.0
         if task_type == 'img2img': ui_inputs['denoise'] = ui_inputs.get('img2img_denoise', 0.7)
         elif task_type == 'hires_fix': ui_inputs['denoise'] = ui_inputs.get('hires_denoise', 0.55)
         
         temp_files_to_clean = []
-        
         if not os.path.exists(INPUT_DIR): os.makedirs(INPUT_DIR)
 
         if task_type == 'img2img':
             input_image_pil = ui_inputs.get('img2img_image')
-            if input_image_pil:
-                temp_file_path = os.path.join(INPUT_DIR, f"temp_input_{random.randint(1000, 9999)}.png")
-                input_image_pil.save(temp_file_path, "PNG")
-                ui_inputs['input_image'] = os.path.basename(temp_file_path)
-                temp_files_to_clean.append(temp_file_path)
-                ui_inputs['width'] = input_image_pil.width
-                ui_inputs['height'] = input_image_pil.height
+            if not input_image_pil:
+                raise gr.Error("Please upload an image for Image-to-Image.")
+            temp_file_path = os.path.join(INPUT_DIR, f"temp_input_{random.randint(1000, 9999)}.png")
+            input_image_pil.save(temp_file_path, "PNG")
+            ui_inputs['input_image'] = os.path.basename(temp_file_path)
+            temp_files_to_clean.append(temp_file_path)
+            ui_inputs['width'] = input_image_pil.width
+            ui_inputs['height'] = input_image_pil.height
         
         elif task_type == 'inpaint':
             inpaint_dict = ui_inputs.get('inpaint_image_dict')
@@ -198,7 +213,6 @@ class SdImagePipeline(BasePipeline):
                 raise gr.Error("Inpainting requires an input image and a drawn mask.")
             
             background_img = inpaint_dict['background'].convert("RGBA")
-            
             composite_mask_pil = Image.new('L', background_img.size, 0)
             for layer in inpaint_dict['layers']:
                 if layer:
@@ -212,25 +226,30 @@ class SdImagePipeline(BasePipeline):
             temp_file_path = os.path.join(INPUT_DIR, f"temp_inpaint_composite_{random.randint(1000, 9999)}.png")
             composite_image_with_mask.save(temp_file_path, "PNG")
             
-            ui_inputs['inpaint_image'] = os.path.basename(temp_file_path)
+            ui_inputs['input_image'] = os.path.basename(temp_file_path)
             temp_files_to_clean.append(temp_file_path)
             ui_inputs.pop('inpaint_mask', None)
 
         elif task_type == 'outpaint':
             input_image_pil = ui_inputs.get('outpaint_image')
-            if input_image_pil:
-                temp_file_path = os.path.join(INPUT_DIR, f"temp_input_{random.randint(1000, 9999)}.png")
-                input_image_pil.save(temp_file_path, "PNG")
-                ui_inputs['input_image'] = os.path.basename(temp_file_path)
-                temp_files_to_clean.append(temp_file_path)
+            if not input_image_pil:
+                raise gr.Error("Please upload an image for Outpainting.")
+            temp_file_path = os.path.join(INPUT_DIR, f"temp_input_{random.randint(1000, 9999)}.png")
+            input_image_pil.save(temp_file_path, "PNG")
+            ui_inputs['input_image'] = os.path.basename(temp_file_path)
+            temp_files_to_clean.append(temp_file_path)
+            
+            ui_inputs['megapixels'] = 0.25
+            ui_inputs['grow_mask_by'] = ui_inputs.get('feathering', 10)
 
         elif task_type == 'hires_fix':
             input_image_pil = ui_inputs.get('hires_image')
-            if input_image_pil:
-                temp_file_path = os.path.join(INPUT_DIR, f"temp_input_{random.randint(1000, 9999)}.png")
-                input_image_pil.save(temp_file_path, "PNG")
-                ui_inputs['input_image'] = os.path.basename(temp_file_path)
-                temp_files_to_clean.append(temp_file_path)
+            if not input_image_pil:
+                raise gr.Error("Please upload an image for Hires Fix.")
+            temp_file_path = os.path.join(INPUT_DIR, f"temp_input_{random.randint(1000, 9999)}.png")
+            input_image_pil.save(temp_file_path, "PNG")
+            ui_inputs['input_image'] = os.path.basename(temp_file_path)
+            temp_files_to_clean.append(temp_file_path)
 
         embedding_data = ui_inputs.get('embedding_data', [])
         embedding_filenames = []
@@ -242,7 +261,7 @@ class SdImagePipeline(BasePipeline):
                     if source == "File":
                         emb_filename = sanitize_filename(emb_id)
                     elif source == "Civitai":
-                        local_path, status = get_embedding_path(source, emb_id, ui_inputs['civitai_api_key'], progress)
+                        local_path, status = get_embedding_path(source, emb_id, os.environ.get("CIVITAI_API_KEY", ""), progress)
                         if local_path: emb_filename = os.path.basename(local_path)
                         else: raise gr.Error(f"Failed to prepare Embedding {emb_id}: {status}")
                     
@@ -263,7 +282,6 @@ class SdImagePipeline(BasePipeline):
             for i in range(len(cn_images)):
                 if cn_images[i] and cn_strengths[i] > 0 and cn_filepaths[i] and cn_filepaths[i] != "None":
                     ensure_controlnet_model_downloaded(cn_filepaths[i], progress)
-
                     if not os.path.exists(INPUT_DIR): os.makedirs(INPUT_DIR)
                     cn_temp_path = os.path.join(INPUT_DIR, f"temp_cn_{i}_{random.randint(1000, 9999)}.png")
                     cn_images[i].save(cn_temp_path, "PNG")
@@ -273,6 +291,23 @@ class SdImagePipeline(BasePipeline):
                         "start_percent": 0.0, "end_percent": 1.0, "control_net_name": cn_filepaths[i]
                     })
         
+        anima_controlnet_lllite_data = ui_inputs.get('anima_controlnet_lllite_data', [])
+        active_anima_controlnets = []
+        if anima_controlnet_lllite_data:
+            (cn_images, _, _, cn_strengths, cn_filepaths, cn_starts, cn_ends) = [anima_controlnet_lllite_data[i::7] for i in range(7)]
+            for i in range(len(cn_images)):
+                if cn_images[i] and cn_strengths[i] > 0 and cn_filepaths[i] and cn_filepaths[i] != "None":
+                    from utils.app_utils import _ensure_model_downloaded
+                    _ensure_model_downloaded(cn_filepaths[i], progress)
+                    if not os.path.exists(INPUT_DIR): os.makedirs(INPUT_DIR)
+                    cn_temp_path = os.path.join(INPUT_DIR, f"temp_anima_cn_{i}_{random.randint(1000, 9999)}.png")
+                    cn_images[i].save(cn_temp_path, "PNG")
+                    temp_files_to_clean.append(cn_temp_path)
+                    active_anima_controlnets.append({
+                        "image": os.path.basename(cn_temp_path), "strength": cn_strengths[i],
+                        "start_percent": cn_starts[i], "end_percent": cn_ends[i], "control_net_name": cn_filepaths[i]
+                    })
+
         diffsynth_controlnet_data = ui_inputs.get('diffsynth_controlnet_data', [])
         active_diffsynth_controlnets = []
         if diffsynth_controlnet_data:
@@ -289,21 +324,130 @@ class SdImagePipeline(BasePipeline):
                         "image": os.path.basename(cn_temp_path), "strength": cn_strengths[i],
                         "control_net_name": cn_filepaths[i]
                     })
-
+        
+        ipadapter_data = ui_inputs.get('ipadapter_data', [])
+        active_ipadapters = []
+        if ipadapter_data:
+            num_ipa_units = (len(ipadapter_data) - 5) // 3
+            final_preset, final_weight, final_lora_strength, final_embeds_scaling, final_combine_method = ipadapter_data[-5:]
+            ipa_images, ipa_weights, ipa_lora_strengths = [ipadapter_data[i*num_ipa_units:(i+1)*num_ipa_units] for i in range(3)]
+            all_presets_to_download = set()
+            for i in range(num_ipa_units):
+                if ipa_images[i] and ipa_weights[i] > 0 and final_preset:
+                    all_presets_to_download.add(final_preset)
+                    if not os.path.exists(INPUT_DIR): os.makedirs(INPUT_DIR)
+                    ipa_temp_path = os.path.join(INPUT_DIR, f"temp_ipa_{i}_{random.randint(1000, 9999)}.png")
+                    ipa_images[i].save(ipa_temp_path, "PNG")
+                    temp_files_to_clean.append(ipa_temp_path)
+                    active_ipadapters.append({
+                        "image": os.path.basename(ipa_temp_path), "preset": final_preset,
+                        "weight": ipa_weights[i], "lora_strength": ipa_lora_strengths[i]
+                    })
+            if active_ipadapters and final_preset:
+                all_presets_to_download.add(final_preset)
+            for preset in all_presets_to_download:
+                ensure_ipadapter_models_downloaded(preset, progress)
+                
+            model_type_key = 'sd15' if workflow_model_type == 'sd15' else 'sdxl'
+            if active_ipadapters:
+                active_ipadapters.append({
+                    'is_final_settings': True, 'model_type': model_type_key, 'final_preset': final_preset,
+                    'final_weight': final_weight, 'final_lora_strength': final_lora_strength,
+                    'final_embeds_scaling': final_embeds_scaling, 'final_combine_method': final_combine_method
+                })
+                
+        flux1_ipadapter_data = ui_inputs.get('flux1_ipadapter_data', [])
+        active_flux1_ipadapters = []
+        if flux1_ipadapter_data:
+            num_units = len(flux1_ipadapter_data) // 4
+            f_images = flux1_ipadapter_data[0*num_units : 1*num_units]
+            f_weights = flux1_ipadapter_data[1*num_units : 2*num_units]
+            f_starts = flux1_ipadapter_data[2*num_units : 3*num_units]
+            f_ends = flux1_ipadapter_data[3*num_units : 4*num_units]
+            for i in range(len(f_images)):
+                if f_images[i] and f_weights[i] > 0:
+                    from utils.app_utils import _ensure_model_downloaded
+                    for filename in ["ip-adapter.bin"]:
+                        _ensure_model_downloaded(filename, progress)
+                    
+                    from huggingface_hub import snapshot_download
+                    progress(0.5, desc="Caching HF SigLIP model...")
+                    snapshot_download(
+                        repo_id="google/siglip-so400m-patch14-384", 
+                        allow_patterns=["*.json", "*.safetensors", "*.txt"], 
+                        ignore_patterns=["*.msgpack", "*.h5", "*.bin"]
+                    )
+
+                    temp_path = os.path.join(INPUT_DIR, f"temp_fipa_{i}_{random.randint(1000, 9999)}.png")
+                    f_images[i].save(temp_path, "PNG")
+                    temp_files_to_clean.append(temp_path)
+                    active_flux1_ipadapters.append({
+                        "image": os.path.basename(temp_path),
+                        "weight": f_weights[i], "start_percent": f_starts[i], "end_percent": f_ends[i]
+                    })
+                    
+        sd3_ipadapter_data = ui_inputs.get('sd3_ipadapter_chain', [])
+        active_sd3_ipadapters = []
+        if sd3_ipadapter_data:
+            num_units = len(sd3_ipadapter_data) // 4
+            s_images = sd3_ipadapter_data[0*num_units : 1*num_units]
+            s_weights = sd3_ipadapter_data[1*num_units : 2*num_units]
+            s_starts = sd3_ipadapter_data[2*num_units : 3*num_units]
+            s_ends = sd3_ipadapter_data[3*num_units : 4*num_units]
+            sd3_ipa_downloaded = False
+            for i in range(len(s_images)):
+                if s_images[i] and s_weights[i] > 0:
+                    if not sd3_ipa_downloaded:
+                        from utils.app_utils import ensure_sd3_ipadapter_models_downloaded
+                        ensure_sd3_ipadapter_models_downloaded(progress)
+                        sd3_ipa_downloaded = True
+                    temp_path = os.path.join(INPUT_DIR, f"temp_s3ipa_{i}_{random.randint(1000, 9999)}.png")
+                    s_images[i].save(temp_path, "PNG")
+                    temp_files_to_clean.append(temp_path)
+                    active_sd3_ipadapters.append({
+                        "image": os.path.basename(temp_path),
+                        "weight": s_weights[i], "start_percent": s_starts[i], "end_percent": s_ends[i]
+                    })
+                    
+        style_data = ui_inputs.get('style_data', [])
+        active_styles = []
+        if style_data:
+            num_units = len(style_data) // 2
+            st_images = style_data[0*num_units : 1*num_units]
+            st_strengths = style_data[1*num_units : 2*num_units]
+            for i in range(len(st_images)):
+                if st_images[i] and st_strengths[i] > 0:
+                    from utils.app_utils import _ensure_model_downloaded
+                    _ensure_model_downloaded("sigclip_vision_patch14_384.safetensors", progress)
+                    temp_path = os.path.join(INPUT_DIR, f"temp_style_{i}_{random.randint(1000, 9999)}.png")
+                    st_images[i].save(temp_path, "PNG")
+                    temp_files_to_clean.append(temp_path)
+                    active_styles.append({
+                        "image": os.path.basename(temp_path), "strength": st_strengths[i]
+                    })
+        
+        reference_latent_data = ui_inputs.get('reference_latent_data', [])
+        active_reference_latents = []
+        if reference_latent_data:
+            for img in reference_latent_data:
+                if img:
+                    if not os.path.exists(INPUT_DIR): os.makedirs(INPUT_DIR)
+                    temp_path = os.path.join(INPUT_DIR, f"temp_ref_{random.randint(1000, 9999)}.png")
+                    img.save(temp_path, "PNG")
+                    temp_files_to_clean.append(temp_path)
+                    active_reference_latents.append(os.path.basename(temp_path))
+        
         from utils.app_utils import get_vae_path
         vae_source = ui_inputs.get('vae_source')
         vae_id = ui_inputs.get('vae_id')
-        vae_file = ui_inputs.get('vae_file')
         vae_name_override = None
-
         if vae_source and vae_source != "None":
             if vae_source == "File":
                 vae_name_override = sanitize_filename(vae_id)
             elif vae_source == "Civitai" and vae_id and vae_id.strip():
-                local_path, status = get_vae_path(vae_source, vae_id, ui_inputs.get('civitai_api_key'), progress)
+                local_path, status = get_vae_path(vae_source, vae_id, os.environ.get("CIVITAI_API_KEY", ""), progress)
                 if local_path: vae_name_override = os.path.basename(local_path)
                 else: raise gr.Error(f"Failed to prepare VAE {vae_id}: {status}")
-        
         if vae_name_override:
             ui_inputs['vae_name'] = vae_name_override
 
@@ -311,22 +455,12 @@ class SdImagePipeline(BasePipeline):
         active_conditioning = []
         if conditioning_data:
             num_units = len(conditioning_data) // 6
-            prompts   = conditioning_data[0*num_units : 1*num_units]
-            widths    = conditioning_data[1*num_units : 2*num_units]
-            heights   = conditioning_data[2*num_units : 3*num_units]
-            xs        = conditioning_data[3*num_units : 4*num_units]
-            ys        = conditioning_data[4*num_units : 5*num_units]
-            strengths = conditioning_data[5*num_units : 6*num_units]
-
+            prompts, widths, heights, xs, ys, strengths = [conditioning_data[i*num_units : (i+1)*num_units] for i in range(6)]
             for i in range(num_units):
                 if prompts[i] and prompts[i].strip():
                     active_conditioning.append({
-                        "prompt": prompts[i],
-                        "width": int(widths[i]),
-                        "height": int(heights[i]),
-                        "x": int(xs[i]),
-                        "y": int(ys[i]),
-                        "strength": float(strengths[i])
+                        "prompt": prompts[i], "width": int(widths[i]), "height": int(heights[i]),
+                        "x": int(xs[i]), "y": int(ys[i]), "strength": float(strengths[i])
                     })
 
         loras_string = f"LoRAs: [{', '.join(active_loras_for_meta)}]" if active_loras_for_meta else ""
@@ -335,45 +469,71 @@ class SdImagePipeline(BasePipeline):
         
         if ui_inputs.get('seed') == -1:
             ui_inputs['seed'] = random.randint(0, 2**32 - 1)
-
-        dynamic_values = {'task_type': ui_inputs['task_type']}
+        
+        model_info = ALL_MODEL_MAP[model_display_name]
+        path_or_components = model_info[1]
+        latent_type = model_info[3] if len(model_info) > 3 and model_info[3] else 'latent'
+        latent_generator_template = "EmptyLatentImage"
+        if latent_type == 'sd3_latent':
+            latent_generator_template = "EmptySD3LatentImage"
+        elif latent_type == 'chroma_radiance_latent':
+            latent_generator_template = "EmptyChromaRadianceLatentImage"
+        elif latent_type == 'hunyuan_latent':
+            latent_generator_template = "EmptyHunyuanImageLatent"
+
+        dynamic_values = {
+            'task_type': ui_inputs['task_type'], 
+            'model_type': workflow_model_type,
+            'latent_type': latent_type,
+            'latent_generator_template': latent_generator_template
+        }
         
         recipe_path = os.path.join(os.path.dirname(__file__), "workflow_recipes", "sd_unified_recipe.yaml")
         assembler = WorkflowAssembler(recipe_path, dynamic_values=dynamic_values)
         
-        model_display_name = ui_inputs['model_display_name']
-        if model_display_name not in ALL_MODEL_MAP:
-            raise gr.Error(f"Model '{model_display_name}' is not configured in model_list.yaml.")
-        
-        _, components, _, _ = ALL_MODEL_MAP[model_display_name]
-        
         workflow_inputs = {
+            **ui_inputs,
             "positive_prompt": ui_inputs['positive_prompt'], "negative_prompt": ui_inputs['negative_prompt'],
             "seed": ui_inputs['seed'], "steps": ui_inputs['num_inference_steps'], "cfg": ui_inputs['guidance_scale'],
             "sampler_name": ui_inputs['sampler'], "scheduler": ui_inputs['scheduler'],
             "batch_size": ui_inputs['batch_size'],
-            "denoise": ui_inputs['denoise'], 
-            "input_image": ui_inputs.get('input_image'),
-            "inpaint_image": ui_inputs.get('inpaint_image'),
-            "inpaint_mask": ui_inputs.get('inpaint_mask'),
-            "left": ui_inputs.get('outpaint_left'), "top": ui_inputs.get('outpaint_top'),
-            "right": ui_inputs.get('outpaint_right'), "bottom": ui_inputs.get('outpaint_bottom'),
-            "hires_upscaler": ui_inputs.get('hires_upscaler'), "hires_scale_by": ui_inputs.get('hires_scale_by'),
-            "unet_name": components['unet'],
-            "clip_name": components['clip'],
-            "vae_name": ui_inputs.get('vae_name', components['vae']),
+            "clip_skip": ui_inputs['clip_skip'],
+            "denoise": ui_inputs['denoise'],
+            "vae_name": ui_inputs.get('vae_name'),
+            "guidance": ui_inputs.get('guidance', 3.5),
             "lora_chain": active_loras_for_gpu,
-            "controlnet_chain": active_controlnets,
+            "controlnet_chain": active_controlnets if not active_anima_controlnets else [],
+            "anima_controlnet_lllite_chain": active_anima_controlnets,
             "diffsynth_controlnet_chain": active_diffsynth_controlnets,
+            "ipadapter_chain": active_ipadapters,
+            "flux1_ipadapter_chain": active_flux1_ipadapters,
+            "sd3_ipadapter_chain": active_sd3_ipadapters,
+            "style_chain": active_styles,
             "conditioning_chain": active_conditioning,
+            "reference_latent_chain": active_reference_latents,
+            "vae_chain": [ui_inputs.get('vae_name')] if ui_inputs.get('vae_name') else [],
         }
+
+        if isinstance(path_or_components, dict):
+            workflow_inputs.update({
+                'unet_name': path_or_components.get('unet'),
+                'vae_name': ui_inputs.get('vae_name') or path_or_components.get('vae'),
+                'clip_name': path_or_components.get('clip'),
+                'clip1_name': path_or_components.get('clip1'),
+                'clip2_name': path_or_components.get('clip2'),
+                'clip3_name': path_or_components.get('clip3'),
+                'clip4_name': path_or_components.get('clip4'),
+                'lora_name': path_or_components.get('lora'),
+            })
+        else:
+            workflow_inputs['model_name'] = path_or_components
         
         if task_type == 'txt2img':
             workflow_inputs['width'] = ui_inputs['width']
             workflow_inputs['height'] = ui_inputs['height']
         
         workflow = assembler.assemble(workflow_inputs)
-        
+
         progress(1.0, desc="All models ready. Requesting GPU for generation...")
 
         try:
@@ -384,7 +544,6 @@ class SdImagePipeline(BasePipeline):
                 task_name=f"ImageGen ({task_type})",
                 ui_inputs=ui_inputs,
                 loras_string=loras_string,
-                required_models_for_gpu=required_models,
                 workflow=workflow,
                 assembler=assembler, 
                 progress=progress

core/pipelines/workflow_recipes/_partials/_base_sampler_sd.yaml

@@ -0,0 +1,36 @@
+nodes:
+  pos_prompt:
+    class_type: CLIPTextEncode
+    title: "CLIP Text Encode (Positive)"
+  neg_prompt:
+    class_type: CLIPTextEncode
+    title: "CLIP Text Encode (Negative)"
+  ksampler:
+    class_type: KSampler
+    title: "KSampler"
+    params:
+      denoise: 1.0
+  vae_decode:
+    class_type: VAEDecode
+    title: "VAE Decode"
+  save_image:
+    class_type: SaveImage
+    title: "Save Image"
+    params: {}
+
+connections:
+  - from: "ksampler:0"
+    to: "vae_decode:samples"
+  - from: "vae_decode:0"
+    to: "save_image:images"
+
+ui_map:
+  positive_prompt: "pos_prompt:text"
+  negative_prompt: "neg_prompt:text"
+  seed: "ksampler:seed"
+  steps: "ksampler:steps"
+  cfg: "ksampler:cfg"
+  sampler_name: "ksampler:sampler_name"
+  scheduler: "ksampler:scheduler"
+  denoise: "ksampler:denoise"
+  filename_prefix: "save_image:filename_prefix"

core/pipelines/workflow_recipes/_partials/conditioning/anima.yaml

@@ -0,0 +1,58 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load CLIP"
+    params:
+      type: "stable_diffusion"
+      device: "default"
+
+connections:
+  - from: "unet_loader:0"
+    to: "ksampler:model"
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    output_map:
+      "unet_loader:0": "model"
+      "clip_loader:0": "clip"
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["ksampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_anima_controlnet_lllite_chains:
+  anima_controlnet_lllite_chain:
+    ksampler_node: "ksampler"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip_name: "clip_loader:clip_name"

core/pipelines/workflow_recipes/_partials/conditioning/chroma1-radiance.yaml

@@ -0,0 +1,59 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+    params:
+      vae_name: "pixel_space"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load CLIP"
+    params:
+      type: "chroma"
+      device: "default"
+  t5_tokenizer:
+    class_type: T5TokenizerOptions
+    title: "T5TokenizerOptions"
+    params:
+      min_padding: 0
+      min_length: 3
+  model_sampler:
+    class_type: ModelSamplingAuraFlow
+    params:
+      shift: 3.0
+
+connections:
+  - from: "unet_loader:0"
+    to: "model_sampler:model"
+  - from: "model_sampler:0"
+    to: "ksampler:model"
+  
+  - from: "clip_loader:0"
+    to: "t5_tokenizer:clip"
+  - from: "t5_tokenizer:0"
+    to: "pos_prompt:clip"
+  - from: "t5_tokenizer:0"
+    to: "neg_prompt:clip"
+    
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+  
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "t5_tokenizer:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  clip_name: "clip_loader:clip_name"

core/pipelines/workflow_recipes/_partials/conditioning/chroma1.yaml

@@ -0,0 +1,61 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load CLIP"
+    params:
+      type: "chroma"
+      device: "default"
+  t5_tokenizer:
+    class_type: T5TokenizerOptions
+    title: "T5TokenizerOptions"
+    params:
+      min_padding: 1
+      min_length: 0
+  fresca:
+    class_type: FreSca
+    title: "FreSca"
+    params:
+      scale_low: 1.0
+      scale_high: 2.5
+      freq_cutoff: 30
+
+connections:
+  - from: "unet_loader:0"
+    to: "fresca:model"
+  - from: "fresca:0"
+    to: "ksampler:model"
+  
+  - from: "clip_loader:0"
+    to: "t5_tokenizer:clip"
+  - from: "t5_tokenizer:0"
+    to: "pos_prompt:clip"
+  - from: "t5_tokenizer:0"
+    to: "neg_prompt:clip"
+    
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+  
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "t5_tokenizer:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip_name: "clip_loader:clip_name"

core/pipelines/workflow_recipes/_partials/conditioning/ernie-image.yaml

@@ -0,0 +1,54 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load CLIP"
+    params:
+      type: "flux2"
+      device: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+
+connections:
+  - from: "unet_loader:0"
+    to: "ksampler:model"
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    output_map:
+      "unet_loader:0": "model"
+      "clip_loader:0": "clip"
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["ksampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  clip_name: "clip_loader:clip_name"
+  vae_name: "vae_loader:vae_name"

core/pipelines/workflow_recipes/_partials/conditioning/flux1.yaml

@@ -0,0 +1,64 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load FLUX UNET"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load FLUX VAE"
+  clip_loader:
+    class_type: DualCLIPLoader
+    title: "Load FLUX Dual CLIP"
+    params:
+      type: "flux"
+      device: "default"
+  flux_guidance:
+    class_type: FluxGuidance
+    title: "FluxGuidance"
+
+connections:
+  - from: "unet_loader:0"
+    to: "ksampler:model"
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+  - from: "pos_prompt:0"
+    to: "flux_guidance:conditioning"
+  - from: "flux_guidance:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+
+dynamic_controlnet_chains:
+  controlnet_chain:
+    template: "ControlNetApplyAdvanced"
+    ksampler_node: "ksampler"
+    vae_source: "vae_loader:0"
+
+dynamic_flux1_ipadapter_chains:
+  flux1_ipadapter_chain:
+    ksampler_node: "ksampler"
+
+dynamic_style_chains:
+  style_chain:
+    flux_guidance_node: "flux_guidance"
+    ksampler_node: "ksampler"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    flux_guidance_node: "flux_guidance"
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip1_name: "clip_loader:clip_name1"
+  clip2_name: "clip_loader:clip_name2"
+  guidance: "flux_guidance:guidance"

core/pipelines/workflow_recipes/_partials/conditioning/flux2-kv.yaml

@@ -0,0 +1,104 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load CLIP"
+    params:
+      type: "flux2"
+      device: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+    
+  flux_kv_cache:
+    class_type: FluxKVCache
+    title: "Flux KV Cache"
+  
+  pos_prompt:
+    class_type: CLIPTextEncode
+    title: "CLIP Text Encode (Positive)"
+  neg_prompt:
+    class_type: CLIPTextEncode
+    title: "CLIP Text Encode (Negative)"
+  
+  ksampler:
+    class_type: KSampler
+    title: "KSampler"
+    params:
+      denoise: 1.0
+
+  vae_decode:
+    class_type: VAEDecode
+    title: "VAE Decode"
+    
+  save_image:
+    class_type: SaveImage
+    title: "Save Image"
+
+connections:
+  - from: "unet_loader:0"
+    to: "flux_kv_cache:model"
+  - from: "flux_kv_cache:0"
+    to: "ksampler:model"
+    
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+    
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+  
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+    
+  - from: "latent_source:0"
+    to: "ksampler:latent_image"
+    
+  - from: "ksampler:0"
+    to: "vae_decode:samples"
+  - from: "vae_decode:0"
+    to: "save_image:images"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    output_map:
+      "unet_loader:0": "model"
+      "clip_loader:0": "clip"
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["flux_kv_cache:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_reference_latent_chains:
+  reference_latent_chain:
+    ksampler_node: "ksampler"
+    vae_node: "vae_loader"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  clip_name: "clip_loader:clip_name"
+  vae_name: "vae_loader:vae_name"
+  
+  positive_prompt: "pos_prompt:text"
+  negative_prompt: "neg_prompt:text"
+  
+  seed: "ksampler:seed"
+  steps: "ksampler:steps"
+  cfg: "ksampler:cfg"
+  sampler_name: "ksampler:sampler_name"
+  scheduler: "ksampler:scheduler"
+  denoise: "ksampler:denoise"
+  
+  filename_prefix: "save_image:filename_prefix"

core/pipelines/workflow_recipes/_partials/conditioning/flux2.yaml

@@ -0,0 +1,96 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load CLIP"
+    params:
+      type: "flux2"
+      device: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+  
+  pos_prompt:
+    class_type: CLIPTextEncode
+    title: "CLIP Text Encode (Positive)"
+  neg_prompt:
+    class_type: CLIPTextEncode
+    title: "CLIP Text Encode (Negative)"
+  
+  ksampler:
+    class_type: KSampler
+    title: "KSampler"
+    params:
+      denoise: 1.0
+
+  vae_decode:
+    class_type: VAEDecode
+    title: "VAE Decode"
+    
+  save_image:
+    class_type: SaveImage
+    title: "Save Image"
+
+connections:
+  - from: "unet_loader:0"
+    to: "ksampler:model"
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+  
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+    
+  - from: "latent_source:0"
+    to: "ksampler:latent_image"
+    
+  - from: "ksampler:0"
+    to: "vae_decode:samples"
+  - from: "vae_decode:0"
+    to: "save_image:images"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    output_map:
+      "unet_loader:0": "model"
+      "clip_loader:0": "clip"
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["ksampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_reference_latent_chains:
+  reference_latent_chain:
+    ksampler_node: "ksampler"
+    vae_node: "vae_loader"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  clip_name: "clip_loader:clip_name"
+  vae_name: "vae_loader:vae_name"
+  
+  positive_prompt: "pos_prompt:text"
+  negative_prompt: "neg_prompt:text"
+  
+  seed: "ksampler:seed"
+  steps: "ksampler:steps"
+  cfg: "ksampler:cfg"
+  sampler_name: "ksampler:sampler_name"
+  scheduler: "ksampler:scheduler"
+  denoise: "ksampler:denoise"
+  
+  filename_prefix: "save_image:filename_prefix"

core/pipelines/workflow_recipes/_partials/conditioning/hidream.yaml

@@ -0,0 +1,53 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load HiDream UNET"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load HiDream VAE"
+  clip_loader:
+    class_type: QuadrupleCLIPLoader
+    title: "Load HiDream Quadruple CLIP"
+  
+  model_sampler:
+    class_type: ModelSamplingSD3
+    title: "ModelSamplingSD3"
+    params:
+      shift: 6.0
+
+connections:
+  - from: "unet_loader:0"
+    to: "model_sampler:model"
+    
+  - from: "model_sampler:0"
+    to: "ksampler:model"
+    
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+    
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip1_name: "clip_loader:clip_name1"
+  clip2_name: "clip_loader:clip_name2"
+  clip3_name: "clip_loader:clip_name3"
+  clip4_name: "clip_loader:clip_name4"

core/pipelines/workflow_recipes/_partials/conditioning/hunyuanimage.yaml

@@ -0,0 +1,42 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Hunyuan UNET"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load Hunyuan VAE"
+  clip_loader:
+    class_type: DualCLIPLoader
+    title: "Load Hunyuan Dual CLIP"
+    params:
+      type: "hunyuan_image"
+      device: "default"
+
+connections:
+  - from: "unet_loader:0"
+    to: "ksampler:model"
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip1_name: "clip_loader:clip_name1"
+  clip2_name: "clip_loader:clip_name2"

core/pipelines/workflow_recipes/_partials/conditioning/longcat-image.yaml

@@ -0,0 +1,83 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load CLIP"
+    params:
+      type: "longcat_image"
+      device: "default"
+  
+  cfg_norm:
+    class_type: CFGNorm
+    title: "CFGNorm"
+    params:
+      strength: 1.0
+
+  flux_guidance_pos:
+    class_type: FluxGuidance
+    title: "FluxGuidance (Positive)"
+    params:
+      guidance: 4.0
+
+  flux_guidance_neg:
+    class_type: FluxGuidance
+    title: "FluxGuidance (Negative)"
+    params:
+      guidance: 4.0
+
+connections:
+  - from: "unet_loader:0"
+    to: "cfg_norm:model"
+  - from: "cfg_norm:0"
+    to: "ksampler:model"
+  
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+    
+  - from: "pos_prompt:0"
+    to: "flux_guidance_pos:conditioning"
+  - from: "neg_prompt:0"
+    to: "flux_guidance_neg:conditioning"
+    
+  - from: "flux_guidance_pos:0"
+    to: "ksampler:positive"
+  - from: "flux_guidance_neg:0"
+    to: "ksampler:negative"
+  
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    output_map:
+      "unet_loader:0": "model"
+      "clip_loader:0": "clip"
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["cfg_norm:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    flux_guidance_node: "flux_guidance_pos"
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip_name: "clip_loader:clip_name"

core/pipelines/workflow_recipes/_partials/conditioning/lumina.yaml

@@ -0,0 +1,57 @@
+nodes:
+  ckpt_loader:
+    class_type: CheckpointLoaderSimple
+    title: "Load Checkpoint"
+  model_sampler:
+    class_type: ModelSamplingAuraFlow
+    title: "ModelSamplingAuraFlow"
+    params:
+      shift: 4.0
+
+connections:
+  - from: "ckpt_loader:0"
+    to: "model_sampler:model"
+  - from: "model_sampler:0"
+    to: "ksampler:model"
+  
+  - from: "ckpt_loader:1"
+    to: "pos_prompt:clip"
+  - from: "ckpt_loader:1"
+    to: "neg_prompt:clip"
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+    
+  - from: "ckpt_loader:2"
+    to: "vae_decode:vae"
+  - from: "ckpt_loader:2"
+    to: "vae_encode:vae"
+
+dynamic_vae_chains:
+  vae_chain:
+    targets:
+      - "vae_decode:vae"
+      - "vae_encode:vae"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    start: "ckpt_loader" 
+    output_map:
+      "0": "model" 
+      "1": "clip"  
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["model_sampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "ckpt_loader:1"
+
+ui_map:
+  model_name: "ckpt_loader:ckpt_name"

core/pipelines/workflow_recipes/_partials/conditioning/newbie-image.yaml

@@ -0,0 +1,65 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+  clip_loader:
+    class_type: DualCLIPLoader
+    title: "Load Dual CLIP"
+    params:
+      type: "newbie"
+      device: "default"
+  model_sampler:
+    class_type: ModelSamplingAuraFlow
+    title: "ModelSamplingAuraFlow"
+    params:
+      shift: 6
+
+connections:
+  - from: "unet_loader:0"
+    to: "model_sampler:model"
+  - from: "model_sampler:0"
+    to: "ksampler:model"
+  
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+    
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+  
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+
+dynamic_newbie_lora_chains:
+  lora_chain:
+    template: "NewBieLoraLoader"
+    output_map:
+      "unet_loader:0": "model"
+      "clip_loader:0": "clip"
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["model_sampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip1_name: "clip_loader:clip_name1"
+  clip2_name: "clip_loader:clip_name2"

core/pipelines/workflow_recipes/_partials/conditioning/omnigen2.yaml

@@ -0,0 +1,59 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load CLIP"
+    params:
+      type: "omnigen2"
+      device: "default"
+
+connections:
+  - from: "unet_loader:0"
+    to: "ksampler:model"
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    output_map:
+      "unet_loader:0": "model"
+      "clip_loader:0": "clip"
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["ksampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+dynamic_reference_latent_chains:
+  reference_latent_chain:
+    ksampler_node: "ksampler"
+    vae_node: "vae_loader"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip_name: "clip_loader:clip_name"

core/pipelines/workflow_recipes/_partials/conditioning/ovis-image.yaml

@@ -0,0 +1,50 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Diffusion Model"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load VAE"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load CLIP"
+    params:
+      type: "ovis"
+      device: "default"
+  model_sampler:
+    class_type: ModelSamplingAuraFlow
+    params:
+      shift: 3.0
+
+connections:
+  - from: "unet_loader:0"
+    to: "model_sampler:model"
+  - from: "model_sampler:0"
+    to: "ksampler:model"
+  
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+    
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+  
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip_name: "clip_loader:clip_name"

core/pipelines/workflow_recipes/_partials/conditioning/qwen-image.yaml

@@ -0,0 +1,80 @@
+nodes:
+  unet_loader:
+    class_type: UNETLoader
+    title: "Load Qwen UNET"
+    params:
+      weight_dtype: "default"
+  vae_loader:
+    class_type: VAELoader
+    title: "Load Qwen VAE"
+  clip_loader:
+    class_type: CLIPLoader
+    title: "Load Qwen CLIP"
+    params:
+      type: "qwen_image"
+      device: "default"
+  
+  lora_loader:
+    class_type: LoraLoaderModelOnly
+    title: "Load Qwen Lightning LoRA"
+    params:
+      strength_model: 1.0
+  model_sampler:
+    class_type: ModelSamplingAuraFlow
+    title: "ModelSamplingAuraFlow"
+    params:
+      shift: 3.1
+
+connections:
+  - from: "unet_loader:0"
+    to: "lora_loader:model"
+  - from: "lora_loader:0"
+    to: "model_sampler:model"
+
+  - from: "model_sampler:0"
+    to: "ksampler:model"
+    
+  - from: "clip_loader:0"
+    to: "pos_prompt:clip"
+  - from: "clip_loader:0"
+    to: "neg_prompt:clip"
+    
+  - from: "vae_loader:0"
+    to: "vae_decode:vae"
+  - from: "vae_loader:0"
+    to: "vae_encode:vae"
+
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    output_map:
+      "lora_loader:0": "model"
+      "clip_loader:0": "clip"
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["model_sampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_controlnet_chains:
+  controlnet_chain:
+    template: "ControlNetApplyAdvanced"
+    ksampler_node: "ksampler"
+    vae_source: "vae_loader:0"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "clip_loader:0"
+
+ui_map:
+  unet_name: "unet_loader:unet_name"
+  vae_name: "vae_loader:vae_name"
+  clip_name: "clip_loader:clip_name"
+  lora_name: "lora_loader:lora_name"

core/pipelines/workflow_recipes/_partials/conditioning/sd15.yaml

@@ -0,0 +1,69 @@
+nodes:
+  ckpt_loader:
+    class_type: CheckpointLoaderSimple
+    title: "Load Checkpoint"
+  clip_set_last_layer:
+    class_type: CLIPSetLastLayer
+    title: "CLIP Set Last Layer"
+
+connections:
+  - from: "ckpt_loader:0"
+    to: "ksampler:model"
+  - from: "ckpt_loader:1"
+    to: "clip_set_last_layer:clip"
+  - from: "clip_set_last_layer:0"
+    to: "pos_prompt:clip"
+  - from: "clip_set_last_layer:0"
+    to: "neg_prompt:clip"
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+  - from: "ckpt_loader:2"
+    to: "vae_decode:vae"
+  - from: "ckpt_loader:2"
+    to: "vae_encode:vae"
+
+dynamic_vae_chains:
+  vae_chain:
+    targets:
+      - "vae_decode:vae"
+      - "vae_encode:vae"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    start: "clip_set_last_layer"
+    output_map:
+      "ckpt_loader:0": "model" 
+      "0": "clip"             
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["ksampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_controlnet_chains:
+  controlnet_chain:
+    template: "ControlNetApplyAdvanced"
+    ksampler_node: "ksampler"
+    vae_source: "ckpt_loader:2"
+
+dynamic_ipadapter_chains:
+  ipadapter_chain:
+    end: "ksampler"
+    final_preset: "{{ ipadapter_final_preset }}"
+    final_weight: "{{ ipadapter_final_weight }}"
+    final_embeds_scaling: "{{ ipadapter_embeds_scaling }}"
+    final_loader_type: "{{ ipadapter_final_loader_type }}"
+    final_lora_strength: "{{ ipadapter_final_lora_strength }}"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "clip_set_last_layer:0"
+
+ui_map:
+  model_name: "ckpt_loader:ckpt_name"
+  clip_skip: "clip_set_last_layer:stop_at_clip_layer"

core/pipelines/workflow_recipes/_partials/conditioning/sd35.yaml

@@ -0,0 +1,58 @@
+nodes:
+  ckpt_loader:
+    class_type: CheckpointLoaderSimple
+    title: "Load Checkpoint"
+
+connections:
+  - from: "ckpt_loader:0"
+    to: "ksampler:model"
+  - from: "ckpt_loader:1"
+    to: "pos_prompt:clip"
+  - from: "ckpt_loader:1"
+    to: "neg_prompt:clip"
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+  - from: "ckpt_loader:2"
+    to: "vae_decode:vae"
+  - from: "ckpt_loader:2"
+    to: "vae_encode:vae"
+
+dynamic_vae_chains:
+  vae_chain:
+    targets:
+      - "vae_decode:vae"
+      - "vae_encode:vae"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    start: "ckpt_loader" 
+    output_map:
+      "0": "model" 
+      "1": "clip"  
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["ksampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_controlnet_chains:
+  controlnet_chain:
+    template: "ControlNetApplyAdvanced"
+    ksampler_node: "ksampler"
+    vae_source: "ckpt_loader:2"
+
+dynamic_sd3_ipadapter_chains:
+  sd3_ipadapter_chain:
+    ksampler_node: "ksampler"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "ckpt_loader:1"
+
+ui_map:
+  model_name: "ckpt_loader:ckpt_name"

core/pipelines/workflow_recipes/_partials/conditioning/sdxl.yaml

@@ -0,0 +1,63 @@
+nodes:
+  ckpt_loader:
+    class_type: CheckpointLoaderSimple
+    title: "Load Checkpoint"
+
+connections:
+  - from: "ckpt_loader:0"
+    to: "ksampler:model"
+  - from: "ckpt_loader:1"
+    to: "pos_prompt:clip"
+  - from: "ckpt_loader:1"
+    to: "neg_prompt:clip"
+  - from: "pos_prompt:0"
+    to: "ksampler:positive"
+  - from: "neg_prompt:0"
+    to: "ksampler:negative"
+  - from: "ckpt_loader:2"
+    to: "vae_decode:vae"
+  - from: "ckpt_loader:2"
+    to: "vae_encode:vae"
+
+dynamic_vae_chains:
+  vae_chain:
+    targets:
+      - "vae_decode:vae"
+      - "vae_encode:vae"
+
+dynamic_lora_chains:
+  lora_chain:
+    template: "LoraLoader"
+    start: "ckpt_loader" 
+    output_map:
+      "0": "model" 
+      "1": "clip"  
+    input_map:
+      "model": "model"
+      "clip": "clip"
+    end_input_map:
+      "model": ["ksampler:model"]
+      "clip": ["pos_prompt:clip", "neg_prompt:clip"]
+
+dynamic_controlnet_chains:
+  controlnet_chain:
+    template: "ControlNetApplyAdvanced"
+    ksampler_node: "ksampler"
+    vae_source: "ckpt_loader:2"
+
+dynamic_ipadapter_chains:
+  ipadapter_chain:
+    end: "ksampler"
+    final_preset: "{{ ipadapter_final_preset }}"
+    final_weight: "{{ ipadapter_final_weight }}"
+    final_embeds_scaling: "{{ ipadapter_embeds_scaling }}"
+    final_loader_type: "{{ ipadapter_final_loader_type }}"
+    final_lora_strength: "{{ ipadapter_final_lora_strength }}"
+
+dynamic_conditioning_chains:
+  conditioning_chain:
+    ksampler_node: "ksampler"
+    clip_source: "ckpt_loader:1"
+
+ui_map:
+  model_name: "ckpt_loader:ckpt_name"

core/pipelines/workflow_recipes/_partials/conditioning/z-image.yaml

@@ -17,12 +17,6 @@ nodes:
     class_type: ModelSamplingAuraFlow
     params:
       shift: 3.0
-  pos_prompt:
-    class_type: CLIPTextEncode
-    title: "Positive Prompt Encoder"
-  neg_prompt:
-    class_type: CLIPTextEncode
-    title: "Negative Prompt Encoder"
 
 connections:
   - from: "unet_loader:0"
@@ -65,14 +59,7 @@ dynamic_diffsynth_controlnet_chains:
     ksampler_node: "ksampler"
     vae_source: "vae_loader:0"
 
-dynamic_conditioning_chains:
-  conditioning_chain:
-    ksampler_node: "ksampler"
-    clip_source: "clip_loader:0"
-
 ui_map:
   unet_name: "unet_loader:unet_name"
   vae_name: "vae_loader:vae_name"
-  clip_name: "clip_loader:clip_name"
-  positive_prompt: "pos_prompt:text"
-  negative_prompt: "neg_prompt:text"
+  clip_name: "clip_loader:clip_name"

core/pipelines/workflow_recipes/_partials/input/hires_fix.yaml

@@ -1,15 +1,16 @@
 nodes:
   input_image_loader:
     class_type: LoadImage
-    
+    title: "Load Input Image"
   vae_encode:
     class_type: VAEEncode
-    
+    title: "VAE Encode (Hires Pre-step)"
   latent_upscaler:
     class_type: LatentUpscaleBy
-
+    title: "Upscale Latent By"
   latent_source:
     class_type: RepeatLatentBatch
+    title: "Repeat Latent Batch for Hires"
 
 connections:
   - from: "input_image_loader:0"

core/pipelines/workflow_recipes/_partials/input/img2img.yaml

@@ -1,12 +1,13 @@
 nodes:
   input_image_loader:
     class_type: LoadImage
-  
+    title: "Load Input Image"
   vae_encode:
     class_type: VAEEncode
-
+    title: "VAE Encode (Img2Img)"
   latent_source:
     class_type: RepeatLatentBatch
+    title: "Repeat Latent Batch"
 
 connections:
   - from: "input_image_loader:0"

core/pipelines/workflow_recipes/_partials/input/inpaint.yaml

@@ -2,24 +2,22 @@ nodes:
   inpaint_loader:
     class_type: LoadImage
     title: "Load Inpaint Image+Mask"
-      
   vae_encode:
     class_type: VAEEncodeForInpaint
-    params:
-      grow_mask_by: 6
-      
+    title: "VAE Encode (for Inpainting)"
   latent_source:
     class_type: RepeatLatentBatch
-      
+    title: "Repeat Latent Batch"
+
 connections:
   - from: "inpaint_loader:0"
     to: "vae_encode:pixels"
   - from: "inpaint_loader:1"
     to: "vae_encode:mask"
-    
   - from: "vae_encode:0"
     to: "latent_source:samples"
 
 ui_map:
-  inpaint_image: "inpaint_loader:image"
-  batch_size: "latent_source:amount"
+  input_image: "inpaint_loader:image"
+  batch_size: "latent_source:amount"
+  grow_mask_by: "vae_encode:grow_mask_by"

core/pipelines/workflow_recipes/_partials/input/outpaint.yaml

@@ -1,38 +1,41 @@
 nodes:
   input_image_loader:
     class_type: LoadImage
-    
+    title: "Load Image for Outpaint"
+  scale_image:
+    class_type: ImageScaleToTotalPixels
+    title: "Scale Image to Total Pixels"
+    params:
+      upscale_method: "nearest-exact"
   pad_image:
     class_type: ImagePadForOutpaint
-    params:
-      feathering: 10
-      
+    title: "Pad Image for Outpainting"
   vae_encode:
     class_type: VAEEncodeForInpaint
-    params:
-      grow_mask_by: 6
-
+    title: "VAE Encode (for Inpainting)"
   latent_source:
     class_type: RepeatLatentBatch
+    title: "Repeat Latent Batch"
 
 connections:
   - from: "input_image_loader:0"
+    to: "scale_image:image"
+  - from: "scale_image:0"
     to: "pad_image:image"
-    
   - from: "pad_image:0"
     to: "vae_encode:pixels"
   - from: "pad_image:1"
     to: "vae_encode:mask"
-
   - from: "vae_encode:0"
     to: "latent_source:samples"
 
 ui_map:
   input_image: "input_image_loader:image"
-  
+  megapixels: "scale_image:megapixels"
   left: "pad_image:left"
   top: "pad_image:top"
   right: "pad_image:right"
   bottom: "pad_image:bottom"
-  
+  feathering: "pad_image:feathering"
+  grow_mask_by: "vae_encode:grow_mask_by"
   batch_size: "latent_source:amount"

core/pipelines/workflow_recipes/_partials/input/txt2img.yaml

@@ -1,8 +1,2 @@
-nodes:
-  latent_source:
-    class_type: EmptySD3LatentImage
-
-ui_map:
-  width: "latent_source:width"
-  height: "latent_source:height"
-  batch_size: "latent_source:batch_size"
+imports:
+  - "txt2img_{{ latent_type }}.yaml"

core/pipelines/workflow_recipes/_partials/input/txt2img_chroma_radiance_latent.yaml

@@ -0,0 +1,11 @@
+nodes:
+  latent_source:
+    class_type: "EmptyChromaRadianceLatentImage"
+    title: "EmptyChromaRadianceLatentImage"
+
+connections: []
+
+ui_map:
+  width: "latent_source:width"
+  height: "latent_source:height"
+  batch_size: "latent_source:batch_size"

core/pipelines/workflow_recipes/_partials/input/txt2img_flux2_latent.yaml

@@ -0,0 +1,11 @@
+nodes:
+  latent_source:
+    class_type: "EmptyFlux2LatentImage"
+    title: "Empty Flux 2 Latent"
+
+connections: []
+
+ui_map:
+  width: "latent_source:width"
+  height: "latent_source:height"
+  batch_size: "latent_source:batch_size"

core/pipelines/workflow_recipes/_partials/input/txt2img_hunyuan_latent.yaml

@@ -0,0 +1,11 @@
+nodes:
+  latent_source:
+    class_type: "EmptyHunyuanImageLatent"
+    title: "EmptyHunyuanImageLatent"
+
+connections: []
+
+ui_map:
+  width: "latent_source:width"
+  height: "latent_source:height"
+  batch_size: "latent_source:batch_size"

core/pipelines/workflow_recipes/_partials/input/txt2img_latent.yaml

@@ -0,0 +1,11 @@
+nodes:
+  latent_source:
+    class_type: "{{ latent_generator_template }}"
+    title: "Empty Latent Image"
+
+connections: []
+
+ui_map:
+  width: "latent_source:width"
+  height: "latent_source:height"
+  batch_size: "latent_source:batch_size"

core/pipelines/workflow_recipes/_partials/input/txt2img_sd3_latent.yaml

@@ -0,0 +1,11 @@
+nodes:
+  latent_source:
+    class_type: "EmptySD3LatentImage"
+    title: "EmptySD3LatentImage"
+
+connections: []
+
+ui_map:
+  width: "latent_source:width"
+  height: "latent_source:height"
+  batch_size: "latent_source:batch_size"

core/pipelines/workflow_recipes/sd_unified_recipe.yaml

@@ -1,7 +1,7 @@
 imports:
-  - "_partials/_base_sampler.yaml"
+  - "_partials/_base_sampler_sd.yaml"
   - "_partials/input/{{ task_type }}.yaml"
-  - "_partials/conditioning/z-image.yaml"
+  - "_partials/conditioning/{{ model_type }}.yaml"
 
 connections:
   - from: "latent_source:0"

core/settings.py

@@ -10,16 +10,37 @@ MODEL_PATCHES_DIR = "models/model_patches"
 DIFFUSION_MODELS_DIR = "models/diffusion_models"
 VAE_DIR = "models/vae"
 TEXT_ENCODERS_DIR = "models/text_encoders"
+STYLE_MODELS_DIR = "models/style_models"
+CLIP_VISION_DIR = "models/clip_vision"
+IPADAPTER_DIR = "models/ipadapter"
+IPADAPTER_FLUX_DIR = "models/ipadapter-flux"
 INPUT_DIR = "input"
 OUTPUT_DIR = "output"
 
+CATEGORY_TO_DIR_MAP = {
+    "diffusion_models": DIFFUSION_MODELS_DIR,
+    "text_encoders": TEXT_ENCODERS_DIR,
+    "vae": VAE_DIR,
+    "checkpoints": CHECKPOINT_DIR,
+    "loras": LORA_DIR,
+    "controlnet": CONTROLNET_DIR,
+    "model_patches": MODEL_PATCHES_DIR,
+    "embeddings": EMBEDDING_DIR,
+    "style_models": STYLE_MODELS_DIR,
+    "clip_vision": CLIP_VISION_DIR,
+    "ipadapter": IPADAPTER_DIR,
+    "ipadapter-flux": IPADAPTER_FLUX_DIR
+}
+
 _PROJECT_ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 _MODEL_LIST_PATH = os.path.join(_PROJECT_ROOT, 'yaml', 'model_list.yaml')
 _FILE_LIST_PATH = os.path.join(_PROJECT_ROOT, 'yaml', 'file_list.yaml')
+_IPADAPTER_LIST_PATH = os.path.join(_PROJECT_ROOT, 'yaml', 'ipadapter.yaml')
 _CONSTANTS_PATH = os.path.join(_PROJECT_ROOT, 'yaml', 'constants.yaml')
+_MODEL_ARCHITECTURES_PATH = os.path.join(_PROJECT_ROOT, 'yaml', 'model_architectures.yaml')
+_IMAGE_GEN_FEATURES_PATH = os.path.join(_PROJECT_ROOT, 'yaml', 'image_gen_features.yaml')
 _MODEL_DEFAULTS_PATH = os.path.join(_PROJECT_ROOT, 'yaml', 'model_defaults.yaml')
 
-
 def load_constants_from_yaml(filepath=_CONSTANTS_PATH):
     if not os.path.exists(filepath):
         print(f"Warning: Constants file not found at {filepath}. Using fallback values.")
@@ -27,6 +48,27 @@ def load_constants_from_yaml(filepath=_CONSTANTS_PATH):
     with open(filepath, 'r', encoding='utf-8') as f:
         return yaml.safe_load(f)
 
+def load_architectures_config(filepath=_MODEL_ARCHITECTURES_PATH):
+    if not os.path.exists(filepath):
+        print(f"Warning: Architectures file not found at {filepath}.")
+        return {}
+    with open(filepath, 'r', encoding='utf-8') as f:
+        return yaml.safe_load(f)
+
+def load_features_config(filepath=_IMAGE_GEN_FEATURES_PATH):
+    if not os.path.exists(filepath):
+        print(f"Warning: Features file not found at {filepath}.")
+        return {}
+    with open(filepath, 'r', encoding='utf-8') as f:
+        return yaml.safe_load(f)
+
+def load_model_defaults(filepath=_MODEL_DEFAULTS_PATH):
+    if not os.path.exists(filepath):
+        print(f"Warning: Model defaults file not found at {filepath}.")
+        return {}
+    with open(filepath, 'r', encoding='utf-8') as f:
+        return yaml.safe_load(f)
+
 def load_file_download_map(filepath=_FILE_LIST_PATH):
     if not os.path.exists(filepath):
         raise FileNotFoundError(f"The file list (for downloads) was not found at: {filepath}")
@@ -59,50 +101,86 @@ def load_models_from_yaml(model_list_filepath=_MODEL_LIST_PATH, download_map=Non
     }
     category_map_names = {
         "Checkpoint": "MODEL_MAP_CHECKPOINT",
+        "Checkpoints": "MODEL_MAP_CHECKPOINT"
     }
     
-    for category, models in model_data.items():
+    for category, architectures in model_data.items():
         if category in category_map_names:
             map_name = category_map_names[category]
-            if not isinstance(models, list): continue
-            for model in models:
-                display_name = model['display_name']
-                components = model.get('components', {})
-
-                model_tuple = (
-                    None,
-                    components,
-                    "SDXL",
-                    None
-                )
-                model_maps[map_name][display_name] = model_tuple
-                model_maps["ALL_MODEL_MAP"][display_name] = model_tuple
+            if not isinstance(architectures, dict): continue
+            
+            for arch, arch_data in architectures.items():
+                if not isinstance(arch_data, dict): continue
+                
+                latent_type = arch_data.get('latent_type', 'latent')
+                models = arch_data.get('models', [])
+                if not isinstance(models, list): continue
+                
+                for model in models:
+                    display_name = model['display_name']
+                    path_or_components = model.get('path') or model.get('components')
+                    mod_category = model.get('category', None)
+                    
+                    repo_id = ''
+                    if isinstance(path_or_components, str):
+                        download_info = download_map.get(path_or_components, {})
+                        repo_id = download_info.get('repo_id', '')
+                    
+                    model_tuple = (
+                        repo_id,
+                        path_or_components,
+                        arch,
+                        latent_type,
+                        mod_category
+                    )
+                    model_maps[map_name][display_name] = model_tuple
+                    model_maps["ALL_MODEL_MAP"][display_name] = model_tuple
     
     return model_maps
 
-def load_model_defaults(filepath=_MODEL_DEFAULTS_PATH):
-    if not os.path.exists(filepath):
-        print(f"Warning: Model defaults file not found at {filepath}. Using empty defaults.")
-        return {}
-    with open(filepath, 'r', encoding='utf-8') as f:
-        return yaml.safe_load(f)
-
 try:
     ALL_FILE_DOWNLOAD_MAP = load_file_download_map()
     loaded_maps = load_models_from_yaml(download_map=ALL_FILE_DOWNLOAD_MAP)
     MODEL_MAP_CHECKPOINT = loaded_maps["MODEL_MAP_CHECKPOINT"]
     ALL_MODEL_MAP = loaded_maps["ALL_MODEL_MAP"]
     
+    category_to_model_type = {
+        "diffusion_models": "UNET",
+        "text_encoders": "TEXT_ENCODER",
+        "vae": "VAE",
+        "checkpoints": "SDXL",
+        "loras": "LORA",
+        "controlnet": "CONTROLNET",
+        "model_patches": "MODEL_PATCH",
+        "style_models": "STYLE",
+        "clip_vision": "CLIP_VISION",
+        "ipadapter": "IPADAPTER",
+        "ipadapter-flux": "IPADAPTER_FLUX"
+    }
+    for filename, file_info in ALL_FILE_DOWNLOAD_MAP.items():
+        if filename not in ALL_MODEL_MAP:
+            category = file_info.get('category')
+            model_type = category_to_model_type.get(category, 'UNKNOWN')
+            repo_id = file_info.get('repo_id', '')
+            ALL_MODEL_MAP[filename] = (repo_id, filename, model_type, None, None)
+
     MODEL_TYPE_MAP = {k: v[2] for k, v in ALL_MODEL_MAP.items()}
-    
-    ALL_MODEL_DEFAULTS = load_model_defaults()
+
+    ARCH_CATEGORIES_MAP = {}
+    for display_name, info in MODEL_MAP_CHECKPOINT.items():
+        arch = info[2]
+        cat = info[4] if len(info) > 4 else None
+        if arch not in ARCH_CATEGORIES_MAP:
+            ARCH_CATEGORIES_MAP[arch] = []
+        if cat and cat not in ARCH_CATEGORIES_MAP[arch]:
+            ARCH_CATEGORIES_MAP[arch].append(cat)
 
 except Exception as e:
     print(f"FATAL: Could not load model configuration from YAML. Error: {e}")
     ALL_FILE_DOWNLOAD_MAP = {}
     MODEL_MAP_CHECKPOINT, ALL_MODEL_MAP = {}, {}
     MODEL_TYPE_MAP = {}
-    ALL_MODEL_DEFAULTS = {}
+    ARCH_CATEGORIES_MAP = {}
 
 
 try:
@@ -111,13 +189,17 @@ try:
     MAX_EMBEDDINGS = _constants.get('MAX_EMBEDDINGS', 5)
     MAX_CONDITIONINGS = _constants.get('MAX_CONDITIONINGS', 10)
     MAX_CONTROLNETS = _constants.get('MAX_CONTROLNETS', 5)
-    LORA_SOURCE_CHOICES = _constants.get('LORA_SOURCE_CHOICES', ["Civitai", "Custom URL", "File"])
+    MAX_IPADAPTERS = _constants.get('MAX_IPADAPTERS', 5)
+    LORA_SOURCE_CHOICES = _constants.get('LORA_SOURCE_CHOICES', ["Civitai", "File"])
     RESOLUTION_MAP = _constants.get('RESOLUTION_MAP', {})
+    ARCHITECTURES_CONFIG = load_architectures_config()
+    FEATURES_CONFIG = load_features_config()
+    MODEL_DEFAULTS_CONFIG = load_model_defaults()
 except Exception as e:
     print(f"FATAL: Could not load constants from YAML. Error: {e}")
-    MAX_LORAS, MAX_EMBEDDINGS, MAX_CONDITIONINGS, MAX_CONTROLNETS = 5, 5, 10, 5
-    LORA_SOURCE_CHOICES = ["Civitai", "Custom URL", "File"]
+    MAX_LORAS, MAX_EMBEDDINGS, MAX_CONDITIONINGS, MAX_CONTROLNETS, MAX_IPADAPTERS = 5, 5, 10, 5, 5
+    LORA_SOURCE_CHOICES = ["Civitai", "File"]
     RESOLUTION_MAP = {}
-
-
-DEFAULT_NEGATIVE_PROMPT = ""
+    ARCHITECTURES_CONFIG = {}
+    FEATURES_CONFIG = {}
+    MODEL_DEFAULTS_CONFIG = {}

requirements.txt

@@ -1,10 +1,10 @@
-comfyui-frontend-package==1.42.14
-comfyui-workflow-templates==0.9.59
-comfyui-embedded-docs==0.4.3
-torch==2.10.0
+comfyui-frontend-package==1.43.18
+comfyui-workflow-templates==0.9.77
+comfyui-embedded-docs==0.5.0
+torch
 torchsde
-torchvision==0.25.0
-torchaudio==2.10.0
+torchvision
+torchaudio
 numpy>=1.25.0
 einops
 transformers>=4.50.3
@@ -19,11 +19,11 @@ scipy
 tqdm
 psutil
 alembic
-SQLAlchemy>=2.0
+SQLAlchemy>=2.0.0
 filelock
 av>=14.2.0
 comfy-kitchen>=0.2.8
-comfy-aimdo>=0.2.12
+comfy-aimdo==0.3.0
 requests
 simpleeval>=1.0.0
 blake3
@@ -58,4 +58,5 @@ svglib
 trimesh[easy]
 yacs
 yapf
-onnxruntime-gpu
+onnxruntime-gpu
+diffusers