app

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

@@ -1,6 +1,6 @@
 ---
 title: YOLOR
-emoji: 📉
+emoji: 🚀
 colorFrom: gray
 colorTo: purple
 sdk: gradio
@@ -35,3 +35,226 @@ Path is relative to the root of the repository.
 
 `pinned`: _boolean_  
 Whether the Space stays on top of your list.
+
+
+# YOLOR
+implementation of paper - [You Only Learn One Representation: Unified Network for Multiple Tasks](https://arxiv.org/abs/2105.04206)
+
+[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/you-only-learn-one-representation-unified/real-time-object-detection-on-coco)](https://paperswithcode.com/sota/real-time-object-detection-on-coco?p=you-only-learn-one-representation-unified)
+
+![Unified Network](https://github.com/WongKinYiu/yolor/blob/main/figure/unifued_network.png)
+
+<img src="https://github.com/WongKinYiu/yolor/blob/main/figure/performance.png" height="480">
+
+To get the results on the table, please use [this branch](https://github.com/WongKinYiu/yolor/tree/paper).
+
+| Model | Test Size | AP<sup>test</sup> | AP<sub>50</sub><sup>test</sup> | AP<sub>75</sub><sup>test</sup> | batch1 throughput | batch32 inference |
+| :-- | :-: | :-: | :-: | :-: | :-: | :-: |
+| **YOLOR-P6** | 1280 | **54.1%** | **71.8%** | **59.3%** | 49 *fps* | 8.3 *ms* |
+| **YOLOR-W6** | 1280 | **55.5%** | **73.2%** | **60.6%** | 47 *fps* | 10.7 *ms* |
+| **YOLOR-E6** | 1280 | **56.4%** | **74.1%** | **61.6%** | 37 *fps* | 17.1 *ms* |
+| **YOLOR-D6** | 1280 | **57.3%** | **75.0%** | **62.7%** | 30 *fps* | 21.8 *ms* |
+| **YOLOR-D6*** | 1280 | **57.8%** | **75.5%** | **63.3%** | 30 *fps* | 21.8 *ms* |
+|  |  |  |  |  |  |  |
+| **YOLOv4-P5** | 896 | **51.8%** | **70.3%** | **56.6%** | 41 *fps* | - |
+| **YOLOv4-P6** | 1280 | **54.5%** | **72.6%** | **59.8%** | 30 *fps* | - |
+| **YOLOv4-P7** | 1536 | **55.5%** | **73.4%** | **60.8%** | 16 *fps* | - |
+|  |  |  |  |  |  |  |
+
+To reproduce the inference speed, please see [darknet](https://github.com/WongKinYiu/yolor/tree/main/darknet).
+
+| Model | Test Size | AP<sup>val</sup> | AP<sub>50</sub><sup>val</sup> | AP<sub>75</sub><sup>val</sup> | AP<sub>S</sub><sup>val</sup> | AP<sub>M</sub><sup>val</sup> | AP<sub>L</sub><sup>val</sup> | batch1 throughput |
+| :-- | :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: | 
+| [**YOLOv4-CSP**](/cfg/yolov4_csp.cfg) | 640 | **49.1%** | **67.7%** | **53.8%** | **32.1%** | **54.4%** | **63.2%** | 76 *fps* |
+| [**YOLOR-CSP**](/cfg/yolor_csp.cfg) | 640 | **49.2%** | **67.6%** | **53.7%** | **32.9%** | **54.4%** | **63.0%** | [weights](https://drive.google.com/file/d/1ZEqGy4kmZyD-Cj3tEFJcLSZenZBDGiyg/view?usp=sharing) |
+| [**YOLOR-CSP***](/cfg/yolor_csp.cfg) | 640 | **50.0%** | **68.7%** | **54.3%** | **34.2%** | **55.1%** | **64.3%** | [weights](https://drive.google.com/file/d/1OJKgIasELZYxkIjFoiqyn555bcmixUP2/view?usp=sharing) |
+|  |  |  |  |  |  |  |
+| [**YOLOv4-CSP-X**](/cfg/yolov4_csp_x.cfg) | 640 | **50.9%** | **69.3%** | **55.4%** | **35.3%** | **55.8%** | **64.8%** | 53 *fps* |
+| [**YOLOR-CSP-X**](/cfg/yolor_csp_x.cfg) | 640 | **51.1%** | **69.6%** | **55.7%** | **35.7%** | **56.0%** | **65.2%** | [weights](https://drive.google.com/file/d/1L29rfIPNH1n910qQClGftknWpTBgAv6c/view?usp=sharing) |
+| [**YOLOR-CSP-X***](/cfg/yolor_csp_x.cfg) | 640 | **51.5%** | **69.9%** | **56.1%** | **35.8%** | **56.8%** | **66.1%** | [weights](https://drive.google.com/file/d/1NbMG3ivuBQ4S8kEhFJ0FIqOQXevGje_w/view?usp=sharing) |
+|  |  |  |  |  |  |  |
+
+Developing...
+
+| Model | Test Size | AP<sup>test</sup> | AP<sub>50</sub><sup>test</sup> | AP<sub>75</sub><sup>test</sup> | AP<sub>S</sub><sup>test</sup> | AP<sub>M</sub><sup>test</sup> | AP<sub>L</sub><sup>test</sup> |
+| :-- | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
+| **YOLOR-CSP** | 640 | **51.1%** | **69.6%** | **55.7%** | **31.7%** | **55.3%** | **64.7%** |
+| **YOLOR-CSP-X** | 640 | **53.0%** | **71.4%** | **57.9%** | **33.7%** | **57.1%** | **66.8%** |
+
+Train from scratch for 300 epochs...
+
+| Model | Info | Test Size | AP |
+| :-- | :-- | :-: | :-: |
+| **YOLOR-CSP** | [evolution](https://github.com/ultralytics/yolov3/issues/392) | 640 | **48.0%** |
+| **YOLOR-CSP** | [strategy](https://openaccess.thecvf.com/content/ICCV2021W/LPCV/html/Wang_Exploring_the_Power_of_Lightweight_YOLOv4_ICCVW_2021_paper.html) | 640 | **50.0%** |
+| **YOLOR-CSP** | [strategy](https://openaccess.thecvf.com/content/ICCV2021W/LPCV/html/Wang_Exploring_the_Power_of_Lightweight_YOLOv4_ICCVW_2021_paper.html) + [simOTA](https://arxiv.org/abs/2107.08430) | 640 | **51.1%** |
+|  |  |  |  |
+| **YOLOR-CSP-X** | [strategy](https://openaccess.thecvf.com/content/ICCV2021W/LPCV/html/Wang_Exploring_the_Power_of_Lightweight_YOLOv4_ICCVW_2021_paper.html) | 640 | **51.5%** |
+| **YOLOR-CSP-X** | [strategy](https://openaccess.thecvf.com/content/ICCV2021W/LPCV/html/Wang_Exploring_the_Power_of_Lightweight_YOLOv4_ICCVW_2021_paper.html) + [simOTA](https://arxiv.org/abs/2107.08430) | 640 | **53.0%** |
+
+## Installation
+
+Docker environment (recommended)
+<details><summary> <b>Expand</b> </summary>
+
+```
+# create the docker container, you can change the share memory size if you have more.
+nvidia-docker run --name yolor -it -v your_coco_path/:/coco/ -v your_code_path/:/yolor --shm-size=64g nvcr.io/nvidia/pytorch:20.11-py3
+
+# apt install required packages
+apt update
+apt install -y zip htop screen libgl1-mesa-glx
+
+# pip install required packages
+pip install seaborn thop
+
+# install mish-cuda if you want to use mish activation
+# https://github.com/thomasbrandon/mish-cuda
+# https://github.com/JunnYu/mish-cuda
+cd /
+git clone https://github.com/JunnYu/mish-cuda
+cd mish-cuda
+python setup.py build install
+
+# install pytorch_wavelets if you want to use dwt down-sampling module
+# https://github.com/fbcotter/pytorch_wavelets
+cd /
+git clone https://github.com/fbcotter/pytorch_wavelets
+cd pytorch_wavelets
+pip install .
+
+# go to code folder
+cd /yolor
+```
+
+</details>
+
+Colab environment
+<details><summary> <b>Expand</b> </summary>
+  
+```
+git clone https://github.com/WongKinYiu/yolor
+cd yolor
+
+# pip install required packages
+pip install -qr requirements.txt
+
+# install mish-cuda if you want to use mish activation
+# https://github.com/thomasbrandon/mish-cuda
+# https://github.com/JunnYu/mish-cuda
+git clone https://github.com/JunnYu/mish-cuda
+cd mish-cuda
+python setup.py build install
+cd ..
+
+# install pytorch_wavelets if you want to use dwt down-sampling module
+# https://github.com/fbcotter/pytorch_wavelets
+git clone https://github.com/fbcotter/pytorch_wavelets
+cd pytorch_wavelets
+pip install .
+cd ..
+```
+
+</details>
+
+Prepare COCO dataset
+<details><summary> <b>Expand</b> </summary>
+
+```
+cd /yolor
+bash scripts/get_coco.sh
+```
+
+</details>
+
+Prepare pretrained weight
+<details><summary> <b>Expand</b> </summary>
+
+```
+cd /yolor
+bash scripts/get_pretrain.sh
+```
+
+</details>
+
+## Testing
+
+[`yolor_p6.pt`](https://drive.google.com/file/d/1Tdn3yqpZ79X7R1Ql0zNlNScB1Dv9Fp76/view?usp=sharing)
+
+```
+python test.py --data data/coco.yaml --img 1280 --batch 32 --conf 0.001 --iou 0.65 --device 0 --cfg cfg/yolor_p6.cfg --weights yolor_p6.pt --name yolor_p6_val
+```
+
+You will get the results:
+
+```
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.52510
+ Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.70718
+ Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.57520
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.37058
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.56878
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.66102
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.39181
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.65229
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.71441
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.57755
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.75337
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.84013
+```
+
+## Training
+
+Single GPU training:
+
+```
+python train.py --batch-size 8 --img 1280 1280 --data coco.yaml --cfg cfg/yolor_p6.cfg --weights '' --device 0 --name yolor_p6 --hyp hyp.scratch.1280.yaml --epochs 300
+```
+
+Multiple GPU training:
+
+```
+python -m torch.distributed.launch --nproc_per_node 2 --master_port 9527 train.py --batch-size 16 --img 1280 1280 --data coco.yaml --cfg cfg/yolor_p6.cfg --weights '' --device 0,1 --sync-bn --name yolor_p6 --hyp hyp.scratch.1280.yaml --epochs 300
+```
+
+Training schedule in the paper:
+
+```
+python -m torch.distributed.launch --nproc_per_node 8 --master_port 9527 train.py --batch-size 64 --img 1280 1280 --data data/coco.yaml --cfg cfg/yolor_p6.cfg --weights '' --device 0,1,2,3,4,5,6,7 --sync-bn --name yolor_p6 --hyp hyp.scratch.1280.yaml --epochs 300
+python -m torch.distributed.launch --nproc_per_node 8 --master_port 9527 tune.py --batch-size 64 --img 1280 1280 --data data/coco.yaml --cfg cfg/yolor_p6.cfg --weights 'runs/train/yolor_p6/weights/last_298.pt' --device 0,1,2,3,4,5,6,7 --sync-bn --name yolor_p6-tune --hyp hyp.finetune.1280.yaml --epochs 450
+python -m torch.distributed.launch --nproc_per_node 8 --master_port 9527 train.py --batch-size 64 --img 1280 1280 --data data/coco.yaml --cfg cfg/yolor_p6.cfg --weights 'runs/train/yolor_p6-tune/weights/epoch_424.pt' --device 0,1,2,3,4,5,6,7 --sync-bn --name yolor_p6-fine --hyp hyp.finetune.1280.yaml --epochs 450
+```
+
+## Inference
+
+[`yolor_p6.pt`](https://drive.google.com/file/d/1Tdn3yqpZ79X7R1Ql0zNlNScB1Dv9Fp76/view?usp=sharing)
+
+```
+python detect.py --source inference/images/horses.jpg --cfg cfg/yolor_p6.cfg --weights yolor_p6.pt --conf 0.25 --img-size 1280 --device 0
+```
+
+You will get the results:
+
+![horses](https://github.com/WongKinYiu/yolor/blob/main/inference/output/horses.jpg)
+
+## Citation
+
+```
+@article{wang2021you,
+  title={You Only Learn One Representation: Unified Network for Multiple Tasks},
+  author={Wang, Chien-Yao and Yeh, I-Hau and Liao, Hong-Yuan Mark},
+  journal={arXiv preprint arXiv:2105.04206},
+  year={2021}
+}
+```
+
+## Acknowledgements
+
+<details><summary> <b>Expand</b> </summary>
+
+* [https://github.com/AlexeyAB/darknet](https://github.com/AlexeyAB/darknet)
+* [https://github.com/WongKinYiu/PyTorch_YOLOv4](https://github.com/WongKinYiu/PyTorch_YOLOv4)
+* [https://github.com/WongKinYiu/ScaledYOLOv4](https://github.com/WongKinYiu/ScaledYOLOv4)
+* [https://github.com/ultralytics/yolov3](https://github.com/ultralytics/yolov3)
+* [https://github.com/ultralytics/yolov5](https://github.com/ultralytics/yolov5)
+
+</details>

app.py

@@ -0,0 +1,104 @@
+from PIL import Image
+import cv2
+import torch
+from numpy import random
+
+from utils.general import (non_max_suppression, scale_coords)
+from utils.plots import plot_one_box
+
+from models.models import *
+from utils.datasets import *
+from utils.general import *
+
+import gradio as gr
+import requests
+
+import gdown
+
+
+url = 'https://drive.google.com/u/0/uc?id=1Tdn3yqpZ79X7R1Ql0zNlNScB1Dv9Fp76&export=download'
+output = 'yolor_p6.pt'
+gdown.download(url, output, quiet=False)
+
+url1 = 'https://cdn.pixabay.com/photo/2014/09/07/21/52/city-438393_1280.jpg'
+r = requests.get(url1, allow_redirects=True)
+open("city1.jpg", 'wb').write(r.content)
+url2 = 'https://cdn.pixabay.com/photo/2016/02/19/11/36/canal-1209808_1280.jpg'
+r = requests.get(url2, allow_redirects=True)
+open("city2.jpg", 'wb').write(r.content)
+
+
+conf_thres = 0.4
+iou_thres = 0.5
+
+
+def load_classes(path):
+    # Loads *.names file at 'path'
+    with open(path, 'r') as f:
+        names = f.read().split('\n')
+    return list(filter(None, names))  # filter removes empty strings (such as last line)
+
+def detect(pil_img,names):
+    img_np = np.array(pil_img)
+    img = torch.from_numpy(img_np)
+    img = img.float()
+    img /= 255.0  # 0 - 255 to 0.0 - 1.0
+
+    # Inference
+    pred = model(img.unsqueeze(0).permute(0,3,1,2), augment=False)[0]
+
+    # Apply NMS
+    pred = non_max_suppression(pred, conf_thres, iou_thres, classes=None, agnostic=False)
+
+    # Process detections
+    for i, det in enumerate(pred):  # detections per image
+        if det is not None and len(det):
+            # Rescale boxes from img_size to im0 size
+            det[:, :4] = scale_coords(img_np.shape, det[:, :4], img_np.shape).round()
+
+            # Print results
+            for c in det[:, -1].unique():
+                n = (det[:, -1] == c).sum()  # detections per class
+
+            # Write results
+            for *xyxy, conf, cls in det:
+                label = '%s %.2f' % (names[int(cls)], conf)
+                plot_one_box(xyxy, img_np, label=label, color=colors[int(cls)], line_thickness=3)
+    cv2.imwrite('/tmp/aaa.jpg',img_np[:,:,::-1])
+    return Image.fromarray(img_np)
+    
+
+with torch.no_grad():
+    cfg = 'cfg/yolor_p6.cfg'
+    imgsz = 1280
+    names = 'data/coco.names'
+    weights = 'yolor_p6.pt'
+
+    # Load model
+    model = Darknet(cfg, imgsz)
+    model.load_state_dict(torch.load(weights)['model'])
+    model.eval()
+
+    # Get names and colors
+    names = load_classes(names)
+    colors = [[random.randint(0, 255) for _ in range(3)] for _ in range(len(names))]
+
+    def inference(image):
+        image = image.resize(size=(imgsz, imgsz))
+        return detect(image, names)
+
+    title = "YOLOR P6"
+    description = "demo for YOLOR. To use it, simply upload your image, or click one of the examples to load them. Read more at the links below.\nModel: YOLOR-P6"
+    article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2105.04206'>You Only Learn One Representation: Unified Network for Multiple Tasks</a> | <a href='https://github.com/WongKinYiu/yolor'>Github Repo</a></p>"
+
+    gr.Interface(
+        inference, 
+        [gr.inputs.Image(type="pil", label="Input")], 
+        gr.outputs.Image(type="numpy", label="Output"),
+        title=title,
+        description=description,
+        article=article,
+        examples=[
+                ["city1.jpg"],
+                ["city2.jpg"]
+            ]).launch()

cfg/yolor_csp.cfg

@@ -0,0 +1,1376 @@
+[net]
+# Testing
+#batch=1
+#subdivisions=1
+# Training
+batch=64
+subdivisions=8
+width=512
+height=512
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.00261
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+#cutmix=1
+mosaic=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# 0
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=silu
+
+# P1
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+# 4 (previous+1+3k)
+[shortcut]
+from=-3
+activation=linear
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-10
+
+# Transition last
+
+# 17 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(4+3k)]
+
+[route]
+layers = -1,-28
+
+# Transition last
+
+# 48 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-28
+
+# Transition last
+
+# 79 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-16
+
+# Transition last
+
+# 98 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1024
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[route]
+layers = -1, -13
+
+# 113 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# End of CSPSPP
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 79
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -6
+
+# Transition last
+
+# 127 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 48
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -6
+
+# Transition last
+
+# 141 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=256
+activation=silu
+
+[route]
+layers = -1, 127
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[route]
+layers = -1,-6
+
+# Transition last
+
+# 152 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=512
+activation=silu
+
+[route]
+layers = -1, 113
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[route]
+layers = -1,-6
+
+# Transition last
+
+# 163 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Neck ============ #
+
+# 164
+[implicit_add]
+filters=256
+
+# 165
+[implicit_add]
+filters=512
+
+# 166
+[implicit_add]
+filters=1024
+
+# 167
+[implicit_mul]
+filters=255
+
+# 168
+[implicit_mul]
+filters=255
+
+# 169
+[implicit_mul]
+filters=255
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 141
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[shift_channels]
+from=164
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=167
+
+[yolo]
+mask = 0,1,2
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-4
+
+[route]
+layers = 152
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[shift_channels]
+from=165
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=168
+
+[yolo]
+mask = 3,4,5
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-5
+
+[route]
+layers = 163
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1024
+activation=silu
+
+[shift_channels]
+from=166
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=169
+
+[yolo]
+mask = 6,7,8
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6

cfg/yolor_csp_x.cfg

@@ -0,0 +1,1576 @@
+[net]
+# Testing
+#batch=1
+#subdivisions=1
+# Training
+batch=64
+subdivisions=8
+width=512
+height=512
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.00261
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+#cutmix=1
+mosaic=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# 0
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=silu
+
+# P1
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=40
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=silu
+
+# 4 (previous+1+3k)
+[shortcut]
+from=-3
+activation=linear
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-13
+
+# Transition last
+
+# 20 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(4+3k)]
+
+[route]
+layers = -1,-34
+
+# Transition last
+
+# 57 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-34
+
+# Transition last
+
+# 94 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1280
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-19
+
+# Transition last
+
+# 116 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1280
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[route]
+layers = -1, -15
+
+# 133 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# End of CSPSPP
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 94
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 149 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 57
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 165 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=320
+activation=silu
+
+[route]
+layers = -1, 149
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 178 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=640
+activation=silu
+
+[route]
+layers = -1, 133
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 191 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Neck ============ #
+
+# 192
+[implicit_add]
+filters=320
+
+# 193
+[implicit_add]
+filters=640
+
+# 194
+[implicit_add]
+filters=1280
+
+# 195
+[implicit_mul]
+filters=255
+
+# 196
+[implicit_mul]
+filters=255
+
+# 197
+[implicit_mul]
+filters=255
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 165
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[shift_channels]
+from=192
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=195
+
+[yolo]
+mask = 0,1,2
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-4
+
+[route]
+layers = 178
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[shift_channels]
+from=193
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=196
+
+[yolo]
+mask = 3,4,5
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-5
+
+[route]
+layers = 191
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1280
+activation=silu
+
+[shift_channels]
+from=194
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=197
+
+[yolo]
+mask = 6,7,8
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6

cfg/yolor_p6.cfg

@@ -0,0 +1,1760 @@
+[net]
+batch=64
+subdivisions=8
+width=1280
+height=1280
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.00261
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+mosaic=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# P1
+
+# Downsample
+
+# 0
+[reorg]
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=64
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-12
+
+# Transition last
+
+# 16 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=128
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-24
+
+# Transition last
+
+# 43 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=192
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-24
+
+# Transition last
+
+# 70 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=256
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-12
+
+# Transition last
+
+# 85 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# P6
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=320
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-12
+
+# Transition last
+
+# 100 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[route]
+layers = -1, -13
+
+# 115 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# End of CSPSPP
+
+
+# FPN-5
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 85
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 131 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 70
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=192
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=192
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=192
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 147 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 43
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 163 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=192
+activation=silu
+
+[route]
+layers = -1, 147
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=192
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=192
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=192
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 176 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=192
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=256
+activation=silu
+
+[route]
+layers = -1, 131
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 189 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-6
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=320
+activation=silu
+
+[route]
+layers = -1, 115
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 202 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Neck ============ #
+
+# 203
+[implicit_add]
+filters=256
+
+# 204
+[implicit_add]
+filters=384
+
+# 205
+[implicit_add]
+filters=512
+
+# 206
+[implicit_add]
+filters=640
+
+# 207
+[implicit_mul]
+filters=255
+
+# 208
+[implicit_mul]
+filters=255
+
+# 209
+[implicit_mul]
+filters=255
+
+# 210
+[implicit_mul]
+filters=255
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 163
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[shift_channels]
+from=203
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=207
+
+[yolo]
+mask = 0,1,2
+anchors = 19,27,  44,40,  38,94,  96,68,  86,152,  180,137,  140,301,  303,264,  238,542,  436,615,  739,380,  925,792
+classes=80
+num=12
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-4
+
+[route]
+layers = 176
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=384
+activation=silu
+
+[shift_channels]
+from=204
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=208
+
+[yolo]
+mask = 3,4,5
+anchors = 19,27,  44,40,  38,94,  96,68,  86,152,  180,137,  140,301,  303,264,  238,542,  436,615,  739,380,  925,792
+classes=80
+num=12
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-5
+
+[route]
+layers = 189
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[shift_channels]
+from=205
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=209
+
+[yolo]
+mask = 6,7,8
+anchors = 19,27,  44,40,  38,94,  96,68,  86,152,  180,137,  140,301,  303,264,  238,542,  436,615,  739,380,  925,792
+classes=80
+num=12
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-6
+
+[route]
+layers = 202
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[shift_channels]
+from=206
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=210
+
+[yolo]
+mask = 9,10,11
+anchors = 19,27,  44,40,  38,94,  96,68,  86,152,  180,137,  140,301,  303,264,  238,542,  436,615,  739,380,  925,792
+classes=80
+num=12
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+# ============ End of Head ============ #

cfg/yolor_w6.cfg

@@ -0,0 +1,1760 @@
+[net]
+batch=64
+subdivisions=8
+width=1280
+height=1280
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.00261
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+mosaic=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# P1
+
+# Downsample
+
+# 0
+[reorg]
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=64
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-12
+
+# Transition last
+
+# 16 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=128
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-24
+
+# Transition last
+
+# 43 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=256
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-24
+
+# Transition last
+
+# 70 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=768
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=384
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-12
+
+# Transition last
+
+# 85 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=768
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# P6
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+#
+#[convolutional]
+#batch_normalize=1
+#filters=512
+#size=1
+#stride=1
+#pad=1
+#activation=silu
+
+# Merge [-1, -(3k+3)]
+
+[route]
+layers = -1,-12
+
+# Transition last
+
+# 100 (previous+6+3k)
+[convolutional]
+batch_normalize=1
+filters=1024
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[route]
+layers = -1, -13
+
+# 115 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# End of CSPSPP
+
+
+# FPN-5
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 85
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=384
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=384
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=384
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 131 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 70
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 147 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 43
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 163 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=256
+activation=silu
+
+[route]
+layers = -1, 147
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 176 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=384
+activation=silu
+
+[route]
+layers = -1, 131
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=384
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=384
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=384
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 189 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=384
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-6
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=512
+activation=silu
+
+[route]
+layers = -1, 115
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 202 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Neck ============ #
+
+# 203
+[implicit_add]
+filters=256
+
+# 204
+[implicit_add]
+filters=512
+
+# 205
+[implicit_add]
+filters=768
+
+# 206
+[implicit_add]
+filters=1024
+
+# 207
+[implicit_mul]
+filters=255
+
+# 208
+[implicit_mul]
+filters=255
+
+# 209
+[implicit_mul]
+filters=255
+
+# 210
+[implicit_mul]
+filters=255
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 163
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[shift_channels]
+from=203
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=207
+
+[yolo]
+mask = 0,1,2
+anchors = 19,27,  44,40,  38,94,  96,68,  86,152,  180,137,  140,301,  303,264,  238,542,  436,615,  739,380,  925,792
+classes=80
+num=12
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-4
+
+[route]
+layers = 176
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[shift_channels]
+from=204
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=208
+
+[yolo]
+mask = 3,4,5
+anchors = 19,27,  44,40,  38,94,  96,68,  86,152,  180,137,  140,301,  303,264,  238,542,  436,615,  739,380,  925,792
+classes=80
+num=12
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-5
+
+[route]
+layers = 189
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=768
+activation=silu
+
+[shift_channels]
+from=205
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=209
+
+[yolo]
+mask = 6,7,8
+anchors = 19,27,  44,40,  38,94,  96,68,  86,152,  180,137,  140,301,  303,264,  238,542,  436,615,  739,380,  925,792
+classes=80
+num=12
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-6
+
+[route]
+layers = 202
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1024
+activation=silu
+
+[shift_channels]
+from=206
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[control_channels]
+from=210
+
+[yolo]
+mask = 9,10,11
+anchors = 19,27,  44,40,  38,94,  96,68,  86,152,  180,137,  140,301,  303,264,  238,542,  436,615,  739,380,  925,792
+classes=80
+num=12
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+# ============ End of Head ============ #

cfg/yolov4_csp.cfg

@@ -0,0 +1,1334 @@
+[net]
+# Testing
+#batch=1
+#subdivisions=1
+# Training
+batch=64
+subdivisions=8
+width=512
+height=512
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.00261
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+#cutmix=1
+mosaic=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# 0
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=silu
+
+# P1
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+# 4 (previous+1+3k)
+[shortcut]
+from=-3
+activation=linear
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-10
+
+# Transition last
+
+# 17 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(4+3k)]
+
+[route]
+layers = -1,-28
+
+# Transition last
+
+# 48 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-28
+
+# Transition last
+
+# 79 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-16
+
+# Transition last
+
+# 98 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1024
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[route]
+layers = -1, -13
+
+# 113 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# End of CSPSPP
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 79
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -6
+
+# Transition last
+
+# 127 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 48
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -6
+
+# Transition last
+
+# 141 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=256
+activation=silu
+
+[route]
+layers = -1, 127
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[route]
+layers = -1,-6
+
+# Transition last
+
+# 152 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=512
+activation=silu
+
+[route]
+layers = -1, 113
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[route]
+layers = -1,-6
+
+# Transition last
+
+# 163 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Neck ============ #
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 141
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=silu
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[yolo]
+mask = 0,1,2
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-4
+
+[route]
+layers = 152
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=silu
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[yolo]
+mask = 3,4,5
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-5
+
+[route]
+layers = 163
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1024
+activation=silu
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[yolo]
+mask = 6,7,8
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6

cfg/yolov4_csp_x.cfg

@@ -0,0 +1,1534 @@
+[net]
+# Testing
+#batch=1
+#subdivisions=1
+# Training
+batch=64
+subdivisions=8
+width=512
+height=512
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.00261
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+#cutmix=1
+mosaic=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# 0
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=silu
+
+# P1
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=40
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=silu
+
+# 4 (previous+1+3k)
+[shortcut]
+from=-3
+activation=linear
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-13
+
+# Transition last
+
+# 20 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(4+3k)]
+
+[route]
+layers = -1,-34
+
+# Transition last
+
+# 57 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-34
+
+# Transition last
+
+# 94 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1280
+size=3
+stride=2
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=silu
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-19
+
+# Transition last
+
+# 116 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1280
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[route]
+layers = -1, -15
+
+# 133 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# End of CSPSPP
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 94
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 149 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[upsample]
+stride=2
+
+[route]
+layers = 57
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=silu
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 165 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=320
+activation=silu
+
+[route]
+layers = -1, 149
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 178 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=silu
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=640
+activation=silu
+
+[route]
+layers = -1, 133
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 191 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=silu
+
+# ============ End of Neck ============ #
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 165
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=silu
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[yolo]
+mask = 0,1,2
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-4
+
+[route]
+layers = 178
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=silu
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[yolo]
+mask = 3,4,5
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-5
+
+[route]
+layers = 191
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1280
+activation=silu
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[yolo]
+mask = 6,7,8
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6

cfg/yolov4_p6.cfg

@@ -0,0 +1,2260 @@
+[net]
+batch=64
+subdivisions=8
+width=1280
+height=1280
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.00261
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+mosaic=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# 0
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=mish
+
+
+# P1
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-7
+
+# Transition last
+
+# 10 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-13
+
+# Transition last
+
+# 26 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-49
+
+# Transition last
+
+# 78 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-49
+
+# Transition last
+
+# 130 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-25
+
+# Transition last
+
+# 158 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1024
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P6
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-25
+
+# Transition last
+
+# 186 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1024
+size=1
+stride=1
+pad=1
+activation=mish
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[route]
+layers = -1, -13
+
+# 201 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+# End of CSPSPP
+
+
+# FPN-5
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[upsample]
+stride=2
+
+[route]
+layers = 158
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 217 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[upsample]
+stride=2
+
+[route]
+layers = 130
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=mish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 233 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[upsample]
+stride=2
+
+[route]
+layers = 78
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=mish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 249 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=256
+activation=mish
+
+[route]
+layers = -1, 233
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=mish
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 262 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=512
+activation=mish
+
+[route]
+layers = -1, 217
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 275 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# PAN-6
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=512
+activation=mish
+
+[route]
+layers = -1, 201
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 288 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=mish
+
+# ============ End of Neck ============ #
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 249
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=mish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=340
+activation=linear
+
+[yolo]
+mask = 0,1,2,3
+anchors = 13,17,  31,25,  24,51, 61,45,  61,45,  48,102,  119,96,  97,189,  97,189,  217,184,  171,384,  324,451,  324,451, 545,357, 616,618, 1024,1024
+classes=80
+num=16
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-4
+
+[route]
+layers = 262
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=mish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=340
+activation=linear
+
+[yolo]
+mask = 4,5,6,7
+anchors = 13,17,  31,25,  24,51, 61,45,  61,45,  48,102,  119,96,  97,189,  97,189,  217,184,  171,384,  324,451,  324,451, 545,357, 616,618, 1024,1024
+classes=80
+num=16
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-5
+
+[route]
+layers = 275
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1024
+activation=mish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=340
+activation=linear
+
+[yolo]
+mask = 8,9,10,11
+anchors = 13,17,  31,25,  24,51, 61,45,  61,45,  48,102,  119,96,  97,189,  97,189,  217,184,  171,384,  324,451,  324,451, 545,357, 616,618, 1024,1024
+classes=80
+num=16
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-6
+
+[route]
+layers = 288
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1024
+activation=mish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=340
+activation=linear
+
+[yolo]
+mask = 12,13,14,15
+anchors = 13,17,  31,25,  24,51, 61,45,  61,45,  48,102,  119,96,  97,189,  97,189,  217,184,  171,384,  324,451,  324,451, 545,357, 616,618, 1024,1024
+classes=80
+num=16
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+# ============ End of Head ============ #

cfg/yolov4_p7.cfg

@@ -0,0 +1,2714 @@
+[net]
+batch=64
+subdivisions=8
+width=1536
+height=1536
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.00261
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+mosaic=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# 0
+[convolutional]
+batch_normalize=1
+filters=40
+size=3
+stride=1
+pad=1
+activation=mish
+
+
+# P1
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=40
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=40
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=40
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=40
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=40
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-7
+
+# Transition last
+
+# 10 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-13
+
+# Transition last
+
+# 26 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-49
+
+# Transition last
+
+# 78 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-49
+
+# Transition last
+
+# 130 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1280
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-25
+
+# Transition last
+
+# 158 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1280
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P6
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1280
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-25
+
+# Transition last
+
+# 186 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1280
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# P7
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1280
+size=3
+stride=2
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=mish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-25
+
+# Transition last
+
+# 214 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1280
+size=1
+stride=1
+pad=1
+activation=mish
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[route]
+layers = -1, -13
+
+# 229 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# End of CSPSPP
+
+
+# FPN-6
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[upsample]
+stride=2
+
+[route]
+layers = 186
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 245 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# FPN-5
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[upsample]
+stride=2
+
+[route]
+layers = 158
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 261 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[upsample]
+stride=2
+
+[route]
+layers = 130
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=mish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 277 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[upsample]
+stride=2
+
+[route]
+layers = 78
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=mish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 293 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=320
+activation=mish
+
+[route]
+layers = -1, 277
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=mish
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 306 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=640
+activation=mish
+
+[route]
+layers = -1, 261
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 319 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# PAN-6
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=640
+activation=mish
+
+[route]
+layers = -1, 245
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 332 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+
+# PAN-7
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=640
+activation=mish
+
+[route]
+layers = -1, 229
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 345 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=mish
+
+# ============ End of Neck ============ #
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 293
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=mish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=340
+activation=linear
+
+[yolo]
+mask = 0,1,2,3
+anchors = 13,17,  22,25,  27,66, 55,41,  57,88,  112,69,  69,177,  136,138,  136,138,  287,114,  134,275,  268,248,  268,248,  232,504, 445,416,  640,640,  812,393,  477,808,  1070,908,  1408,1408
+classes=80
+num=20
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-4
+
+[route]
+layers = 306
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=mish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=340
+activation=linear
+
+[yolo]
+mask = 4,5,6,7
+anchors = 13,17,  22,25,  27,66, 55,41,  57,88,  112,69,  69,177,  136,138,  136,138,  287,114,  134,275,  268,248,  268,248,  232,504, 445,416,  640,640,  812,393,  477,808,  1070,908,  1408,1408
+classes=80
+num=20
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-5
+
+[route]
+layers = 319
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1280
+activation=mish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=340
+activation=linear
+
+[yolo]
+mask = 8,9,10,11
+anchors = 13,17,  22,25,  27,66, 55,41,  57,88,  112,69,  69,177,  136,138,  136,138,  287,114,  134,275,  268,248,  268,248,  232,504, 445,416,  640,640,  812,393,  477,808,  1070,908,  1408,1408
+classes=80
+num=20
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-6
+
+[route]
+layers = 332
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1280
+activation=mish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=340
+activation=linear
+
+[yolo]
+mask = 12,13,14,15
+anchors = 13,17,  22,25,  27,66, 55,41,  57,88,  112,69,  69,177,  136,138,  136,138,  287,114,  134,275,  268,248,  268,248,  232,504, 445,416,  640,640,  812,393,  477,808,  1070,908,  1408,1408
+classes=80
+num=20
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+
+# YOLO-7
+
+[route]
+layers = 345
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1280
+activation=mish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=340
+activation=linear
+
+[yolo]
+mask = 16,17,18,19
+anchors = 13,17,  22,25,  27,66, 55,41,  57,88,  112,69,  69,177,  136,138,  136,138,  287,114,  134,275,  268,248,  268,248,  232,504, 445,416,  640,640,  812,393,  477,808,  1070,908,  1408,1408
+classes=80
+num=20
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+scale_x_y = 1.05
+iou_thresh=0.213
+cls_normalizer=1.0
+iou_normalizer=0.07
+iou_loss=ciou
+nms_kind=greedynms
+beta_nms=0.6
+
+# ============ End of Head ============ #

darknet/README.md

@@ -0,0 +1,63 @@
+## Model Zoo
+
+| Model | Test Size | AP<sup>val</sup> | AP<sub>50</sub><sup>val</sup> | AP<sub>75</sub><sup>val</sup> | AP<sub>S</sub><sup>val</sup> | AP<sub>M</sub><sup>val</sup> | AP<sub>L</sub><sup>val</sup> | batch1 throughput |
+| :-- | :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: | 
+| **YOLOv4-CSP** | 640 | **49.1%** | **67.7%** | **53.8%** | **32.1%** | **54.4%** | **63.2%** | 76 *fps* |
+| **YOLOR-CSP** | 640 | **49.2%** | **67.6%** | **53.7%** | **32.9%** | **54.4%** | **63.0%** | - |
+|  |  |  |  |  |  |  |
+| **YOLOv4-CSP-X** | 640 | **50.9%** | **69.3%** | **55.4%** | **35.3%** | **55.8%** | **64.8%** | 53 *fps* |
+| **YOLOR-CSP-X** | 640 | **51.1%** | **69.6%** | **55.7%** | **35.7%** | **56.0%** | **65.2%** | - |
+|  |  |  |  |  |  |  |
+
+## Installation
+
+https://github.com/AlexeyAB/darknet
+
+Docker environment (recommended)
+<details><summary> <b>Expand</b> </summary>
+
+```
+# get code
+git clone https://github.com/AlexeyAB/darknet
+
+# create the docker container, you can change the share memory size if you have more.
+nvidia-docker run --name yolor -it -v your_coco_path/:/coco/ -v your_code_path/:/yolor --shm-size=64g nvcr.io/nvidia/pytorch:21.02-py3
+
+# apt install required packages
+apt update
+apt install -y libopencv-dev
+
+# edit Makefile
+#GPU=1
+#CUDNN=1
+#CUDNN_HALF=1
+#OPENCV=1
+#AVX=1
+#OPENMP=1
+#LIBSO=1
+#ZED_CAMERA=0
+#ZED_CAMERA_v2_8=0
+#
+#USE_CPP=0
+#DEBUG=0
+#
+#ARCH= -gencode arch=compute_52,code=[sm_70,compute_70] \
+#      -gencode arch=compute_61,code=[sm_75,compute_75] \
+#      -gencode arch=compute_61,code=[sm_80,compute_80] \
+#      -gencode arch=compute_61,code=[sm_86,compute_86]
+#
+#...
+
+# build
+make -j8
+```
+
+</details>
+
+## Testing
+
+To reproduce inference speed, using:
+
+```
+CUDA_VISIBLE_DEVICES=0 ./darknet detector demo cfg/coco.data cfg/yolov4-csp.cfg weights/yolov4-csp.weights source/test.mp4 -dont_show -benchmark 
+```

darknet/cfg/yolov4-csp-x.cfg

@@ -0,0 +1,1555 @@
+  
+[net]
+# Testing
+#batch=1
+#subdivisions=1
+# Training
+batch=64
+subdivisions=8
+width=640
+height=640
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.001
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+mosaic=1
+
+letter_box=1
+
+ema_alpha=0.9998
+
+#optimized_memory=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# 0
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=swish
+
+# P1
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=40
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=swish
+
+# 4 (previous+1+3k)
+[shortcut]
+from=-3
+activation=linear
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=80
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-13
+
+# Transition last
+
+# 20 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Merge [-1 -(4+3k)]
+
+[route]
+layers = -1,-34
+
+# Transition last
+
+# 57 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-34
+
+# Transition last
+
+# 94 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1280
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-19
+
+# Transition last
+
+# 116 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1280
+size=1
+stride=1
+pad=1
+activation=swish
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=swish
+
+[route]
+layers = -1, -15
+
+# 133 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+# End of CSPSPP
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[upsample]
+stride=2
+
+[route]
+layers = 94
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=swish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 149 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[upsample]
+stride=2
+
+[route]
+layers = 57
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=160
+activation=swish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -8
+
+# Transition last
+
+# 165 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=160
+size=1
+stride=1
+pad=1
+activation=swish
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=320
+activation=swish
+
+[route]
+layers = -1, 149
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=swish
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 178 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=320
+size=1
+stride=1
+pad=1
+activation=swish
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=640
+activation=swish
+
+[route]
+layers = -1, 133
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=swish
+
+[route]
+layers = -1,-8
+
+# Transition last
+
+# 191 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=640
+size=1
+stride=1
+pad=1
+activation=swish
+
+# ============ End of Neck ============ #
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 165
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=320
+activation=swish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=logistic
+
+[yolo]
+mask = 0,1,2
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.1
+scale_x_y = 2.0
+objectness_smooth=1
+ignore_thresh = .7
+truth_thresh = 1
+#random=1
+resize=1.5
+iou_thresh=0.2
+iou_normalizer=0.05
+cls_normalizer=0.5
+obj_normalizer=0.4
+iou_loss=ciou
+nms_kind=diounms
+beta_nms=0.6
+new_coords=1
+max_delta=2
+
+
+# YOLO-4
+
+[route]
+layers = 178
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=640
+activation=swish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=logistic
+
+[yolo]
+mask = 3,4,5
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.1
+scale_x_y = 2.0
+objectness_smooth=1
+ignore_thresh = .7
+truth_thresh = 1
+#random=1
+resize=1.5
+iou_thresh=0.2
+iou_normalizer=0.05
+cls_normalizer=0.5
+obj_normalizer=0.4
+iou_loss=ciou
+nms_kind=diounms
+beta_nms=0.6
+new_coords=1
+max_delta=2
+
+
+# YOLO-5
+
+[route]
+layers = 191
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1280
+activation=swish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=logistic
+
+[yolo]
+mask = 6,7,8
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.1
+scale_x_y = 2.0
+objectness_smooth=1
+ignore_thresh = .7
+truth_thresh = 1
+#random=1
+resize=1.5
+iou_thresh=0.2
+iou_normalizer=0.05
+cls_normalizer=0.5
+obj_normalizer=0.4
+iou_loss=ciou
+nms_kind=diounms
+beta_nms=0.6
+new_coords=1
+max_delta=2

darknet/cfg/yolov4-csp.cfg

@@ -0,0 +1,1354 @@
+[net]
+# Testing
+#batch=1
+#subdivisions=1
+# Training
+batch=64
+subdivisions=8
+width=640
+height=640
+channels=3
+momentum=0.949
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.001
+burn_in=1000
+max_batches = 500500
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+mosaic=1
+
+letter_box=1
+
+ema_alpha=0.9998
+
+#optimized_memory=1
+
+
+# ============ Backbone ============ #
+
+# Stem 
+
+# 0
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=swish
+
+# P1
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=swish
+
+# 4 (previous+1+3k)
+[shortcut]
+from=-3
+activation=linear
+
+# P2
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Merge [-1, -(3k+4)]
+
+[route]
+layers = -1,-10
+
+# Transition last
+
+# 17 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+# P3
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Merge [-1 -(4+3k)]
+
+[route]
+layers = -1,-28
+
+# Transition last
+
+# 48 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+# P4
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-28
+
+# Transition last
+
+# 79 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+# P5
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=2
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Residual Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=swish
+
+[shortcut]
+from=-3
+activation=linear
+
+# Transition first
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Merge [-1 -(3k+4)]
+
+[route]
+layers = -1,-16
+
+# Transition last
+
+# 98 (previous+7+3k)
+[convolutional]
+batch_normalize=1
+filters=1024
+size=1
+stride=1
+pad=1
+activation=swish
+
+# ============ End of Backbone ============ #
+
+# ============ Neck ============ #
+
+# CSPSPP
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+### SPP ###
+[maxpool]
+stride=1
+size=5
+
+[route]
+layers=-2
+
+[maxpool]
+stride=1
+size=9
+
+[route]
+layers=-4
+
+[maxpool]
+stride=1
+size=13
+
+[route]
+layers=-1,-3,-5,-6
+### End SPP ###
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=swish
+
+[route]
+layers = -1, -13
+
+# 113 (previous+6+5+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+# End of CSPSPP
+
+
+# FPN-4
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[upsample]
+stride=2
+
+[route]
+layers = 79
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=swish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -6
+
+# Transition last
+
+# 127 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+
+# FPN-3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[upsample]
+stride=2
+
+[route]
+layers = 48
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -1, -3
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=128
+activation=swish
+
+# Merge [-1, -(2k+2)]
+
+[route]
+layers = -1, -6
+
+# Transition last
+
+# 141 (previous+6+4+2k)
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=swish
+
+
+# PAN-4
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=256
+activation=swish
+
+[route]
+layers = -1, 127
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=swish
+
+[route]
+layers = -1,-6
+
+# Transition last
+
+# 152 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=swish
+
+
+# PAN-5
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=2
+pad=1
+filters=512
+activation=swish
+
+[route]
+layers = -1, 113
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+# Split
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[route]
+layers = -2
+
+# Plain Block
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=swish
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=swish
+
+[route]
+layers = -1,-6
+
+# Transition last
+
+# 163 (previous+3+4+2k)
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=swish
+
+# ============ End of Neck ============ #
+
+# ============ Head ============ #
+
+# YOLO-3
+
+[route]
+layers = 141
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=swish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=logistic
+
+[yolo]
+mask = 0,1,2
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.1
+scale_x_y = 2.0
+objectness_smooth=1
+ignore_thresh = .7
+truth_thresh = 1
+#random=1
+resize=1.5
+iou_thresh=0.2
+iou_normalizer=0.05
+cls_normalizer=0.5
+obj_normalizer=0.4
+iou_loss=ciou
+nms_kind=diounms
+beta_nms=0.6
+new_coords=1
+max_delta=2
+
+
+# YOLO-4
+
+[route]
+layers = 152
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=swish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=logistic
+
+[yolo]
+mask = 3,4,5
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.1
+scale_x_y = 2.0
+objectness_smooth=1
+ignore_thresh = .7
+truth_thresh = 1
+#random=1
+resize=1.5
+iou_thresh=0.2
+iou_normalizer=0.05
+cls_normalizer=0.5
+obj_normalizer=0.4
+iou_loss=ciou
+nms_kind=diounms
+beta_nms=0.6
+new_coords=1
+max_delta=2
+
+
+# YOLO-5
+
+[route]
+layers = 163
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1024
+activation=swish
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=logistic
+
+[yolo]
+mask = 6,7,8
+anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
+classes=80
+num=9
+jitter=.1
+scale_x_y = 2.0
+objectness_smooth=1
+ignore_thresh = .7
+truth_thresh = 1
+#random=1
+resize=1.5
+iou_thresh=0.2
+iou_normalizer=0.05
+cls_normalizer=0.5
+obj_normalizer=0.4
+iou_loss=ciou
+nms_kind=diounms
+beta_nms=0.6
+new_coords=1
+max_delta=2

darknet/new_layers.md

@@ -0,0 +1,329 @@
+![Implicit Modeling](https://github.com/WongKinYiu/yolor/blob/main/figure/implicit_modeling.png)
+
+### 1. silence layer
+
+Usage:
+
+```
+[silence]
+```
+
+PyTorch code:
+
+``` python
+class Silence(nn.Module):
+    def __init__(self):
+        super(Silence, self).__init__()
+    def forward(self, x):    
+        return x
+```
+
+
+### 2. implicit_add layer
+
+Usage:
+
+```
+[implicit_add]
+filters=128
+```
+
+PyTorch code:
+
+``` python
+class ImplicitA(nn.Module):
+    def __init__(self, channel):
+        super(ImplicitA, self).__init__()
+        self.channel = channel
+        self.implicit = nn.Parameter(torch.zeros(1, channel, 1, 1))
+        nn.init.normal_(self.implicit, std=.02)
+
+    def forward(self):
+        return self.implicit
+```
+
+
+### 3. shift_channels layer
+
+Usage:
+
+```
+[shift_channels]
+from=101
+```
+
+PyTorch code:
+
+``` python
+class ShiftChannel(nn.Module):
+    def __init__(self, layers):
+        super(ShiftChannel, self).__init__()
+        self.layers = layers  # layer indices
+
+    def forward(self, x, outputs):
+        a = outputs[self.layers[0]]
+        return a.expand_as(x) + x
+```
+
+
+### 4. implicit_mul layer
+
+Usage:
+
+```
+[implicit_mul]
+filters=128
+```
+
+PyTorch code:
+
+``` python
+class ImplicitM(nn.Module):
+    def __init__(self, channel):
+        super(ImplicitM, self).__init__()
+        self.channel = channel
+        self.implicit = nn.Parameter(torch.ones(1, channel, 1, 1))
+        nn.init.normal_(self.implicit, mean=1., std=.02)
+
+    def forward(self):
+        return self.implicit
+```
+
+
+### 5. control_channels layer
+
+Usage:
+
+```
+[control_channels]
+from=101
+```
+
+PyTorch code:
+
+``` python
+class ControlChannel(nn.Module):
+    def __init__(self, layers):
+        super(ControlChannel, self).__init__()
+        self.layers = layers  # layer indices
+
+    def forward(self, x, outputs):
+        a = outputs[self.layers[0]]
+        return a.expand_as(x) * x
+```
+
+
+### 6. implicit_cat layer
+
+Usage:
+
+```
+[implicit_cat]
+filters=128
+```
+
+PyTorch code: (same as ImplicitA)
+
+``` python
+class ImplicitC(nn.Module):
+    def __init__(self, channel):
+        super(ImplicitC, self).__init__()
+        self.channel = channel
+        self.implicit = nn.Parameter(torch.zeros(1, channel, 1, 1))
+        nn.init.normal_(self.implicit, std=.02)
+
+    def forward(self):
+        return self.implicit
+```
+
+
+### 7. alternate_channels layer
+
+Usage:
+
+```
+[alternate_channels]
+from=101
+```
+
+PyTorch code:
+
+``` python
+class AlternateChannel(nn.Module):
+    def __init__(self, layers):
+        super(AlternateChannel, self).__init__()
+        self.layers = layers  # layer indices
+
+    def forward(self, x, outputs):
+        a = outputs[self.layers[0]]
+        return torch.cat([a.expand_as(x), x], dim=1)
+```
+
+
+### 8. implicit_add_2d layer
+
+Usage:
+
+```
+[implicit_add_2d]
+filters=128
+atoms=128
+```
+
+PyTorch code:
+
+``` python
+class Implicit2DA(nn.Module):
+    def __init__(self, atom, channel):
+        super(Implicit2DA, self).__init__()
+        self.channel = channel
+        self.implicit = nn.Parameter(torch.zeros(1, atom, channel, 1))
+        nn.init.normal_(self.implicit, std=.02)
+
+    def forward(self):
+        return self.implicit
+```
+
+
+### 9. shift_channels_2d layer
+
+Usage:
+
+```
+[shift_channels_2d]
+from=101
+```
+
+PyTorch code:
+
+``` python
+class ShiftChannel2D(nn.Module):
+    def __init__(self, layers):
+        super(ShiftChannel2D, self).__init__()
+        self.layers = layers  # layer indices
+
+    def forward(self, x, outputs):
+        a = outputs[self.layers[0]].view(1,-1,1,1)
+        return a.expand_as(x) + x
+```
+
+
+### 10. implicit_mul_2d layer
+
+Usage:
+
+```
+[implicit_mul_2d]
+filters=128
+atoms=128
+```
+
+PyTorch code:
+
+``` python
+class Implicit2DM(nn.Module):
+    def __init__(self, atom, channel):
+        super(Implicit2DM, self).__init__()
+        self.channel = channel
+        self.implicit = nn.Parameter(torch.ones(1, atom, channel, 1))
+        nn.init.normal_(self.implicit, mean=1., std=.02)
+
+    def forward(self):
+        return self.implicit
+```
+
+
+### 11. control_channels_2d layer
+
+Usage:
+
+```
+[control_channels_2d]
+from=101
+```
+
+PyTorch code:
+
+``` python
+class ControlChannel2D(nn.Module):
+    def __init__(self, layers):
+        super(ControlChannel2D, self).__init__()
+        self.layers = layers  # layer indices
+
+    def forward(self, x, outputs):
+        a = outputs[self.layers[0]].view(1,-1,1,1)
+        return a.expand_as(x) * x
+```
+
+
+### 12. implicit_cat_2d layer
+
+Usage:
+
+```
+[implicit_cat_2d]
+filters=128
+atoms=128
+```
+
+PyTorch code: (same as Implicit2DA)
+
+``` python
+class Implicit2DC(nn.Module):
+    def __init__(self, atom, channel):
+        super(Implicit2DC, self).__init__()
+        self.channel = channel
+        self.implicit = nn.Parameter(torch.zeros(1, atom, channel, 1))
+        nn.init.normal_(self.implicit, std=.02)
+
+    def forward(self):
+        return self.implicit
+```
+
+
+### 13. alternate_channels_2d layer
+
+Usage:
+
+```
+[alternate_channels_2d]
+from=101
+```
+
+PyTorch code:
+
+``` python
+class AlternateChannel2D(nn.Module):
+    def __init__(self, layers):
+        super(AlternateChannel2D, self).__init__()
+        self.layers = layers  # layer indices
+
+    def forward(self, x, outputs):
+        a = outputs[self.layers[0]].view(1,-1,1,1)
+        return torch.cat([a.expand_as(x), x], dim=1)
+```
+
+
+### 14. dwt layer
+
+Usage:
+
+```
+[dwt]
+```
+
+PyTorch code:
+
+``` python
+# https://github.com/fbcotter/pytorch_wavelets
+from pytorch_wavelets import DWTForward, DWTInverse
+class DWT(nn.Module):
+    def __init__(self):
+        super(DWT, self).__init__()
+        self.xfm = DWTForward(J=1, wave='db1', mode='zero')
+
+    def forward(self, x):
+        b,c,w,h = x.shape
+        yl, yh = self.xfm(x)
+        return torch.cat([yl/2., yh[0].view(b,-1,w//2,h//2)/2.+.5], 1)
+```

data/coco.names

@@ -0,0 +1,80 @@
+person
+bicycle
+car
+motorcycle
+airplane
+bus
+train
+truck
+boat
+traffic light
+fire hydrant
+stop sign
+parking meter
+bench
+bird
+cat
+dog
+horse
+sheep
+cow
+elephant
+bear
+zebra
+giraffe
+backpack
+umbrella
+handbag
+tie
+suitcase
+frisbee
+skis
+snowboard
+sports ball
+kite
+baseball bat
+baseball glove
+skateboard
+surfboard
+tennis racket
+bottle
+wine glass
+cup
+fork
+knife
+spoon
+bowl
+banana
+apple
+sandwich
+orange
+broccoli
+carrot
+hot dog
+pizza
+donut
+cake
+chair
+couch
+potted plant
+bed
+dining table
+toilet
+tv
+laptop
+mouse
+remote
+keyboard
+cell phone
+microwave
+oven
+toaster
+sink
+refrigerator
+book
+clock
+vase
+scissors
+teddy bear
+hair drier
+toothbrush

data/coco.yaml

@@ -0,0 +1,18 @@
+# train and val datasets (image directory or *.txt file with image paths)
+train: ../coco/train2017.txt  # 118k images
+val: ../coco/val2017.txt  # 5k images
+test: ../coco/test-dev2017.txt  # 20k images for submission to https://competitions.codalab.org/competitions/20794
+
+# number of classes
+nc: 80
+
+# class names
+names: ['person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'traffic light',
+        'fire hydrant', 'stop sign', 'parking meter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow',
+        'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee',
+        'skis', 'snowboard', 'sports ball', 'kite', 'baseball bat', 'baseball glove', 'skateboard', 'surfboard',
+        'tennis racket', 'bottle', 'wine glass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple',
+        'sandwich', 'orange', 'broccoli', 'carrot', 'hot dog', 'pizza', 'donut', 'cake', 'chair', 'couch',
+        'potted plant', 'bed', 'dining table', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cell phone',
+        'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddy bear',
+        'hair drier', 'toothbrush']

data/hyp.finetune.1280.yaml

@@ -0,0 +1,28 @@
+lr0: 0.01  # initial learning rate (SGD=1E-2, Adam=1E-3)
+lrf: 0.2  # final OneCycleLR learning rate (lr0 * lrf)
+momentum: 0.937  # SGD momentum/Adam beta1
+weight_decay: 0.0005  # optimizer weight decay 5e-4
+warmup_epochs: 3.0  # warmup epochs (fractions ok)
+warmup_momentum: 0.8  # warmup initial momentum
+warmup_bias_lr: 0.1  # warmup initial bias lr
+box: 0.05  # box loss gain
+cls: 0.5  # cls loss gain
+cls_pw: 1.0  # cls BCELoss positive_weight
+obj: 1.0  # obj loss gain (scale with pixels)
+obj_pw: 1.0  # obj BCELoss positive_weight
+iou_t: 0.20  # IoU training threshold
+anchor_t: 4.0  # anchor-multiple threshold
+# anchors: 3  # anchors per output layer (0 to ignore)
+fl_gamma: 0.0  # focal loss gamma (efficientDet default gamma=1.5)
+hsv_h: 0.015  # image HSV-Hue augmentation (fraction)
+hsv_s: 0.7  # image HSV-Saturation augmentation (fraction)
+hsv_v: 0.4  # image HSV-Value augmentation (fraction)
+degrees: 0.0  # image rotation (+/- deg)
+translate: 0.5  # image translation (+/- fraction)
+scale: 0.8  # image scale (+/- gain)
+shear: 0.0  # image shear (+/- deg)
+perspective: 0.0  # image perspective (+/- fraction), range 0-0.001
+flipud: 0.0  # image flip up-down (probability)
+fliplr: 0.5  # image flip left-right (probability)
+mosaic: 1.0  # image mosaic (probability)
+mixup: 0.2  # image mixup (probability)

data/hyp.scratch.1280.yaml

@@ -0,0 +1,28 @@
+lr0: 0.01  # initial learning rate (SGD=1E-2, Adam=1E-3)
+lrf: 0.2  # final OneCycleLR learning rate (lr0 * lrf)
+momentum: 0.937  # SGD momentum/Adam beta1
+weight_decay: 0.0005  # optimizer weight decay 5e-4
+warmup_epochs: 3.0  # warmup epochs (fractions ok)
+warmup_momentum: 0.8  # warmup initial momentum
+warmup_bias_lr: 0.1  # warmup initial bias lr
+box: 0.05  # box loss gain
+cls: 0.5  # cls loss gain
+cls_pw: 1.0  # cls BCELoss positive_weight
+obj: 1.0  # obj loss gain (scale with pixels)
+obj_pw: 1.0  # obj BCELoss positive_weight
+iou_t: 0.20  # IoU training threshold
+anchor_t: 4.0  # anchor-multiple threshold
+# anchors: 3  # anchors per output layer (0 to ignore)
+fl_gamma: 0.0  # focal loss gamma (efficientDet default gamma=1.5)
+hsv_h: 0.015  # image HSV-Hue augmentation (fraction)
+hsv_s: 0.7  # image HSV-Saturation augmentation (fraction)
+hsv_v: 0.4  # image HSV-Value augmentation (fraction)
+degrees: 0.0  # image rotation (+/- deg)
+translate: 0.5  # image translation (+/- fraction)
+scale: 0.5  # image scale (+/- gain)
+shear: 0.0  # image shear (+/- deg)
+perspective: 0.0  # image perspective (+/- fraction), range 0-0.001
+flipud: 0.0  # image flip up-down (probability)
+fliplr: 0.5  # image flip left-right (probability)
+mosaic: 1.0  # image mosaic (probability)
+mixup: 0.0  # image mixup (probability)

data/hyp.scratch.640.yaml

@@ -0,0 +1,28 @@
+lr0: 0.01  # initial learning rate (SGD=1E-2, Adam=1E-3)
+lrf: 0.2  # final OneCycleLR learning rate (lr0 * lrf)
+momentum: 0.937  # SGD momentum/Adam beta1
+weight_decay: 0.0005  # optimizer weight decay 5e-4
+warmup_epochs: 3.0  # warmup epochs (fractions ok)
+warmup_momentum: 0.8  # warmup initial momentum
+warmup_bias_lr: 0.1  # warmup initial bias lr
+box: 0.05  # box loss gain
+cls: 0.3  # cls loss gain
+cls_pw: 1.0  # cls BCELoss positive_weight
+obj: 0.7  # obj loss gain (scale with pixels)
+obj_pw: 1.0  # obj BCELoss positive_weight
+iou_t: 0.20  # IoU training threshold
+anchor_t: 4.0  # anchor-multiple threshold
+# anchors: 3  # anchors per output layer (0 to ignore)
+fl_gamma: 0.0  # focal loss gamma (efficientDet default gamma=1.5)
+hsv_h: 0.015  # image HSV-Hue augmentation (fraction)
+hsv_s: 0.7  # image HSV-Saturation augmentation (fraction)
+hsv_v: 0.4  # image HSV-Value augmentation (fraction)
+degrees: 0.0  # image rotation (+/- deg)
+translate: 0.1  # image translation (+/- fraction)
+scale: 0.9  # image scale (+/- gain)
+shear: 0.0  # image shear (+/- deg)
+perspective: 0.0  # image perspective (+/- fraction), range 0-0.001
+flipud: 0.0  # image flip up-down (probability)
+fliplr: 0.5  # image flip left-right (probability)
+mosaic: 1.0  # image mosaic (probability)
+mixup: 0.0  # image mixup (probability)

models/__init__.py

@@ -0,0 +1 @@
+

models/export.py

@@ -0,0 +1,68 @@
+import argparse
+
+import torch
+
+from utils.google_utils import attempt_download
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--weights', type=str, default='./yolov4.pt', help='weights path')
+    parser.add_argument('--img-size', nargs='+', type=int, default=[640, 640], help='image size')
+    parser.add_argument('--batch-size', type=int, default=1, help='batch size')
+    opt = parser.parse_args()
+    opt.img_size *= 2 if len(opt.img_size) == 1 else 1  # expand
+    print(opt)
+
+    # Input
+    img = torch.zeros((opt.batch_size, 3, *opt.img_size))  # image size(1,3,320,192) iDetection
+
+    # Load PyTorch model
+    attempt_download(opt.weights)
+    model = torch.load(opt.weights, map_location=torch.device('cpu'))['model'].float()
+    model.eval()
+    model.model[-1].export = True  # set Detect() layer export=True
+    y = model(img)  # dry run
+
+    # TorchScript export
+    try:
+        print('\nStarting TorchScript export with torch %s...' % torch.__version__)
+        f = opt.weights.replace('.pt', '.torchscript.pt')  # filename
+        ts = torch.jit.trace(model, img)
+        ts.save(f)
+        print('TorchScript export success, saved as %s' % f)
+    except Exception as e:
+        print('TorchScript export failure: %s' % e)
+
+    # ONNX export
+    try:
+        import onnx
+
+        print('\nStarting ONNX export with onnx %s...' % onnx.__version__)
+        f = opt.weights.replace('.pt', '.onnx')  # filename
+        model.fuse()  # only for ONNX
+        torch.onnx.export(model, img, f, verbose=False, opset_version=12, input_names=['images'],
+                          output_names=['classes', 'boxes'] if y is None else ['output'])
+
+        # Checks
+        onnx_model = onnx.load(f)  # load onnx model
+        onnx.checker.check_model(onnx_model)  # check onnx model
+        print(onnx.helper.printable_graph(onnx_model.graph))  # print a human readable model
+        print('ONNX export success, saved as %s' % f)
+    except Exception as e:
+        print('ONNX export failure: %s' % e)
+
+    # CoreML export
+    try:
+        import coremltools as ct
+
+        print('\nStarting CoreML export with coremltools %s...' % ct.__version__)
+        # convert model from torchscript and apply pixel scaling as per detect.py
+        model = ct.convert(ts, inputs=[ct.ImageType(name='images', shape=img.shape, scale=1 / 255.0, bias=[0, 0, 0])])
+        f = opt.weights.replace('.pt', '.mlmodel')  # filename
+        model.save(f)
+        print('CoreML export success, saved as %s' % f)
+    except Exception as e:
+        print('CoreML export failure: %s' % e)
+
+    # Finish
+    print('\nExport complete. Visualize with https://github.com/lutzroeder/netron.')

models/models.py

@@ -0,0 +1,761 @@
+from utils.google_utils import *
+from utils.layers import *
+from utils.parse_config import *
+from utils import torch_utils
+
+ONNX_EXPORT = False
+
+
+def create_modules(module_defs, img_size, cfg):
+    # Constructs module list of layer blocks from module configuration in module_defs
+
+    img_size = [img_size] * 2 if isinstance(img_size, int) else img_size  # expand if necessary
+    _ = module_defs.pop(0)  # cfg training hyperparams (unused)
+    output_filters = [3]  # input channels
+    module_list = nn.ModuleList()
+    routs = []  # list of layers which rout to deeper layers
+    yolo_index = -1
+
+    for i, mdef in enumerate(module_defs):
+        modules = nn.Sequential()
+
+        if mdef['type'] == 'convolutional':
+            bn = mdef['batch_normalize']
+            filters = mdef['filters']
+            k = mdef['size']  # kernel size
+            stride = mdef['stride'] if 'stride' in mdef else (mdef['stride_y'], mdef['stride_x'])
+            if isinstance(k, int):  # single-size conv
+                modules.add_module('Conv2d', nn.Conv2d(in_channels=output_filters[-1],
+                                                       out_channels=filters,
+                                                       kernel_size=k,
+                                                       stride=stride,
+                                                       padding=k // 2 if mdef['pad'] else 0,
+                                                       groups=mdef['groups'] if 'groups' in mdef else 1,
+                                                       bias=not bn))
+            else:  # multiple-size conv
+                modules.add_module('MixConv2d', MixConv2d(in_ch=output_filters[-1],
+                                                          out_ch=filters,
+                                                          k=k,
+                                                          stride=stride,
+                                                          bias=not bn))
+
+            if bn:
+                modules.add_module('BatchNorm2d', nn.BatchNorm2d(filters, momentum=0.03, eps=1E-4))
+            else:
+                routs.append(i)  # detection output (goes into yolo layer)
+
+            if mdef['activation'] == 'leaky':  # activation study https://github.com/ultralytics/yolov3/issues/441
+                modules.add_module('activation', nn.LeakyReLU(0.1, inplace=True))
+            elif mdef['activation'] == 'swish':
+                modules.add_module('activation', Swish())
+            elif mdef['activation'] == 'mish':
+                modules.add_module('activation', Mish())
+            elif mdef['activation'] == 'emb':
+                modules.add_module('activation', F.normalize())
+            elif mdef['activation'] == 'logistic':
+                modules.add_module('activation', nn.Sigmoid())
+            elif mdef['activation'] == 'silu':
+                modules.add_module('activation', nn.SiLU())
+
+        elif mdef['type'] == 'deformableconvolutional':
+            bn = mdef['batch_normalize']
+            filters = mdef['filters']
+            k = mdef['size']  # kernel size
+            stride = mdef['stride'] if 'stride' in mdef else (mdef['stride_y'], mdef['stride_x'])
+            if isinstance(k, int):  # single-size conv
+                modules.add_module('DeformConv2d', DeformConv2d(output_filters[-1],
+                                                       filters,
+                                                       kernel_size=k,
+                                                       padding=k // 2 if mdef['pad'] else 0,
+                                                       stride=stride,
+                                                       bias=not bn,
+                                                       modulation=True))
+            else:  # multiple-size conv
+                modules.add_module('MixConv2d', MixConv2d(in_ch=output_filters[-1],
+                                                          out_ch=filters,
+                                                          k=k,
+                                                          stride=stride,
+                                                          bias=not bn))
+
+            if bn:
+                modules.add_module('BatchNorm2d', nn.BatchNorm2d(filters, momentum=0.03, eps=1E-4))
+            else:
+                routs.append(i)  # detection output (goes into yolo layer)
+
+            if mdef['activation'] == 'leaky':  # activation study https://github.com/ultralytics/yolov3/issues/441
+                modules.add_module('activation', nn.LeakyReLU(0.1, inplace=True))
+            elif mdef['activation'] == 'swish':
+                modules.add_module('activation', Swish())
+            elif mdef['activation'] == 'mish':
+                modules.add_module('activation', Mish())
+            elif mdef['activation'] == 'silu':
+                modules.add_module('activation', nn.SiLU())
+                
+        elif mdef['type'] == 'dropout':
+            p = mdef['probability']
+            modules = nn.Dropout(p)
+
+        elif mdef['type'] == 'avgpool':
+            modules = GAP()
+
+        elif mdef['type'] == 'silence':
+            filters = output_filters[-1]
+            modules = Silence()
+
+        elif mdef['type'] == 'scale_channels':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1]
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = ScaleChannel(layers=layers)
+
+        elif mdef['type'] == 'shift_channels':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1]
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = ShiftChannel(layers=layers)
+
+        elif mdef['type'] == 'shift_channels_2d':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1]
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = ShiftChannel2D(layers=layers)
+
+        elif mdef['type'] == 'control_channels':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1]
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = ControlChannel(layers=layers)
+
+        elif mdef['type'] == 'control_channels_2d':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1]
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = ControlChannel2D(layers=layers)
+
+        elif mdef['type'] == 'alternate_channels':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1] * 2
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = AlternateChannel(layers=layers)
+
+        elif mdef['type'] == 'alternate_channels_2d':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1] * 2
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = AlternateChannel2D(layers=layers)
+
+        elif mdef['type'] == 'select_channels':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1]
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = SelectChannel(layers=layers)
+
+        elif mdef['type'] == 'select_channels_2d':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1]
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = SelectChannel2D(layers=layers)
+
+        elif mdef['type'] == 'sam':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1]
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = ScaleSpatial(layers=layers)
+
+        elif mdef['type'] == 'BatchNorm2d':
+            filters = output_filters[-1]
+            modules = nn.BatchNorm2d(filters, momentum=0.03, eps=1E-4)
+            if i == 0 and filters == 3:  # normalize RGB image
+                # imagenet mean and var https://pytorch.org/docs/stable/torchvision/models.html#classification
+                modules.running_mean = torch.tensor([0.485, 0.456, 0.406])
+                modules.running_var = torch.tensor([0.0524, 0.0502, 0.0506])
+
+        elif mdef['type'] == 'maxpool':
+            k = mdef['size']  # kernel size
+            stride = mdef['stride']
+            maxpool = nn.MaxPool2d(kernel_size=k, stride=stride, padding=(k - 1) // 2)
+            if k == 2 and stride == 1:  # yolov3-tiny
+                modules.add_module('ZeroPad2d', nn.ZeroPad2d((0, 1, 0, 1)))
+                modules.add_module('MaxPool2d', maxpool)
+            else:
+                modules = maxpool
+
+        elif mdef['type'] == 'local_avgpool':
+            k = mdef['size']  # kernel size
+            stride = mdef['stride']
+            avgpool = nn.AvgPool2d(kernel_size=k, stride=stride, padding=(k - 1) // 2)
+            if k == 2 and stride == 1:  # yolov3-tiny
+                modules.add_module('ZeroPad2d', nn.ZeroPad2d((0, 1, 0, 1)))
+                modules.add_module('AvgPool2d', avgpool)
+            else:
+                modules = avgpool
+
+        elif mdef['type'] == 'upsample':
+            if ONNX_EXPORT:  # explicitly state size, avoid scale_factor
+                g = (yolo_index + 1) * 2 / 32  # gain
+                modules = nn.Upsample(size=tuple(int(x * g) for x in img_size))  # img_size = (320, 192)
+            else:
+                modules = nn.Upsample(scale_factor=mdef['stride'])
+
+        elif mdef['type'] == 'route':  # nn.Sequential() placeholder for 'route' layer
+            layers = mdef['layers']
+            filters = sum([output_filters[l + 1 if l > 0 else l] for l in layers])
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = FeatureConcat(layers=layers)
+
+        elif mdef['type'] == 'route2':  # nn.Sequential() placeholder for 'route' layer
+            layers = mdef['layers']
+            filters = sum([output_filters[l + 1 if l > 0 else l] for l in layers])
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = FeatureConcat2(layers=layers)
+
+        elif mdef['type'] == 'route3':  # nn.Sequential() placeholder for 'route' layer
+            layers = mdef['layers']
+            filters = sum([output_filters[l + 1 if l > 0 else l] for l in layers])
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = FeatureConcat3(layers=layers)
+
+        elif mdef['type'] == 'route_lhalf':  # nn.Sequential() placeholder for 'route' layer
+            layers = mdef['layers']
+            filters = sum([output_filters[l + 1 if l > 0 else l] for l in layers])//2
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = FeatureConcat_l(layers=layers)
+
+        elif mdef['type'] == 'shortcut':  # nn.Sequential() placeholder for 'shortcut' layer
+            layers = mdef['from']
+            filters = output_filters[-1]
+            routs.extend([i + l if l < 0 else l for l in layers])
+            modules = WeightedFeatureFusion(layers=layers, weight='weights_type' in mdef)
+
+        elif mdef['type'] == 'reorg3d':  # yolov3-spp-pan-scale
+            pass
+
+        elif mdef['type'] == 'reorg':  # yolov3-spp-pan-scale
+            filters = 4 * output_filters[-1]
+            modules.add_module('Reorg', Reorg())
+
+        elif mdef['type'] == 'dwt':  # yolov3-spp-pan-scale
+            filters = 4 * output_filters[-1]
+            modules.add_module('DWT', DWT())
+
+        elif mdef['type'] == 'implicit_add':  # yolov3-spp-pan-scale
+            filters = mdef['filters']
+            modules = ImplicitA(channel=filters)
+
+        elif mdef['type'] == 'implicit_mul':  # yolov3-spp-pan-scale
+            filters = mdef['filters']
+            modules = ImplicitM(channel=filters)
+
+        elif mdef['type'] == 'implicit_cat':  # yolov3-spp-pan-scale
+            filters = mdef['filters']
+            modules = ImplicitC(channel=filters)
+
+        elif mdef['type'] == 'implicit_add_2d':  # yolov3-spp-pan-scale
+            channels = mdef['filters']
+            filters = mdef['atoms']
+            modules = Implicit2DA(atom=filters, channel=channels)
+
+        elif mdef['type'] == 'implicit_mul_2d':  # yolov3-spp-pan-scale
+            channels = mdef['filters']
+            filters = mdef['atoms']
+            modules = Implicit2DM(atom=filters, channel=channels)
+
+        elif mdef['type'] == 'implicit_cat_2d':  # yolov3-spp-pan-scale
+            channels = mdef['filters']
+            filters = mdef['atoms']
+            modules = Implicit2DC(atom=filters, channel=channels)
+
+        elif mdef['type'] == 'yolo':
+            yolo_index += 1
+            stride = [8, 16, 32, 64, 128]  # P3, P4, P5, P6, P7 strides
+            if any(x in cfg for x in ['yolov4-tiny', 'fpn', 'yolov3']):  # P5, P4, P3 strides
+                stride = [32, 16, 8]
+            layers = mdef['from'] if 'from' in mdef else []
+            modules = YOLOLayer(anchors=mdef['anchors'][mdef['mask']],  # anchor list
+                                nc=mdef['classes'],  # number of classes
+                                img_size=img_size,  # (416, 416)
+                                yolo_index=yolo_index,  # 0, 1, 2...
+                                layers=layers,  # output layers
+                                stride=stride[yolo_index])
+
+            # Initialize preceding Conv2d() bias (https://arxiv.org/pdf/1708.02002.pdf section 3.3)
+            try:
+                j = layers[yolo_index] if 'from' in mdef else -2
+                bias_ = module_list[j][0].bias  # shape(255,)
+                bias = bias_[:modules.no * modules.na].view(modules.na, -1)  # shape(3,85)
+                #bias[:, 4] += -4.5  # obj
+                bias.data[:, 4] += math.log(8 / (640 / stride[yolo_index]) ** 2)  # obj (8 objects per 640 image)
+                bias.data[:, 5:] += math.log(0.6 / (modules.nc - 0.99))  # cls (sigmoid(p) = 1/nc)
+                module_list[j][0].bias = torch.nn.Parameter(bias_, requires_grad=bias_.requires_grad)
+                
+                #j = [-2, -5, -8]
+                #for sj in j:
+                #    bias_ = module_list[sj][0].bias
+                #    bias = bias_[:modules.no * 1].view(1, -1)
+                #    bias.data[:, 4] += math.log(8 / (640 / stride[yolo_index]) ** 2)
+                #    bias.data[:, 5:] += math.log(0.6 / (modules.nc - 0.99))
+                #    module_list[sj][0].bias = torch.nn.Parameter(bias_, requires_grad=bias_.requires_grad)
+            except:
+                print('WARNING: smart bias initialization failure.')
+
+        elif mdef['type'] == 'jde':
+            yolo_index += 1
+            stride = [8, 16, 32, 64, 128]  # P3, P4, P5, P6, P7 strides
+            if any(x in cfg for x in ['yolov4-tiny', 'fpn', 'yolov3']):  # P5, P4, P3 strides
+                stride = [32, 16, 8]
+            layers = mdef['from'] if 'from' in mdef else []
+            modules = JDELayer(anchors=mdef['anchors'][mdef['mask']],  # anchor list
+                                nc=mdef['classes'],  # number of classes
+                                img_size=img_size,  # (416, 416)
+                                yolo_index=yolo_index,  # 0, 1, 2...
+                                layers=layers,  # output layers
+                                stride=stride[yolo_index])
+
+            # Initialize preceding Conv2d() bias (https://arxiv.org/pdf/1708.02002.pdf section 3.3)
+            try:
+                j = layers[yolo_index] if 'from' in mdef else -1
+                bias_ = module_list[j][0].bias  # shape(255,)
+                bias = bias_[:modules.no * modules.na].view(modules.na, -1)  # shape(3,85)
+                #bias[:, 4] += -4.5  # obj
+                bias.data[:, 4] += math.log(8 / (640 / stride[yolo_index]) ** 2)  # obj (8 objects per 640 image)
+                bias.data[:, 5:] += math.log(0.6 / (modules.nc - 0.99))  # cls (sigmoid(p) = 1/nc)
+                module_list[j][0].bias = torch.nn.Parameter(bias_, requires_grad=bias_.requires_grad)
+            except:
+                print('WARNING: smart bias initialization failure.')
+
+        else:
+            print('Warning: Unrecognized Layer Type: ' + mdef['type'])
+
+        # Register module list and number of output filters
+        module_list.append(modules)
+        output_filters.append(filters)
+
+    routs_binary = [False] * (i + 1)
+    for i in routs:
+        routs_binary[i] = True
+    return module_list, routs_binary
+
+
+class YOLOLayer(nn.Module):
+    def __init__(self, anchors, nc, img_size, yolo_index, layers, stride):
+        super(YOLOLayer, self).__init__()
+        self.anchors = torch.Tensor(anchors)
+        self.index = yolo_index  # index of this layer in layers
+        self.layers = layers  # model output layer indices
+        self.stride = stride  # layer stride
+        self.nl = len(layers)  # number of output layers (3)
+        self.na = len(anchors)  # number of anchors (3)
+        self.nc = nc  # number of classes (80)
+        self.no = nc + 5  # number of outputs (85)
+        self.nx, self.ny, self.ng = 0, 0, 0  # initialize number of x, y gridpoints
+        self.anchor_vec = self.anchors / self.stride
+        self.anchor_wh = self.anchor_vec.view(1, self.na, 1, 1, 2)
+
+        if ONNX_EXPORT:
+            self.training = False
+            self.create_grids((img_size[1] // stride, img_size[0] // stride))  # number x, y grid points
+
+    def create_grids(self, ng=(13, 13), device='cpu'):
+        self.nx, self.ny = ng  # x and y grid size
+        self.ng = torch.tensor(ng, dtype=torch.float)
+
+        # build xy offsets
+        if not self.training:
+            yv, xv = torch.meshgrid([torch.arange(self.ny, device=device), torch.arange(self.nx, device=device)])
+            self.grid = torch.stack((xv, yv), 2).view((1, 1, self.ny, self.nx, 2)).float()
+
+        if self.anchor_vec.device != device:
+            self.anchor_vec = self.anchor_vec.to(device)
+            self.anchor_wh = self.anchor_wh.to(device)
+
+    def forward(self, p, out):
+        ASFF = False  # https://arxiv.org/abs/1911.09516
+        if ASFF:
+            i, n = self.index, self.nl  # index in layers, number of layers
+            p = out[self.layers[i]]
+            bs, _, ny, nx = p.shape  # bs, 255, 13, 13
+            if (self.nx, self.ny) != (nx, ny):
+                self.create_grids((nx, ny), p.device)
+
+            # outputs and weights
+            # w = F.softmax(p[:, -n:], 1)  # normalized weights
+            w = torch.sigmoid(p[:, -n:]) * (2 / n)  # sigmoid weights (faster)
+            # w = w / w.sum(1).unsqueeze(1)  # normalize across layer dimension
+
+            # weighted ASFF sum
+            p = out[self.layers[i]][:, :-n] * w[:, i:i + 1]
+            for j in range(n):
+                if j != i:
+                    p += w[:, j:j + 1] * \
+                         F.interpolate(out[self.layers[j]][:, :-n], size=[ny, nx], mode='bilinear', align_corners=False)
+
+        elif ONNX_EXPORT:
+            bs = 1  # batch size
+        else:
+            bs, _, ny, nx = p.shape  # bs, 255, 13, 13
+            if (self.nx, self.ny) != (nx, ny):
+                self.create_grids((nx, ny), p.device)
+
+        # p.view(bs, 255, 13, 13) -- > (bs, 3, 13, 13, 85)  # (bs, anchors, grid, grid, classes + xywh)
+        p = p.view(bs, self.na, self.no, self.ny, self.nx).permute(0, 1, 3, 4, 2).contiguous()  # prediction
+
+        if self.training:
+            return p
+
+        elif ONNX_EXPORT:
+            # Avoid broadcasting for ANE operations
+            m = self.na * self.nx * self.ny
+            ng = 1. / self.ng.repeat(m, 1)
+            grid = self.grid.repeat(1, self.na, 1, 1, 1).view(m, 2)
+            anchor_wh = self.anchor_wh.repeat(1, 1, self.nx, self.ny, 1).view(m, 2) * ng
+
+            p = p.view(m, self.no)
+            xy = torch.sigmoid(p[:, 0:2]) + grid  # x, y
+            wh = torch.exp(p[:, 2:4]) * anchor_wh  # width, height
+            p_cls = torch.sigmoid(p[:, 4:5]) if self.nc == 1 else \
+                torch.sigmoid(p[:, 5:self.no]) * torch.sigmoid(p[:, 4:5])  # conf
+            return p_cls, xy * ng, wh
+
+        else:  # inference
+            io = p.sigmoid()
+            io[..., :2] = (io[..., :2] * 2. - 0.5 + self.grid)
+            io[..., 2:4] = (io[..., 2:4] * 2) ** 2 * self.anchor_wh
+            io[..., :4] *= self.stride
+            #io = p.clone()  # inference output
+            #io[..., :2] = torch.sigmoid(io[..., :2]) + self.grid  # xy
+            #io[..., 2:4] = torch.exp(io[..., 2:4]) * self.anchor_wh  # wh yolo method
+            #io[..., :4] *= self.stride
+            #torch.sigmoid_(io[..., 4:])
+            return io.view(bs, -1, self.no), p  # view [1, 3, 13, 13, 85] as [1, 507, 85]
+
+
+class JDELayer(nn.Module):
+    def __init__(self, anchors, nc, img_size, yolo_index, layers, stride):
+        super(JDELayer, self).__init__()
+        self.anchors = torch.Tensor(anchors)
+        self.index = yolo_index  # index of this layer in layers
+        self.layers = layers  # model output layer indices
+        self.stride = stride  # layer stride
+        self.nl = len(layers)  # number of output layers (3)
+        self.na = len(anchors)  # number of anchors (3)
+        self.nc = nc  # number of classes (80)
+        self.no = nc + 5  # number of outputs (85)
+        self.nx, self.ny, self.ng = 0, 0, 0  # initialize number of x, y gridpoints
+        self.anchor_vec = self.anchors / self.stride
+        self.anchor_wh = self.anchor_vec.view(1, self.na, 1, 1, 2)
+
+        if ONNX_EXPORT:
+            self.training = False
+            self.create_grids((img_size[1] // stride, img_size[0] // stride))  # number x, y grid points
+
+    def create_grids(self, ng=(13, 13), device='cpu'):
+        self.nx, self.ny = ng  # x and y grid size
+        self.ng = torch.tensor(ng, dtype=torch.float)
+
+        # build xy offsets
+        if not self.training:
+            yv, xv = torch.meshgrid([torch.arange(self.ny, device=device), torch.arange(self.nx, device=device)])
+            self.grid = torch.stack((xv, yv), 2).view((1, 1, self.ny, self.nx, 2)).float()
+
+        if self.anchor_vec.device != device:
+            self.anchor_vec = self.anchor_vec.to(device)
+            self.anchor_wh = self.anchor_wh.to(device)
+
+    def forward(self, p, out):
+        ASFF = False  # https://arxiv.org/abs/1911.09516
+        if ASFF:
+            i, n = self.index, self.nl  # index in layers, number of layers
+            p = out[self.layers[i]]
+            bs, _, ny, nx = p.shape  # bs, 255, 13, 13
+            if (self.nx, self.ny) != (nx, ny):
+                self.create_grids((nx, ny), p.device)
+
+            # outputs and weights
+            # w = F.softmax(p[:, -n:], 1)  # normalized weights
+            w = torch.sigmoid(p[:, -n:]) * (2 / n)  # sigmoid weights (faster)
+            # w = w / w.sum(1).unsqueeze(1)  # normalize across layer dimension
+
+            # weighted ASFF sum
+            p = out[self.layers[i]][:, :-n] * w[:, i:i + 1]
+            for j in range(n):
+                if j != i:
+                    p += w[:, j:j + 1] * \
+                         F.interpolate(out[self.layers[j]][:, :-n], size=[ny, nx], mode='bilinear', align_corners=False)
+
+        elif ONNX_EXPORT:
+            bs = 1  # batch size
+        else:
+            bs, _, ny, nx = p.shape  # bs, 255, 13, 13
+            if (self.nx, self.ny) != (nx, ny):
+                self.create_grids((nx, ny), p.device)
+
+        # p.view(bs, 255, 13, 13) -- > (bs, 3, 13, 13, 85)  # (bs, anchors, grid, grid, classes + xywh)
+        p = p.view(bs, self.na, self.no, self.ny, self.nx).permute(0, 1, 3, 4, 2).contiguous()  # prediction
+
+        if self.training:
+            return p
+
+        elif ONNX_EXPORT:
+            # Avoid broadcasting for ANE operations
+            m = self.na * self.nx * self.ny
+            ng = 1. / self.ng.repeat(m, 1)
+            grid = self.grid.repeat(1, self.na, 1, 1, 1).view(m, 2)
+            anchor_wh = self.anchor_wh.repeat(1, 1, self.nx, self.ny, 1).view(m, 2) * ng
+
+            p = p.view(m, self.no)
+            xy = torch.sigmoid(p[:, 0:2]) + grid  # x, y
+            wh = torch.exp(p[:, 2:4]) * anchor_wh  # width, height
+            p_cls = torch.sigmoid(p[:, 4:5]) if self.nc == 1 else \
+                torch.sigmoid(p[:, 5:self.no]) * torch.sigmoid(p[:, 4:5])  # conf
+            return p_cls, xy * ng, wh
+
+        else:  # inference
+            #io = p.sigmoid()
+            #io[..., :2] = (io[..., :2] * 2. - 0.5 + self.grid)
+            #io[..., 2:4] = (io[..., 2:4] * 2) ** 2 * self.anchor_wh
+            #io[..., :4] *= self.stride
+            io = p.clone()  # inference output
+            io[..., :2] = torch.sigmoid(io[..., :2]) * 2. - 0.5 + self.grid  # xy
+            io[..., 2:4] = (torch.sigmoid(io[..., 2:4]) * 2) ** 2 * self.anchor_wh  # wh yolo method
+            io[..., :4] *= self.stride
+            io[..., 4:] = F.softmax(io[..., 4:])
+            return io.view(bs, -1, self.no), p  # view [1, 3, 13, 13, 85] as [1, 507, 85]
+
+class Darknet(nn.Module):
+    # YOLOv3 object detection model
+
+    def __init__(self, cfg, img_size=(416, 416), verbose=False):
+        super(Darknet, self).__init__()
+
+        self.module_defs = parse_model_cfg(cfg)
+        self.module_list, self.routs = create_modules(self.module_defs, img_size, cfg)
+        self.yolo_layers = get_yolo_layers(self)
+        # torch_utils.initialize_weights(self)
+
+        # Darknet Header https://github.com/AlexeyAB/darknet/issues/2914#issuecomment-496675346
+        self.version = np.array([0, 2, 5], dtype=np.int32)  # (int32) version info: major, minor, revision
+        self.seen = np.array([0], dtype=np.int64)  # (int64) number of images seen during training
+        self.info(verbose) if not ONNX_EXPORT else None  # print model description
+
+    def forward(self, x, augment=False, verbose=False):
+
+        if not augment:
+            return self.forward_once(x)
+        else:  # Augment images (inference and test only) https://github.com/ultralytics/yolov3/issues/931
+            img_size = x.shape[-2:]  # height, width
+            s = [0.83, 0.67]  # scales
+            y = []
+            for i, xi in enumerate((x,
+                                    torch_utils.scale_img(x.flip(3), s[0], same_shape=False),  # flip-lr and scale
+                                    torch_utils.scale_img(x, s[1], same_shape=False),  # scale
+                                    )):
+                # cv2.imwrite('img%g.jpg' % i, 255 * xi[0].numpy().transpose((1, 2, 0))[:, :, ::-1])
+                y.append(self.forward_once(xi)[0])
+
+            y[1][..., :4] /= s[0]  # scale
+            y[1][..., 0] = img_size[1] - y[1][..., 0]  # flip lr
+            y[2][..., :4] /= s[1]  # scale
+
+            # for i, yi in enumerate(y):  # coco small, medium, large = < 32**2 < 96**2 <
+            #     area = yi[..., 2:4].prod(2)[:, :, None]
+            #     if i == 1:
+            #         yi *= (area < 96. ** 2).float()
+            #     elif i == 2:
+            #         yi *= (area > 32. ** 2).float()
+            #     y[i] = yi
+
+            y = torch.cat(y, 1)
+            return y, None
+
+    def forward_once(self, x, augment=False, verbose=False):
+        img_size = x.shape[-2:]  # height, width
+        yolo_out, out = [], []
+        if verbose:
+            print('0', x.shape)
+            str = ''
+
+        # Augment images (inference and test only)
+        if augment:  # https://github.com/ultralytics/yolov3/issues/931
+            nb = x.shape[0]  # batch size
+            s = [0.83, 0.67]  # scales
+            x = torch.cat((x,
+                           torch_utils.scale_img(x.flip(3), s[0]),  # flip-lr and scale
+                           torch_utils.scale_img(x, s[1]),  # scale
+                           ), 0)
+
+        for i, module in enumerate(self.module_list):
+            name = module.__class__.__name__
+            #print(name)
+            if name in ['WeightedFeatureFusion', 'FeatureConcat', 'FeatureConcat2', 'FeatureConcat3', 'FeatureConcat_l', 'ScaleChannel', 'ShiftChannel', 'ShiftChannel2D', 'ControlChannel', 'ControlChannel2D', 'AlternateChannel', 'AlternateChannel2D', 'SelectChannel', 'SelectChannel2D', 'ScaleSpatial']:  # sum, concat
+                if verbose:
+                    l = [i - 1] + module.layers  # layers
+                    sh = [list(x.shape)] + [list(out[i].shape) for i in module.layers]  # shapes
+                    str = ' >> ' + ' + '.join(['layer %g %s' % x for x in zip(l, sh)])
+                x = module(x, out)  # WeightedFeatureFusion(), FeatureConcat()
+            elif name in ['ImplicitA', 'ImplicitM', 'ImplicitC', 'Implicit2DA', 'Implicit2DM', 'Implicit2DC']:
+                x = module()
+            elif name == 'YOLOLayer':
+                yolo_out.append(module(x, out))
+            elif name == 'JDELayer':
+                yolo_out.append(module(x, out))
+            else:  # run module directly, i.e. mtype = 'convolutional', 'upsample', 'maxpool', 'batchnorm2d' etc.
+                #print(module)
+                #print(x.shape)
+                x = module(x)
+
+            out.append(x if self.routs[i] else [])
+            if verbose:
+                print('%g/%g %s -' % (i, len(self.module_list), name), list(x.shape), str)
+                str = ''
+
+        if self.training:  # train
+            return yolo_out
+        elif ONNX_EXPORT:  # export
+            x = [torch.cat(x, 0) for x in zip(*yolo_out)]
+            return x[0], torch.cat(x[1:3], 1)  # scores, boxes: 3780x80, 3780x4
+        else:  # inference or test
+            x, p = zip(*yolo_out)  # inference output, training output
+            x = torch.cat(x, 1)  # cat yolo outputs
+            if augment:  # de-augment results
+                x = torch.split(x, nb, dim=0)
+                x[1][..., :4] /= s[0]  # scale
+                x[1][..., 0] = img_size[1] - x[1][..., 0]  # flip lr
+                x[2][..., :4] /= s[1]  # scale
+                x = torch.cat(x, 1)
+            return x, p
+
+    def fuse(self):
+        # Fuse Conv2d + BatchNorm2d layers throughout model
+        print('Fusing layers...')
+        fused_list = nn.ModuleList()
+        for a in list(self.children())[0]:
+            if isinstance(a, nn.Sequential):
+                for i, b in enumerate(a):
+                    if isinstance(b, nn.modules.batchnorm.BatchNorm2d):
+                        # fuse this bn layer with the previous conv2d layer
+                        conv = a[i - 1]
+                        fused = torch_utils.fuse_conv_and_bn(conv, b)
+                        a = nn.Sequential(fused, *list(a.children())[i + 1:])
+                        break
+            fused_list.append(a)
+        self.module_list = fused_list
+        self.info() if not ONNX_EXPORT else None  # yolov3-spp reduced from 225 to 152 layers
+
+    def info(self, verbose=False):
+        torch_utils.model_info(self, verbose)
+
+
+def get_yolo_layers(model):
+    return [i for i, m in enumerate(model.module_list) if m.__class__.__name__ in ['YOLOLayer', 'JDELayer']]  # [89, 101, 113]
+
+
+def load_darknet_weights(self, weights, cutoff=-1):
+    # Parses and loads the weights stored in 'weights'
+
+    # Establish cutoffs (load layers between 0 and cutoff. if cutoff = -1 all are loaded)
+    file = Path(weights).name
+    if file == 'darknet53.conv.74':
+        cutoff = 75
+    elif file == 'yolov3-tiny.conv.15':
+        cutoff = 15
+
+    # Read weights file
+    with open(weights, 'rb') as f:
+        # Read Header https://github.com/AlexeyAB/darknet/issues/2914#issuecomment-496675346
+        self.version = np.fromfile(f, dtype=np.int32, count=3)  # (int32) version info: major, minor, revision
+        self.seen = np.fromfile(f, dtype=np.int64, count=1)  # (int64) number of images seen during training
+
+        weights = np.fromfile(f, dtype=np.float32)  # the rest are weights
+
+    ptr = 0
+    for i, (mdef, module) in enumerate(zip(self.module_defs[:cutoff], self.module_list[:cutoff])):
+        if mdef['type'] == 'convolutional':
+            conv = module[0]
+            if mdef['batch_normalize']:
+                # Load BN bias, weights, running mean and running variance
+                bn = module[1]
+                nb = bn.bias.numel()  # number of biases
+                # Bias
+                bn.bias.data.copy_(torch.from_numpy(weights[ptr:ptr + nb]).view_as(bn.bias))
+                ptr += nb
+                # Weight
+                bn.weight.data.copy_(torch.from_numpy(weights[ptr:ptr + nb]).view_as(bn.weight))
+                ptr += nb
+                # Running Mean
+                bn.running_mean.data.copy_(torch.from_numpy(weights[ptr:ptr + nb]).view_as(bn.running_mean))
+                ptr += nb
+                # Running Var
+                bn.running_var.data.copy_(torch.from_numpy(weights[ptr:ptr + nb]).view_as(bn.running_var))
+                ptr += nb
+            else:
+                # Load conv. bias
+                nb = conv.bias.numel()
+                conv_b = torch.from_numpy(weights[ptr:ptr + nb]).view_as(conv.bias)
+                conv.bias.data.copy_(conv_b)
+                ptr += nb
+            # Load conv. weights
+            nw = conv.weight.numel()  # number of weights
+            conv.weight.data.copy_(torch.from_numpy(weights[ptr:ptr + nw]).view_as(conv.weight))
+            ptr += nw
+
+
+def save_weights(self, path='model.weights', cutoff=-1):
+    # Converts a PyTorch model to Darket format (*.pt to *.weights)
+    # Note: Does not work if model.fuse() is applied
+    with open(path, 'wb') as f:
+        # Write Header https://github.com/AlexeyAB/darknet/issues/2914#issuecomment-496675346
+        self.version.tofile(f)  # (int32) version info: major, minor, revision
+        self.seen.tofile(f)  # (int64) number of images seen during training
+
+        # Iterate through layers
+        for i, (mdef, module) in enumerate(zip(self.module_defs[:cutoff], self.module_list[:cutoff])):
+            if mdef['type'] == 'convolutional':
+                conv_layer = module[0]
+                # If batch norm, load bn first
+                if mdef['batch_normalize']:
+                    bn_layer = module[1]
+                    bn_layer.bias.data.cpu().numpy().tofile(f)
+                    bn_layer.weight.data.cpu().numpy().tofile(f)
+                    bn_layer.running_mean.data.cpu().numpy().tofile(f)
+                    bn_layer.running_var.data.cpu().numpy().tofile(f)
+                # Load conv bias
+                else:
+                    conv_layer.bias.data.cpu().numpy().tofile(f)
+                # Load conv weights
+                conv_layer.weight.data.cpu().numpy().tofile(f)
+
+
+def convert(cfg='cfg/yolov3-spp.cfg', weights='weights/yolov3-spp.weights', saveto='converted.weights'):
+    # Converts between PyTorch and Darknet format per extension (i.e. *.weights convert to *.pt and vice versa)
+    # from models import *; convert('cfg/yolov3-spp.cfg', 'weights/yolov3-spp.weights')
+
+    # Initialize model
+    model = Darknet(cfg)
+    ckpt = torch.load(weights)  # load checkpoint
+    try:
+        ckpt['model'] = {k: v for k, v in ckpt['model'].items() if model.state_dict()[k].numel() == v.numel()}
+        model.load_state_dict(ckpt['model'], strict=False)
+        save_weights(model, path=saveto, cutoff=-1)
+    except KeyError as e:
+        print(e)
+
+def attempt_download(weights):
+    # Attempt to download pretrained weights if not found locally
+    weights = weights.strip()
+    msg = weights + ' missing, try downloading from https://drive.google.com/open?id=1LezFG5g3BCW6iYaV89B2i64cqEUZD7e0'
+
+    if len(weights) > 0 and not os.path.isfile(weights):
+        d = {''}
+
+        file = Path(weights).name
+        if file in d:
+            r = gdrive_download(id=d[file], name=weights)
+        else:  # download from pjreddie.com
+            url = 'https://pjreddie.com/media/files/' + file
+            print('Downloading ' + url)
+            r = os.system('curl -f ' + url + ' -o ' + weights)
+
+        # Error check
+        if not (r == 0 and os.path.exists(weights) and os.path.getsize(weights) > 1E6):  # weights exist and > 1MB
+            os.system('rm ' + weights)  # remove partial downloads
+            raise Exception(msg)

requirements.txt

@@ -0,0 +1,33 @@
+# pip install -qr requirements.txt
+
+# base ----------------------------------------
+Cython
+matplotlib>=3.2.2
+numpy>=1.18.5
+opencv-python>=4.1.2
+Pillow
+PyYAML>=5.3.1
+scipy>=1.4.1
+tensorboard>=1.5
+torch==1.7.0
+torchvision==0.8.1
+tqdm>=4.41.0
+
+# logging -------------------------------------
+# wandb
+
+# plotting ------------------------------------
+seaborn>=0.11.0
+pandas
+
+# export --------------------------------------
+# coremltools>=4.1
+# onnx>=1.8.1
+# scikit-learn==0.19.2  # for coreml quantization
+
+# extras --------------------------------------
+thop  # FLOPS computation
+pycocotools==2.0  # COCO mAP
+
+
+gdown

scripts/get_coco.sh

@@ -0,0 +1,27 @@
+#!/bin/bash
+# Script credit to https://github.com/ultralytics/yolov5
+# COCO 2017 dataset http://cocodataset.org
+# Download command: bash scripts/get_coco.sh
+# Default dataset location is next to /yolor:
+#   /parent_folder
+#     /coco
+#     /yolor
+
+# Download/unzip labels
+d='../' # unzip directory
+url=https://github.com/ultralytics/yolov5/releases/download/v1.0/
+f='coco2017labels.zip' # or 'coco2017labels-segments.zip', 68 MB
+echo 'Downloading' $url$f ' ...'
+curl -L $url$f -o $f && unzip -q $f -d $d && rm $f & # download, unzip, remove in background
+
+# Download/unzip images
+d='../coco/images' # unzip directory
+url=http://images.cocodataset.org/zips/
+f1='train2017.zip' # 19G, 118k images
+f2='val2017.zip'   # 1G, 5k images
+f3='test2017.zip'  # 7G, 41k images (optional)
+for f in $f1 $f2 $f3; do
+  echo 'Downloading' $url$f '...'
+  curl -L $url$f -o $f && unzip -q $f -d $d && rm $f & # download, unzip, remove in background
+done
+wait # finish background tasks

scripts/get_pretrain.sh

@@ -0,0 +1,7 @@
+curl -c ./cookie -s -L "https://drive.google.com/uc?export=download&id=1Tdn3yqpZ79X7R1Ql0zNlNScB1Dv9Fp76" > /dev/null
+curl -Lb ./cookie "https://drive.google.com/uc?export=download&confirm=`awk '/download/ {print $NF}' ./cookie`&id=1Tdn3yqpZ79X7R1Ql0zNlNScB1Dv9Fp76" -o yolor_p6.pt
+rm ./cookie
+
+curl -c ./cookie -s -L "https://drive.google.com/uc?export=download&id=1UflcHlN5ERPdhahMivQYCbWWw7d2wY7U" > /dev/null
+curl -Lb ./cookie "https://drive.google.com/uc?export=download&confirm=`awk '/download/ {print $NF}' ./cookie`&id=1UflcHlN5ERPdhahMivQYCbWWw7d2wY7U" -o yolor_w6.pt
+rm ./cookie

test.py

@@ -0,0 +1,344 @@
+import argparse
+import glob
+import json
+import os
+from pathlib import Path
+
+import numpy as np
+import torch
+import yaml
+from tqdm import tqdm
+
+from utils.google_utils import attempt_load
+from utils.datasets import create_dataloader
+from utils.general import coco80_to_coco91_class, check_dataset, check_file, check_img_size, box_iou, \
+    non_max_suppression, scale_coords, xyxy2xywh, xywh2xyxy, clip_coords, set_logging, increment_path
+from utils.loss import compute_loss
+from utils.metrics import ap_per_class
+from utils.plots import plot_images, output_to_target
+from utils.torch_utils import select_device, time_synchronized
+
+from models.models import *
+
+def load_classes(path):
+    # Loads *.names file at 'path'
+    with open(path, 'r') as f:
+        names = f.read().split('\n')
+    return list(filter(None, names))  # filter removes empty strings (such as last line)
+
+
+def test(data,
+         weights=None,
+         batch_size=16,
+         imgsz=640,
+         conf_thres=0.001,
+         iou_thres=0.6,  # for NMS
+         save_json=False,
+         single_cls=False,
+         augment=False,
+         verbose=False,
+         model=None,
+         dataloader=None,
+         save_dir=Path(''),  # for saving images
+         save_txt=False,  # for auto-labelling
+         save_conf=False,
+         plots=True,
+         log_imgs=0):  # number of logged images
+
+    # Initialize/load model and set device
+    training = model is not None
+    if training:  # called by train.py
+        device = next(model.parameters()).device  # get model device
+
+    else:  # called directly
+        set_logging()
+        device = select_device(opt.device, batch_size=batch_size)
+        save_txt = opt.save_txt  # save *.txt labels
+
+        # Directories
+        save_dir = Path(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok))  # increment run
+        (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True)  # make dir
+
+        # Load model
+        model = Darknet(opt.cfg).to(device)
+
+        # load model
+        try:
+            ckpt = torch.load(weights[0], map_location=device)  # load checkpoint
+            ckpt['model'] = {k: v for k, v in ckpt['model'].items() if model.state_dict()[k].numel() == v.numel()}
+            model.load_state_dict(ckpt['model'], strict=False)
+        except:
+            load_darknet_weights(model, weights[0])
+        imgsz = check_img_size(imgsz, s=64)  # check img_size
+
+    # Half
+    half = device.type != 'cpu'  # half precision only supported on CUDA
+    if half:
+        model.half()
+
+    # Configure
+    model.eval()
+    is_coco = data.endswith('coco.yaml')  # is COCO dataset
+    with open(data) as f:
+        data = yaml.load(f, Loader=yaml.FullLoader)  # model dict
+    check_dataset(data)  # check
+    nc = 1 if single_cls else int(data['nc'])  # number of classes
+    iouv = torch.linspace(0.5, 0.95, 10).to(device)  # iou vector for mAP@0.5:0.95
+    niou = iouv.numel()
+
+    # Logging
+    log_imgs, wandb = min(log_imgs, 100), None  # ceil
+    try:
+        import wandb  # Weights & Biases
+    except ImportError:
+        log_imgs = 0
+
+    # Dataloader
+    if not training:
+        img = torch.zeros((1, 3, imgsz, imgsz), device=device)  # init img
+        _ = model(img.half() if half else img) if device.type != 'cpu' else None  # run once
+        path = data['test'] if opt.task == 'test' else data['val']  # path to val/test images
+        dataloader = create_dataloader(path, imgsz, batch_size, 64, opt, pad=0.5, rect=True)[0]
+
+    seen = 0
+    try:
+        names = model.names if hasattr(model, 'names') else model.module.names
+    except:
+        names = load_classes(opt.names)
+    coco91class = coco80_to_coco91_class()
+    s = ('%20s' + '%12s' * 6) % ('Class', 'Images', 'Targets', 'P', 'R', 'mAP@.5', 'mAP@.5:.95')
+    p, r, f1, mp, mr, map50, map, t0, t1 = 0., 0., 0., 0., 0., 0., 0., 0., 0.
+    loss = torch.zeros(3, device=device)
+    jdict, stats, ap, ap_class, wandb_images = [], [], [], [], []
+    for batch_i, (img, targets, paths, shapes) in enumerate(tqdm(dataloader, desc=s)):
+        img = img.to(device, non_blocking=True)
+        img = img.half() if half else img.float()  # uint8 to fp16/32
+        img /= 255.0  # 0 - 255 to 0.0 - 1.0
+        targets = targets.to(device)
+        nb, _, height, width = img.shape  # batch size, channels, height, width
+        whwh = torch.Tensor([width, height, width, height]).to(device)
+
+        # Disable gradients
+        with torch.no_grad():
+            # Run model
+            t = time_synchronized()
+            inf_out, train_out = model(img, augment=augment)  # inference and training outputs
+            t0 += time_synchronized() - t
+
+            # Compute loss
+            if training:  # if model has loss hyperparameters
+                loss += compute_loss([x.float() for x in train_out], targets, model)[1][:3]  # box, obj, cls
+
+            # Run NMS
+            t = time_synchronized()
+            output = non_max_suppression(inf_out, conf_thres=conf_thres, iou_thres=iou_thres)
+            t1 += time_synchronized() - t
+
+        # Statistics per image
+        for si, pred in enumerate(output):
+            labels = targets[targets[:, 0] == si, 1:]
+            nl = len(labels)
+            tcls = labels[:, 0].tolist() if nl else []  # target class
+            seen += 1
+
+            if len(pred) == 0:
+                if nl:
+                    stats.append((torch.zeros(0, niou, dtype=torch.bool), torch.Tensor(), torch.Tensor(), tcls))
+                continue
+
+            # Append to text file
+            path = Path(paths[si])
+            if save_txt:
+                gn = torch.tensor(shapes[si][0])[[1, 0, 1, 0]]  # normalization gain whwh
+                x = pred.clone()
+                x[:, :4] = scale_coords(img[si].shape[1:], x[:, :4], shapes[si][0], shapes[si][1])  # to original
+                for *xyxy, conf, cls in x:
+                    xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist()  # normalized xywh
+                    line = (cls, *xywh, conf) if save_conf else (cls, *xywh)  # label format
+                    with open(save_dir / 'labels' / (path.stem + '.txt'), 'a') as f:
+                        f.write(('%g ' * len(line)).rstrip() % line + '\n')
+
+            # W&B logging
+            if plots and len(wandb_images) < log_imgs:
+                box_data = [{"position": {"minX": xyxy[0], "minY": xyxy[1], "maxX": xyxy[2], "maxY": xyxy[3]},
+                             "class_id": int(cls),
+                             "box_caption": "%s %.3f" % (names[cls], conf),
+                             "scores": {"class_score": conf},
+                             "domain": "pixel"} for *xyxy, conf, cls in pred.tolist()]
+                boxes = {"predictions": {"box_data": box_data, "class_labels": names}}
+                wandb_images.append(wandb.Image(img[si], boxes=boxes, caption=path.name))
+
+            # Clip boxes to image bounds
+            clip_coords(pred, (height, width))
+
+            # Append to pycocotools JSON dictionary
+            if save_json:
+                # [{"image_id": 42, "category_id": 18, "bbox": [258.15, 41.29, 348.26, 243.78], "score": 0.236}, ...
+                image_id = int(path.stem) if path.stem.isnumeric() else path.stem
+                box = pred[:, :4].clone()  # xyxy
+                scale_coords(img[si].shape[1:], box, shapes[si][0], shapes[si][1])  # to original shape
+                box = xyxy2xywh(box)  # xywh
+                box[:, :2] -= box[:, 2:] / 2  # xy center to top-left corner
+                for p, b in zip(pred.tolist(), box.tolist()):
+                    jdict.append({'image_id': image_id,
+                                  'category_id': coco91class[int(p[5])] if is_coco else int(p[5]),
+                                  'bbox': [round(x, 3) for x in b],
+                                  'score': round(p[4], 5)})
+
+            # Assign all predictions as incorrect
+            correct = torch.zeros(pred.shape[0], niou, dtype=torch.bool, device=device)
+            if nl:
+                detected = []  # target indices
+                tcls_tensor = labels[:, 0]
+
+                # target boxes
+                tbox = xywh2xyxy(labels[:, 1:5]) * whwh
+
+                # Per target class
+                for cls in torch.unique(tcls_tensor):
+                    ti = (cls == tcls_tensor).nonzero(as_tuple=False).view(-1)  # prediction indices
+                    pi = (cls == pred[:, 5]).nonzero(as_tuple=False).view(-1)  # target indices
+
+                    # Search for detections
+                    if pi.shape[0]:
+                        # Prediction to target ious
+                        ious, i = box_iou(pred[pi, :4], tbox[ti]).max(1)  # best ious, indices
+
+                        # Append detections
+                        detected_set = set()
+                        for j in (ious > iouv[0]).nonzero(as_tuple=False):
+                            d = ti[i[j]]  # detected target
+                            if d.item() not in detected_set:
+                                detected_set.add(d.item())
+                                detected.append(d)
+                                correct[pi[j]] = ious[j] > iouv  # iou_thres is 1xn
+                                if len(detected) == nl:  # all targets already located in image
+                                    break
+
+            # Append statistics (correct, conf, pcls, tcls)
+            stats.append((correct.cpu(), pred[:, 4].cpu(), pred[:, 5].cpu(), tcls))
+
+        # Plot images
+        if plots and batch_i < 3:
+            f = save_dir / f'test_batch{batch_i}_labels.jpg'  # filename
+            plot_images(img, targets, paths, f, names)  # labels
+            f = save_dir / f'test_batch{batch_i}_pred.jpg'
+            plot_images(img, output_to_target(output, width, height), paths, f, names)  # predictions
+
+    # Compute statistics
+    stats = [np.concatenate(x, 0) for x in zip(*stats)]  # to numpy
+    if len(stats) and stats[0].any():
+        p, r, ap, f1, ap_class = ap_per_class(*stats, plot=plots, fname=save_dir / 'precision-recall_curve.png')
+        p, r, ap50, ap = p[:, 0], r[:, 0], ap[:, 0], ap.mean(1)  # [P, R, AP@0.5, AP@0.5:0.95]
+        mp, mr, map50, map = p.mean(), r.mean(), ap50.mean(), ap.mean()
+        nt = np.bincount(stats[3].astype(np.int64), minlength=nc)  # number of targets per class
+    else:
+        nt = torch.zeros(1)
+
+    # W&B logging
+    if plots and wandb:
+        wandb.log({"Images": wandb_images})
+        wandb.log({"Validation": [wandb.Image(str(x), caption=x.name) for x in sorted(save_dir.glob('test*.jpg'))]})
+
+    # Print results
+    pf = '%20s' + '%12.3g' * 6  # print format
+    print(pf % ('all', seen, nt.sum(), mp, mr, map50, map))
+
+    # Print results per class
+    if verbose and nc > 1 and len(stats):
+        for i, c in enumerate(ap_class):
+            print(pf % (names[c], seen, nt[c], p[i], r[i], ap50[i], ap[i]))
+
+    # Print speeds
+    t = tuple(x / seen * 1E3 for x in (t0, t1, t0 + t1)) + (imgsz, imgsz, batch_size)  # tuple
+    if not training:
+        print('Speed: %.1f/%.1f/%.1f ms inference/NMS/total per %gx%g image at batch-size %g' % t)
+
+    # Save JSON
+    if save_json and len(jdict):
+        w = Path(weights[0] if isinstance(weights, list) else weights).stem if weights is not None else ''  # weights
+        anno_json = glob.glob('../coco/annotations/instances_val*.json')[0]  # annotations json
+        pred_json = str(save_dir / f"{w}_predictions.json")  # predictions json
+        print('\nEvaluating pycocotools mAP... saving %s...' % pred_json)
+        with open(pred_json, 'w') as f:
+            json.dump(jdict, f)
+
+        try:  # https://github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocoEvalDemo.ipynb
+            from pycocotools.coco import COCO
+            from pycocotools.cocoeval import COCOeval
+
+            anno = COCO(anno_json)  # init annotations api
+            pred = anno.loadRes(pred_json)  # init predictions api
+            eval = COCOeval(anno, pred, 'bbox')
+            if is_coco:
+                eval.params.imgIds = [int(Path(x).stem) for x in dataloader.dataset.img_files]  # image IDs to evaluate
+            eval.evaluate()
+            eval.accumulate()
+            eval.summarize()
+            map, map50 = eval.stats[:2]  # update results (mAP@0.5:0.95, mAP@0.5)
+        except Exception as e:
+            print('ERROR: pycocotools unable to run: %s' % e)
+
+    # Return results
+    if not training:
+        print('Results saved to %s' % save_dir)
+    model.float()  # for training
+    maps = np.zeros(nc) + map
+    for i, c in enumerate(ap_class):
+        maps[c] = ap[i]
+    return (mp, mr, map50, map, *(loss.cpu() / len(dataloader)).tolist()), maps, t
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(prog='test.py')
+    parser.add_argument('--weights', nargs='+', type=str, default='yolor_p6.pt', help='model.pt path(s)')
+    parser.add_argument('--data', type=str, default='data/coco.yaml', help='*.data path')
+    parser.add_argument('--batch-size', type=int, default=32, help='size of each image batch')
+    parser.add_argument('--img-size', type=int, default=1280, help='inference size (pixels)')
+    parser.add_argument('--conf-thres', type=float, default=0.001, help='object confidence threshold')
+    parser.add_argument('--iou-thres', type=float, default=0.65, help='IOU threshold for NMS')
+    parser.add_argument('--task', default='val', help="'val', 'test', 'study'")
+    parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+    parser.add_argument('--single-cls', action='store_true', help='treat as single-class dataset')
+    parser.add_argument('--augment', action='store_true', help='augmented inference')
+    parser.add_argument('--verbose', action='store_true', help='report mAP by class')
+    parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
+    parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
+    parser.add_argument('--save-json', action='store_true', help='save a cocoapi-compatible JSON results file')
+    parser.add_argument('--project', default='runs/test', help='save to project/name')
+    parser.add_argument('--name', default='exp', help='save to project/name')
+    parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+    parser.add_argument('--cfg', type=str, default='cfg/yolor_p6.cfg', help='*.cfg path')
+    parser.add_argument('--names', type=str, default='data/coco.names', help='*.cfg path')
+    opt = parser.parse_args()
+    opt.save_json |= opt.data.endswith('coco.yaml')
+    opt.data = check_file(opt.data)  # check file
+    print(opt)
+
+    if opt.task in ['val', 'test']:  # run normally
+        test(opt.data,
+             opt.weights,
+             opt.batch_size,
+             opt.img_size,
+             opt.conf_thres,
+             opt.iou_thres,
+             opt.save_json,
+             opt.single_cls,
+             opt.augment,
+             opt.verbose,
+             save_txt=opt.save_txt,
+             save_conf=opt.save_conf,
+             )
+
+    elif opt.task == 'study':  # run over a range of settings and save/plot
+        for weights in ['yolor_p6.pt', 'yolor_w6.pt']:
+            f = 'study_%s_%s.txt' % (Path(opt.data).stem, Path(weights).stem)  # filename to save to
+            x = list(range(320, 800, 64))  # x axis
+            y = []  # y axis
+            for i in x:  # img-size
+                print('\nRunning %s point %s...' % (f, i))
+                r, _, t = test(opt.data, weights, opt.batch_size, i, opt.conf_thres, opt.iou_thres, opt.save_json)
+                y.append(r + t)  # results and times
+            np.savetxt(f, y, fmt='%10.4g')  # save
+        os.system('zip -r study.zip study_*.txt')
+        # utils.general.plot_study_txt(f, x)  # plot

train.py

@@ -0,0 +1,619 @@
+import argparse
+import logging
+import math
+import os
+import random
+import time
+from pathlib import Path
+from warnings import warn
+
+import numpy as np
+import torch.distributed as dist
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import torch.optim.lr_scheduler as lr_scheduler
+import torch.utils.data
+import yaml
+from torch.cuda import amp
+from torch.nn.parallel import DistributedDataParallel as DDP
+from torch.utils.tensorboard import SummaryWriter
+from tqdm import tqdm
+
+import test  # import test.py to get mAP after each epoch
+#from models.yolo import Model
+from models.models import *
+from utils.autoanchor import check_anchors
+from utils.datasets import create_dataloader
+from utils.general import labels_to_class_weights, increment_path, labels_to_image_weights, init_seeds, \
+    fitness, fitness_p, fitness_r, fitness_ap50, fitness_ap, fitness_f, strip_optimizer, get_latest_run,\
+    check_dataset, check_file, check_git_status, check_img_size, print_mutation, set_logging
+from utils.google_utils import attempt_download
+from utils.loss import compute_loss
+from utils.plots import plot_images, plot_labels, plot_results, plot_evolution
+from utils.torch_utils import ModelEMA, select_device, intersect_dicts, torch_distributed_zero_first
+
+logger = logging.getLogger(__name__)
+
+try:
+    import wandb
+except ImportError:
+    wandb = None
+    logger.info("Install Weights & Biases for experiment logging via 'pip install wandb' (recommended)")
+
+def train(hyp, opt, device, tb_writer=None, wandb=None):
+    logger.info(f'Hyperparameters {hyp}')
+    save_dir, epochs, batch_size, total_batch_size, weights, rank = \
+        Path(opt.save_dir), opt.epochs, opt.batch_size, opt.total_batch_size, opt.weights, opt.global_rank
+
+    # Directories
+    wdir = save_dir / 'weights'
+    wdir.mkdir(parents=True, exist_ok=True)  # make dir
+    last = wdir / 'last.pt'
+    best = wdir / 'best.pt'
+    results_file = save_dir / 'results.txt'
+
+    # Save run settings
+    with open(save_dir / 'hyp.yaml', 'w') as f:
+        yaml.dump(hyp, f, sort_keys=False)
+    with open(save_dir / 'opt.yaml', 'w') as f:
+        yaml.dump(vars(opt), f, sort_keys=False)
+
+    # Configure
+    plots = not opt.evolve  # create plots
+    cuda = device.type != 'cpu'
+    init_seeds(2 + rank)
+    with open(opt.data) as f:
+        data_dict = yaml.load(f, Loader=yaml.FullLoader)  # data dict
+    with torch_distributed_zero_first(rank):
+        check_dataset(data_dict)  # check
+    train_path = data_dict['train']
+    test_path = data_dict['val']
+    nc, names = (1, ['item']) if opt.single_cls else (int(data_dict['nc']), data_dict['names'])  # number classes, names
+    assert len(names) == nc, '%g names found for nc=%g dataset in %s' % (len(names), nc, opt.data)  # check
+
+    # Model
+    pretrained = weights.endswith('.pt')
+    if pretrained:
+        with torch_distributed_zero_first(rank):
+            attempt_download(weights)  # download if not found locally
+        ckpt = torch.load(weights, map_location=device)  # load checkpoint
+        model = Darknet(opt.cfg).to(device)  # create
+        state_dict = {k: v for k, v in ckpt['model'].items() if model.state_dict()[k].numel() == v.numel()}
+        model.load_state_dict(state_dict, strict=False)
+        print('Transferred %g/%g items from %s' % (len(state_dict), len(model.state_dict()), weights))  # report
+    else:
+        model = Darknet(opt.cfg).to(device) # create
+
+    # Optimizer
+    nbs = 64  # nominal batch size
+    accumulate = max(round(nbs / total_batch_size), 1)  # accumulate loss before optimizing
+    hyp['weight_decay'] *= total_batch_size * accumulate / nbs  # scale weight_decay
+
+    pg0, pg1, pg2 = [], [], []  # optimizer parameter groups
+    for k, v in dict(model.named_parameters()).items():
+        if '.bias' in k:
+            pg2.append(v)  # biases
+        elif 'Conv2d.weight' in k:
+            pg1.append(v)  # apply weight_decay
+        elif 'm.weight' in k:
+            pg1.append(v)  # apply weight_decay
+        elif 'w.weight' in k:
+            pg1.append(v)  # apply weight_decay
+        else:
+            pg0.append(v)  # all else
+
+    if opt.adam:
+        optimizer = optim.Adam(pg0, lr=hyp['lr0'], betas=(hyp['momentum'], 0.999))  # adjust beta1 to momentum
+    else:
+        optimizer = optim.SGD(pg0, lr=hyp['lr0'], momentum=hyp['momentum'], nesterov=True)
+
+    optimizer.add_param_group({'params': pg1, 'weight_decay': hyp['weight_decay']})  # add pg1 with weight_decay
+    optimizer.add_param_group({'params': pg2})  # add pg2 (biases)
+    logger.info('Optimizer groups: %g .bias, %g conv.weight, %g other' % (len(pg2), len(pg1), len(pg0)))
+    del pg0, pg1, pg2
+
+    # Scheduler https://arxiv.org/pdf/1812.01187.pdf
+    # https://pytorch.org/docs/stable/_modules/torch/optim/lr_scheduler.html#OneCycleLR
+    lf = lambda x: ((1 + math.cos(x * math.pi / epochs)) / 2) * (1 - hyp['lrf']) + hyp['lrf']  # cosine
+    scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf)
+    # plot_lr_scheduler(optimizer, scheduler, epochs)
+
+    # Logging
+    if wandb and wandb.run is None:
+        opt.hyp = hyp  # add hyperparameters
+        wandb_run = wandb.init(config=opt, resume="allow",
+                               project='YOLOR' if opt.project == 'runs/train' else Path(opt.project).stem,
+                               name=save_dir.stem,
+                               id=ckpt.get('wandb_id') if 'ckpt' in locals() else None)
+
+    # Resume
+    start_epoch, best_fitness = 0, 0.0
+    best_fitness_p, best_fitness_r, best_fitness_ap50, best_fitness_ap, best_fitness_f = 0.0, 0.0, 0.0, 0.0, 0.0
+    if pretrained:
+        # Optimizer
+        if ckpt['optimizer'] is not None:
+            optimizer.load_state_dict(ckpt['optimizer'])
+            best_fitness = ckpt['best_fitness']
+            best_fitness_p = ckpt['best_fitness_p']
+            best_fitness_r = ckpt['best_fitness_r']
+            best_fitness_ap50 = ckpt['best_fitness_ap50']
+            best_fitness_ap = ckpt['best_fitness_ap']
+            best_fitness_f = ckpt['best_fitness_f']
+
+        # Results
+        if ckpt.get('training_results') is not None:
+            with open(results_file, 'w') as file:
+                file.write(ckpt['training_results'])  # write results.txt
+
+        # Epochs
+        start_epoch = ckpt['epoch'] + 1
+        if opt.resume:
+            assert start_epoch > 0, '%s training to %g epochs is finished, nothing to resume.' % (weights, epochs)
+        if epochs < start_epoch:
+            logger.info('%s has been trained for %g epochs. Fine-tuning for %g additional epochs.' %
+                        (weights, ckpt['epoch'], epochs))
+            epochs += ckpt['epoch']  # finetune additional epochs
+
+        del ckpt, state_dict
+
+    # Image sizes
+    gs = 64 #int(max(model.stride))  # grid size (max stride)
+    imgsz, imgsz_test = [check_img_size(x, gs) for x in opt.img_size]  # verify imgsz are gs-multiples
+
+    # DP mode
+    if cuda and rank == -1 and torch.cuda.device_count() > 1:
+        model = torch.nn.DataParallel(model)
+
+    # SyncBatchNorm
+    if opt.sync_bn and cuda and rank != -1:
+        model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model).to(device)
+        logger.info('Using SyncBatchNorm()')
+
+    # EMA
+    ema = ModelEMA(model) if rank in [-1, 0] else None
+
+    # DDP mode
+    if cuda and rank != -1:
+        model = DDP(model, device_ids=[opt.local_rank], output_device=opt.local_rank)
+
+    # Trainloader
+    dataloader, dataset = create_dataloader(train_path, imgsz, batch_size, gs, opt,
+                                            hyp=hyp, augment=True, cache=opt.cache_images, rect=opt.rect,
+                                            rank=rank, world_size=opt.world_size, workers=opt.workers)
+    mlc = np.concatenate(dataset.labels, 0)[:, 0].max()  # max label class
+    nb = len(dataloader)  # number of batches
+    assert mlc < nc, 'Label class %g exceeds nc=%g in %s. Possible class labels are 0-%g' % (mlc, nc, opt.data, nc - 1)
+
+    # Process 0
+    if rank in [-1, 0]:
+        ema.updates = start_epoch * nb // accumulate  # set EMA updates
+        testloader = create_dataloader(test_path, imgsz_test, batch_size*2, gs, opt,
+                                       hyp=hyp, cache=opt.cache_images and not opt.notest, rect=True,
+                                       rank=-1, world_size=opt.world_size, workers=opt.workers)[0]  # testloader
+
+        if not opt.resume:
+            labels = np.concatenate(dataset.labels, 0)
+            c = torch.tensor(labels[:, 0])  # classes
+            # cf = torch.bincount(c.long(), minlength=nc) + 1.  # frequency
+            # model._initialize_biases(cf.to(device))
+            if plots:
+                plot_labels(labels, save_dir=save_dir)
+                if tb_writer:
+                    tb_writer.add_histogram('classes', c, 0)
+                if wandb:
+                    wandb.log({"Labels": [wandb.Image(str(x), caption=x.name) for x in save_dir.glob('*labels*.png')]})
+
+            # Anchors
+            # if not opt.noautoanchor:
+            #     check_anchors(dataset, model=model, thr=hyp['anchor_t'], imgsz=imgsz)
+
+    # Model parameters
+    hyp['cls'] *= nc / 80.  # scale coco-tuned hyp['cls'] to current dataset
+    model.nc = nc  # attach number of classes to model
+    model.hyp = hyp  # attach hyperparameters to model
+    model.gr = 1.0  # iou loss ratio (obj_loss = 1.0 or iou)
+    model.class_weights = labels_to_class_weights(dataset.labels, nc).to(device)  # attach class weights
+    model.names = names
+
+    # Start training
+    t0 = time.time()
+    nw = max(round(hyp['warmup_epochs'] * nb), 1000)  # number of warmup iterations, max(3 epochs, 1k iterations)
+    # nw = min(nw, (epochs - start_epoch) / 2 * nb)  # limit warmup to < 1/2 of training
+    maps = np.zeros(nc)  # mAP per class
+    results = (0, 0, 0, 0, 0, 0, 0)  # P, R, mAP@.5, mAP@.5-.95, val_loss(box, obj, cls)
+    scheduler.last_epoch = start_epoch - 1  # do not move
+    scaler = amp.GradScaler(enabled=cuda)
+    logger.info('Image sizes %g train, %g test\n'
+                'Using %g dataloader workers\nLogging results to %s\n'
+                'Starting training for %g epochs...' % (imgsz, imgsz_test, dataloader.num_workers, save_dir, epochs))
+    
+    torch.save(model, wdir / 'init.pt')
+    
+    for epoch in range(start_epoch, epochs):  # epoch ------------------------------------------------------------------
+        model.train()
+
+        # Update image weights (optional)
+        if opt.image_weights:
+            # Generate indices
+            if rank in [-1, 0]:
+                cw = model.class_weights.cpu().numpy() * (1 - maps) ** 2  # class weights
+                iw = labels_to_image_weights(dataset.labels, nc=nc, class_weights=cw)  # image weights
+                dataset.indices = random.choices(range(dataset.n), weights=iw, k=dataset.n)  # rand weighted idx
+            # Broadcast if DDP
+            if rank != -1:
+                indices = (torch.tensor(dataset.indices) if rank == 0 else torch.zeros(dataset.n)).int()
+                dist.broadcast(indices, 0)
+                if rank != 0:
+                    dataset.indices = indices.cpu().numpy()
+
+        # Update mosaic border
+        # b = int(random.uniform(0.25 * imgsz, 0.75 * imgsz + gs) // gs * gs)
+        # dataset.mosaic_border = [b - imgsz, -b]  # height, width borders
+
+        mloss = torch.zeros(4, device=device)  # mean losses
+        if rank != -1:
+            dataloader.sampler.set_epoch(epoch)
+        pbar = enumerate(dataloader)
+        logger.info(('\n' + '%10s' * 8) % ('Epoch', 'gpu_mem', 'box', 'obj', 'cls', 'total', 'targets', 'img_size'))
+        if rank in [-1, 0]:
+            pbar = tqdm(pbar, total=nb)  # progress bar
+        optimizer.zero_grad()
+        for i, (imgs, targets, paths, _) in pbar:  # batch -------------------------------------------------------------
+            ni = i + nb * epoch  # number integrated batches (since train start)
+            imgs = imgs.to(device, non_blocking=True).float() / 255.0  # uint8 to float32, 0-255 to 0.0-1.0
+
+            # Warmup
+            if ni <= nw:
+                xi = [0, nw]  # x interp
+                # model.gr = np.interp(ni, xi, [0.0, 1.0])  # iou loss ratio (obj_loss = 1.0 or iou)
+                accumulate = max(1, np.interp(ni, xi, [1, nbs / total_batch_size]).round())
+                for j, x in enumerate(optimizer.param_groups):
+                    # bias lr falls from 0.1 to lr0, all other lrs rise from 0.0 to lr0
+                    x['lr'] = np.interp(ni, xi, [hyp['warmup_bias_lr'] if j == 2 else 0.0, x['initial_lr'] * lf(epoch)])
+                    if 'momentum' in x:
+                        x['momentum'] = np.interp(ni, xi, [hyp['warmup_momentum'], hyp['momentum']])
+
+            # Multi-scale
+            if opt.multi_scale:
+                sz = random.randrange(imgsz * 0.5, imgsz * 1.5 + gs) // gs * gs  # size
+                sf = sz / max(imgs.shape[2:])  # scale factor
+                if sf != 1:
+                    ns = [math.ceil(x * sf / gs) * gs for x in imgs.shape[2:]]  # new shape (stretched to gs-multiple)
+                    imgs = F.interpolate(imgs, size=ns, mode='bilinear', align_corners=False)
+
+            # Forward
+            with amp.autocast(enabled=cuda):
+                pred = model(imgs)  # forward
+                loss, loss_items = compute_loss(pred, targets.to(device), model)  # loss scaled by batch_size
+                if rank != -1:
+                    loss *= opt.world_size  # gradient averaged between devices in DDP mode
+
+            # Backward
+            scaler.scale(loss).backward()
+
+            # Optimize
+            if ni % accumulate == 0:
+                scaler.step(optimizer)  # optimizer.step
+                scaler.update()
+                optimizer.zero_grad()
+                if ema:
+                    ema.update(model)
+
+            # Print
+            if rank in [-1, 0]:
+                mloss = (mloss * i + loss_items) / (i + 1)  # update mean losses
+                mem = '%.3gG' % (torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0)  # (GB)
+                s = ('%10s' * 2 + '%10.4g' * 6) % (
+                    '%g/%g' % (epoch, epochs - 1), mem, *mloss, targets.shape[0], imgs.shape[-1])
+                pbar.set_description(s)
+
+                # Plot
+                if plots and ni < 3:
+                    f = save_dir / f'train_batch{ni}.jpg'  # filename
+                    plot_images(images=imgs, targets=targets, paths=paths, fname=f)
+                    # if tb_writer:
+                    #     tb_writer.add_image(f, result, dataformats='HWC', global_step=epoch)
+                    #     tb_writer.add_graph(model, imgs)  # add model to tensorboard
+                elif plots and ni == 3 and wandb:
+                    wandb.log({"Mosaics": [wandb.Image(str(x), caption=x.name) for x in save_dir.glob('train*.jpg')]})
+
+            # end batch ------------------------------------------------------------------------------------------------
+        # end epoch ----------------------------------------------------------------------------------------------------
+
+        # Scheduler
+        lr = [x['lr'] for x in optimizer.param_groups]  # for tensorboard
+        scheduler.step()
+
+        # DDP process 0 or single-GPU
+        if rank in [-1, 0]:
+            # mAP
+            if ema:
+                ema.update_attr(model)
+            final_epoch = epoch + 1 == epochs
+            if not opt.notest or final_epoch:  # Calculate mAP
+                if epoch >= 3:
+                    results, maps, times = test.test(opt.data,
+                                                 batch_size=batch_size*2,
+                                                 imgsz=imgsz_test,
+                                                 model=ema.ema.module if hasattr(ema.ema, 'module') else ema.ema,
+                                                 single_cls=opt.single_cls,
+                                                 dataloader=testloader,
+                                                 save_dir=save_dir,
+                                                 plots=plots and final_epoch,
+                                                 log_imgs=opt.log_imgs if wandb else 0)
+
+            # Write
+            with open(results_file, 'a') as f:
+                f.write(s + '%10.4g' * 7 % results + '\n')  # P, R, mAP@.5, mAP@.5-.95, val_loss(box, obj, cls)
+            if len(opt.name) and opt.bucket:
+                os.system('gsutil cp %s gs://%s/results/results%s.txt' % (results_file, opt.bucket, opt.name))
+
+            # Log
+            tags = ['train/box_loss', 'train/obj_loss', 'train/cls_loss',  # train loss
+                    'metrics/precision', 'metrics/recall', 'metrics/mAP_0.5', 'metrics/mAP_0.5:0.95',
+                    'val/box_loss', 'val/obj_loss', 'val/cls_loss',  # val loss
+                    'x/lr0', 'x/lr1', 'x/lr2']  # params
+            for x, tag in zip(list(mloss[:-1]) + list(results) + lr, tags):
+                if tb_writer:
+                    tb_writer.add_scalar(tag, x, epoch)  # tensorboard
+                if wandb:
+                    wandb.log({tag: x})  # W&B
+
+            # Update best mAP
+            fi = fitness(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            fi_p = fitness_p(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            fi_r = fitness_r(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            fi_ap50 = fitness_ap50(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            fi_ap = fitness_ap(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            if (fi_p > 0.0) or (fi_r > 0.0):
+                fi_f = fitness_f(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            else:
+                fi_f = 0.0
+            if fi > best_fitness:
+                best_fitness = fi
+            if fi_p > best_fitness_p:
+                best_fitness_p = fi_p
+            if fi_r > best_fitness_r:
+                best_fitness_r = fi_r
+            if fi_ap50 > best_fitness_ap50:
+                best_fitness_ap50 = fi_ap50
+            if fi_ap > best_fitness_ap:
+                best_fitness_ap = fi_ap
+            if fi_f > best_fitness_f:
+                best_fitness_f = fi_f
+
+            # Save model
+            save = (not opt.nosave) or (final_epoch and not opt.evolve)
+            if save:
+                with open(results_file, 'r') as f:  # create checkpoint
+                    ckpt = {'epoch': epoch,
+                            'best_fitness': best_fitness,
+                            'best_fitness_p': best_fitness_p,
+                            'best_fitness_r': best_fitness_r,
+                            'best_fitness_ap50': best_fitness_ap50,
+                            'best_fitness_ap': best_fitness_ap,
+                            'best_fitness_f': best_fitness_f,
+                            'training_results': f.read(),
+                            'model': ema.ema.module.state_dict() if hasattr(ema, 'module') else ema.ema.state_dict(),
+                            'optimizer': None if final_epoch else optimizer.state_dict(),
+                            'wandb_id': wandb_run.id if wandb else None}
+
+                # Save last, best and delete
+                torch.save(ckpt, last)
+                if best_fitness == fi:
+                    torch.save(ckpt, best)
+                if (best_fitness == fi) and (epoch >= 200):
+                    torch.save(ckpt, wdir / 'best_{:03d}.pt'.format(epoch))
+                if best_fitness == fi:
+                    torch.save(ckpt, wdir / 'best_overall.pt')
+                if best_fitness_p == fi_p:
+                    torch.save(ckpt, wdir / 'best_p.pt')
+                if best_fitness_r == fi_r:
+                    torch.save(ckpt, wdir / 'best_r.pt')
+                if best_fitness_ap50 == fi_ap50:
+                    torch.save(ckpt, wdir / 'best_ap50.pt')
+                if best_fitness_ap == fi_ap:
+                    torch.save(ckpt, wdir / 'best_ap.pt')
+                if best_fitness_f == fi_f:
+                    torch.save(ckpt, wdir / 'best_f.pt')
+                if epoch == 0:
+                    torch.save(ckpt, wdir / 'epoch_{:03d}.pt'.format(epoch))
+                if ((epoch+1) % 25) == 0:
+                    torch.save(ckpt, wdir / 'epoch_{:03d}.pt'.format(epoch))
+                if epoch >= (epochs-5):
+                    torch.save(ckpt, wdir / 'last_{:03d}.pt'.format(epoch))
+                elif epoch >= 420: 
+                    torch.save(ckpt, wdir / 'last_{:03d}.pt'.format(epoch))
+                del ckpt
+        # end epoch ----------------------------------------------------------------------------------------------------
+    # end training
+
+    if rank in [-1, 0]:
+        # Strip optimizers
+        n = opt.name if opt.name.isnumeric() else ''
+        fresults, flast, fbest = save_dir / f'results{n}.txt', wdir / f'last{n}.pt', wdir / f'best{n}.pt'
+        for f1, f2 in zip([wdir / 'last.pt', wdir / 'best.pt', results_file], [flast, fbest, fresults]):
+            if f1.exists():
+                os.rename(f1, f2)  # rename
+                if str(f2).endswith('.pt'):  # is *.pt
+                    strip_optimizer(f2)  # strip optimizer
+                    os.system('gsutil cp %s gs://%s/weights' % (f2, opt.bucket)) if opt.bucket else None  # upload
+        # Finish
+        if plots:
+            plot_results(save_dir=save_dir)  # save as results.png
+            if wandb:
+                wandb.log({"Results": [wandb.Image(str(save_dir / x), caption=x) for x in
+                                       ['results.png', 'precision-recall_curve.png']]})
+        logger.info('%g epochs completed in %.3f hours.\n' % (epoch - start_epoch + 1, (time.time() - t0) / 3600))
+    else:
+        dist.destroy_process_group()
+
+    wandb.run.finish() if wandb and wandb.run else None
+    torch.cuda.empty_cache()
+    return results
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--weights', type=str, default='yolor_p6.pt', help='initial weights path')
+    parser.add_argument('--cfg', type=str, default='', help='model.yaml path')
+    parser.add_argument('--data', type=str, default='data/coco.yaml', help='data.yaml path')
+    parser.add_argument('--hyp', type=str, default='data/hyp.scratch.1280.yaml', help='hyperparameters path')
+    parser.add_argument('--epochs', type=int, default=300)
+    parser.add_argument('--batch-size', type=int, default=8, help='total batch size for all GPUs')
+    parser.add_argument('--img-size', nargs='+', type=int, default=[1280, 1280], help='[train, test] image sizes')
+    parser.add_argument('--rect', action='store_true', help='rectangular training')
+    parser.add_argument('--resume', nargs='?', const=True, default=False, help='resume most recent training')
+    parser.add_argument('--nosave', action='store_true', help='only save final checkpoint')
+    parser.add_argument('--notest', action='store_true', help='only test final epoch')
+    parser.add_argument('--noautoanchor', action='store_true', help='disable autoanchor check')
+    parser.add_argument('--evolve', action='store_true', help='evolve hyperparameters')
+    parser.add_argument('--bucket', type=str, default='', help='gsutil bucket')
+    parser.add_argument('--cache-images', action='store_true', help='cache images for faster training')
+    parser.add_argument('--image-weights', action='store_true', help='use weighted image selection for training')
+    parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+    parser.add_argument('--multi-scale', action='store_true', help='vary img-size +/- 50%%')
+    parser.add_argument('--single-cls', action='store_true', help='train as single-class dataset')
+    parser.add_argument('--adam', action='store_true', help='use torch.optim.Adam() optimizer')
+    parser.add_argument('--sync-bn', action='store_true', help='use SyncBatchNorm, only available in DDP mode')
+    parser.add_argument('--local_rank', type=int, default=-1, help='DDP parameter, do not modify')
+    parser.add_argument('--log-imgs', type=int, default=16, help='number of images for W&B logging, max 100')
+    parser.add_argument('--workers', type=int, default=8, help='maximum number of dataloader workers')
+    parser.add_argument('--project', default='runs/train', help='save to project/name')
+    parser.add_argument('--name', default='exp', help='save to project/name')
+    parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+    opt = parser.parse_args()
+
+    # Set DDP variables
+    opt.total_batch_size = opt.batch_size
+    opt.world_size = int(os.environ['WORLD_SIZE']) if 'WORLD_SIZE' in os.environ else 1
+    opt.global_rank = int(os.environ['RANK']) if 'RANK' in os.environ else -1
+    set_logging(opt.global_rank)
+    if opt.global_rank in [-1, 0]:
+        check_git_status()
+
+    # Resume
+    if opt.resume:  # resume an interrupted run
+        ckpt = opt.resume if isinstance(opt.resume, str) else get_latest_run()  # specified or most recent path
+        assert os.path.isfile(ckpt), 'ERROR: --resume checkpoint does not exist'
+        with open(Path(ckpt).parent.parent / 'opt.yaml') as f:
+            opt = argparse.Namespace(**yaml.load(f, Loader=yaml.FullLoader))  # replace
+        opt.cfg, opt.weights, opt.resume = '', ckpt, True
+        logger.info('Resuming training from %s' % ckpt)
+    else:
+        # opt.hyp = opt.hyp or ('hyp.finetune.yaml' if opt.weights else 'hyp.scratch.yaml')
+        opt.data, opt.cfg, opt.hyp = check_file(opt.data), check_file(opt.cfg), check_file(opt.hyp)  # check files
+        assert len(opt.cfg) or len(opt.weights), 'either --cfg or --weights must be specified'
+        opt.img_size.extend([opt.img_size[-1]] * (2 - len(opt.img_size)))  # extend to 2 sizes (train, test)
+        opt.name = 'evolve' if opt.evolve else opt.name
+        opt.save_dir = increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok | opt.evolve)  # increment run
+
+    # DDP mode
+    device = select_device(opt.device, batch_size=opt.batch_size)
+    if opt.local_rank != -1:
+        assert torch.cuda.device_count() > opt.local_rank
+        torch.cuda.set_device(opt.local_rank)
+        device = torch.device('cuda', opt.local_rank)
+        dist.init_process_group(backend='nccl', init_method='env://')  # distributed backend
+        assert opt.batch_size % opt.world_size == 0, '--batch-size must be multiple of CUDA device count'
+        opt.batch_size = opt.total_batch_size // opt.world_size
+
+    # Hyperparameters
+    with open(opt.hyp) as f:
+        hyp = yaml.load(f, Loader=yaml.FullLoader)  # load hyps
+        if 'box' not in hyp:
+            warn('Compatibility: %s missing "box" which was renamed from "giou" in %s' %
+                 (opt.hyp, 'https://github.com/ultralytics/yolov5/pull/1120'))
+            hyp['box'] = hyp.pop('giou')
+
+    # Train
+    logger.info(opt)
+    if not opt.evolve:
+        tb_writer = None  # init loggers
+        if opt.global_rank in [-1, 0]:
+            logger.info(f'Start Tensorboard with "tensorboard --logdir {opt.project}", view at http://localhost:6006/')
+            tb_writer = SummaryWriter(opt.save_dir)  # Tensorboard
+        train(hyp, opt, device, tb_writer, wandb)
+
+    # Evolve hyperparameters (optional)
+    else:
+        # Hyperparameter evolution metadata (mutation scale 0-1, lower_limit, upper_limit)
+        meta = {'lr0': (1, 1e-5, 1e-1),  # initial learning rate (SGD=1E-2, Adam=1E-3)
+                'lrf': (1, 0.01, 1.0),  # final OneCycleLR learning rate (lr0 * lrf)
+                'momentum': (0.3, 0.6, 0.98),  # SGD momentum/Adam beta1
+                'weight_decay': (1, 0.0, 0.001),  # optimizer weight decay
+                'warmup_epochs': (1, 0.0, 5.0),  # warmup epochs (fractions ok)
+                'warmup_momentum': (1, 0.0, 0.95),  # warmup initial momentum
+                'warmup_bias_lr': (1, 0.0, 0.2),  # warmup initial bias lr
+                'box': (1, 0.02, 0.2),  # box loss gain
+                'cls': (1, 0.2, 4.0),  # cls loss gain
+                'cls_pw': (1, 0.5, 2.0),  # cls BCELoss positive_weight
+                'obj': (1, 0.2, 4.0),  # obj loss gain (scale with pixels)
+                'obj_pw': (1, 0.5, 2.0),  # obj BCELoss positive_weight
+                'iou_t': (0, 0.1, 0.7),  # IoU training threshold
+                'anchor_t': (1, 2.0, 8.0),  # anchor-multiple threshold
+                'anchors': (2, 2.0, 10.0),  # anchors per output grid (0 to ignore)
+                'fl_gamma': (0, 0.0, 2.0),  # focal loss gamma (efficientDet default gamma=1.5)
+                'hsv_h': (1, 0.0, 0.1),  # image HSV-Hue augmentation (fraction)
+                'hsv_s': (1, 0.0, 0.9),  # image HSV-Saturation augmentation (fraction)
+                'hsv_v': (1, 0.0, 0.9),  # image HSV-Value augmentation (fraction)
+                'degrees': (1, 0.0, 45.0),  # image rotation (+/- deg)
+                'translate': (1, 0.0, 0.9),  # image translation (+/- fraction)
+                'scale': (1, 0.0, 0.9),  # image scale (+/- gain)
+                'shear': (1, 0.0, 10.0),  # image shear (+/- deg)
+                'perspective': (0, 0.0, 0.001),  # image perspective (+/- fraction), range 0-0.001
+                'flipud': (1, 0.0, 1.0),  # image flip up-down (probability)
+                'fliplr': (0, 0.0, 1.0),  # image flip left-right (probability)
+                'mosaic': (1, 0.0, 1.0),  # image mixup (probability)
+                'mixup': (1, 0.0, 1.0)}  # image mixup (probability)
+
+        assert opt.local_rank == -1, 'DDP mode not implemented for --evolve'
+        opt.notest, opt.nosave = True, True  # only test/save final epoch
+        # ei = [isinstance(x, (int, float)) for x in hyp.values()]  # evolvable indices
+        yaml_file = Path(opt.save_dir) / 'hyp_evolved.yaml'  # save best result here
+        if opt.bucket:
+            os.system('gsutil cp gs://%s/evolve.txt .' % opt.bucket)  # download evolve.txt if exists
+
+        for _ in range(300):  # generations to evolve
+            if Path('evolve.txt').exists():  # if evolve.txt exists: select best hyps and mutate
+                # Select parent(s)
+                parent = 'single'  # parent selection method: 'single' or 'weighted'
+                x = np.loadtxt('evolve.txt', ndmin=2)
+                n = min(5, len(x))  # number of previous results to consider
+                x = x[np.argsort(-fitness(x))][:n]  # top n mutations
+                w = fitness(x) - fitness(x).min()  # weights
+                if parent == 'single' or len(x) == 1:
+                    # x = x[random.randint(0, n - 1)]  # random selection
+                    x = x[random.choices(range(n), weights=w)[0]]  # weighted selection
+                elif parent == 'weighted':
+                    x = (x * w.reshape(n, 1)).sum(0) / w.sum()  # weighted combination
+
+                # Mutate
+                mp, s = 0.8, 0.2  # mutation probability, sigma
+                npr = np.random
+                npr.seed(int(time.time()))
+                g = np.array([x[0] for x in meta.values()])  # gains 0-1
+                ng = len(meta)
+                v = np.ones(ng)
+                while all(v == 1):  # mutate until a change occurs (prevent duplicates)
+                    v = (g * (npr.random(ng) < mp) * npr.randn(ng) * npr.random() * s + 1).clip(0.3, 3.0)
+                for i, k in enumerate(hyp.keys()):  # plt.hist(v.ravel(), 300)
+                    hyp[k] = float(x[i + 7] * v[i])  # mutate
+
+            # Constrain to limits
+            for k, v in meta.items():
+                hyp[k] = max(hyp[k], v[1])  # lower limit
+                hyp[k] = min(hyp[k], v[2])  # upper limit
+                hyp[k] = round(hyp[k], 5)  # significant digits
+
+            # Train mutation
+            results = train(hyp.copy(), opt, device, wandb=wandb)
+
+            # Write mutation results
+            print_mutation(hyp.copy(), results, yaml_file, opt.bucket)
+
+        # Plot results
+        plot_evolution(yaml_file)
+        print(f'Hyperparameter evolution complete. Best results saved as: {yaml_file}\n'
+              f'Command to train a new model with these hyperparameters: $ python train.py --hyp {yaml_file}')

tune.py

@@ -0,0 +1,619 @@
+import argparse
+import logging
+import math
+import os
+import random
+import time
+from pathlib import Path
+from warnings import warn
+
+import numpy as np
+import torch.distributed as dist
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import torch.optim.lr_scheduler as lr_scheduler
+import torch.utils.data
+import yaml
+from torch.cuda import amp
+from torch.nn.parallel import DistributedDataParallel as DDP
+from torch.utils.tensorboard import SummaryWriter
+from tqdm import tqdm
+
+import test  # import test.py to get mAP after each epoch
+#from models.yolo import Model
+from models.models import *
+from utils.autoanchor import check_anchors
+from utils.datasets import create_dataloader9 as create_dataloader
+from utils.general import labels_to_class_weights, increment_path, labels_to_image_weights, init_seeds, \
+    fitness, fitness_p, fitness_r, fitness_ap50, fitness_ap, fitness_f, strip_optimizer, get_latest_run,\
+    check_dataset, check_file, check_git_status, check_img_size, print_mutation, set_logging
+from utils.google_utils import attempt_download
+from utils.loss import compute_loss
+from utils.plots import plot_images, plot_labels, plot_results, plot_evolution
+from utils.torch_utils import ModelEMA, select_device, intersect_dicts, torch_distributed_zero_first
+
+logger = logging.getLogger(__name__)
+
+try:
+    import wandb
+except ImportError:
+    wandb = None
+    logger.info("Install Weights & Biases for experiment logging via 'pip install wandb' (recommended)")
+
+def train(hyp, opt, device, tb_writer=None, wandb=None):
+    logger.info(f'Hyperparameters {hyp}')
+    save_dir, epochs, batch_size, total_batch_size, weights, rank = \
+        Path(opt.save_dir), opt.epochs, opt.batch_size, opt.total_batch_size, opt.weights, opt.global_rank
+
+    # Directories
+    wdir = save_dir / 'weights'
+    wdir.mkdir(parents=True, exist_ok=True)  # make dir
+    last = wdir / 'last.pt'
+    best = wdir / 'best.pt'
+    results_file = save_dir / 'results.txt'
+
+    # Save run settings
+    with open(save_dir / 'hyp.yaml', 'w') as f:
+        yaml.dump(hyp, f, sort_keys=False)
+    with open(save_dir / 'opt.yaml', 'w') as f:
+        yaml.dump(vars(opt), f, sort_keys=False)
+
+    # Configure
+    plots = not opt.evolve  # create plots
+    cuda = device.type != 'cpu'
+    init_seeds(2 + rank)
+    with open(opt.data) as f:
+        data_dict = yaml.load(f, Loader=yaml.FullLoader)  # data dict
+    with torch_distributed_zero_first(rank):
+        check_dataset(data_dict)  # check
+    train_path = data_dict['train']
+    test_path = data_dict['val']
+    nc, names = (1, ['item']) if opt.single_cls else (int(data_dict['nc']), data_dict['names'])  # number classes, names
+    assert len(names) == nc, '%g names found for nc=%g dataset in %s' % (len(names), nc, opt.data)  # check
+
+    # Model
+    pretrained = weights.endswith('.pt')
+    if pretrained:
+        with torch_distributed_zero_first(rank):
+            attempt_download(weights)  # download if not found locally
+        ckpt = torch.load(weights, map_location=device)  # load checkpoint
+        model = Darknet(opt.cfg).to(device)  # create
+        state_dict = {k: v for k, v in ckpt['model'].items() if model.state_dict()[k].numel() == v.numel()}
+        model.load_state_dict(state_dict, strict=False)
+        print('Transferred %g/%g items from %s' % (len(state_dict), len(model.state_dict()), weights))  # report
+    else:
+        model = Darknet(opt.cfg).to(device) # create
+
+    # Optimizer
+    nbs = 64  # nominal batch size
+    accumulate = max(round(nbs / total_batch_size), 1)  # accumulate loss before optimizing
+    hyp['weight_decay'] *= total_batch_size * accumulate / nbs  # scale weight_decay
+
+    pg0, pg1, pg2 = [], [], []  # optimizer parameter groups
+    for k, v in dict(model.named_parameters()).items():
+        if '.bias' in k:
+            pg2.append(v)  # biases
+        elif 'Conv2d.weight' in k:
+            pg1.append(v)  # apply weight_decay
+        elif 'm.weight' in k:
+            pg1.append(v)  # apply weight_decay
+        elif 'w.weight' in k:
+            pg1.append(v)  # apply weight_decay
+        else:
+            pg0.append(v)  # all else
+
+    if opt.adam:
+        optimizer = optim.Adam(pg0, lr=hyp['lr0'], betas=(hyp['momentum'], 0.999))  # adjust beta1 to momentum
+    else:
+        optimizer = optim.SGD(pg0, lr=hyp['lr0'], momentum=hyp['momentum'], nesterov=True)
+
+    optimizer.add_param_group({'params': pg1, 'weight_decay': hyp['weight_decay']})  # add pg1 with weight_decay
+    optimizer.add_param_group({'params': pg2})  # add pg2 (biases)
+    logger.info('Optimizer groups: %g .bias, %g conv.weight, %g other' % (len(pg2), len(pg1), len(pg0)))
+    del pg0, pg1, pg2
+
+    # Scheduler https://arxiv.org/pdf/1812.01187.pdf
+    # https://pytorch.org/docs/stable/_modules/torch/optim/lr_scheduler.html#OneCycleLR
+    lf = lambda x: ((1 + math.cos(x * math.pi / epochs)) / 2) * (1 - hyp['lrf']) + hyp['lrf']  # cosine
+    scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf)
+    # plot_lr_scheduler(optimizer, scheduler, epochs)
+
+    # Logging
+    if wandb and wandb.run is None:
+        opt.hyp = hyp  # add hyperparameters
+        wandb_run = wandb.init(config=opt, resume="allow",
+                               project='YOLOR' if opt.project == 'runs/train' else Path(opt.project).stem,
+                               name=save_dir.stem,
+                               id=ckpt.get('wandb_id') if 'ckpt' in locals() else None)
+
+    # Resume
+    start_epoch, best_fitness = 0, 0.0
+    best_fitness_p, best_fitness_r, best_fitness_ap50, best_fitness_ap, best_fitness_f = 0.0, 0.0, 0.0, 0.0, 0.0
+    if pretrained:
+        # Optimizer
+        if ckpt['optimizer'] is not None:
+            optimizer.load_state_dict(ckpt['optimizer'])
+            best_fitness = ckpt['best_fitness']
+            best_fitness_p = ckpt['best_fitness_p']
+            best_fitness_r = ckpt['best_fitness_r']
+            best_fitness_ap50 = ckpt['best_fitness_ap50']
+            best_fitness_ap = ckpt['best_fitness_ap']
+            best_fitness_f = ckpt['best_fitness_f']
+
+        # Results
+        if ckpt.get('training_results') is not None:
+            with open(results_file, 'w') as file:
+                file.write(ckpt['training_results'])  # write results.txt
+
+        # Epochs
+        start_epoch = ckpt['epoch'] + 1
+        if opt.resume:
+            assert start_epoch > 0, '%s training to %g epochs is finished, nothing to resume.' % (weights, epochs)
+        if epochs < start_epoch:
+            logger.info('%s has been trained for %g epochs. Fine-tuning for %g additional epochs.' %
+                        (weights, ckpt['epoch'], epochs))
+            epochs += ckpt['epoch']  # finetune additional epochs
+
+        del ckpt, state_dict
+
+    # Image sizes
+    gs = 64 #int(max(model.stride))  # grid size (max stride)
+    imgsz, imgsz_test = [check_img_size(x, gs) for x in opt.img_size]  # verify imgsz are gs-multiples
+
+    # DP mode
+    if cuda and rank == -1 and torch.cuda.device_count() > 1:
+        model = torch.nn.DataParallel(model)
+
+    # SyncBatchNorm
+    if opt.sync_bn and cuda and rank != -1:
+        model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model).to(device)
+        logger.info('Using SyncBatchNorm()')
+
+    # EMA
+    ema = ModelEMA(model) if rank in [-1, 0] else None
+
+    # DDP mode
+    if cuda and rank != -1:
+        model = DDP(model, device_ids=[opt.local_rank], output_device=opt.local_rank)
+
+    # Trainloader
+    dataloader, dataset = create_dataloader(train_path, imgsz, batch_size, gs, opt,
+                                            hyp=hyp, augment=True, cache=opt.cache_images, rect=opt.rect,
+                                            rank=rank, world_size=opt.world_size, workers=opt.workers)
+    mlc = np.concatenate(dataset.labels, 0)[:, 0].max()  # max label class
+    nb = len(dataloader)  # number of batches
+    assert mlc < nc, 'Label class %g exceeds nc=%g in %s. Possible class labels are 0-%g' % (mlc, nc, opt.data, nc - 1)
+
+    # Process 0
+    if rank in [-1, 0]:
+        ema.updates = start_epoch * nb // accumulate  # set EMA updates
+        testloader = create_dataloader(test_path, imgsz_test, batch_size*2, gs, opt,
+                                       hyp=hyp, cache=opt.cache_images and not opt.notest, rect=True,
+                                       rank=-1, world_size=opt.world_size, workers=opt.workers)[0]  # testloader
+
+        if not opt.resume:
+            labels = np.concatenate(dataset.labels, 0)
+            c = torch.tensor(labels[:, 0])  # classes
+            # cf = torch.bincount(c.long(), minlength=nc) + 1.  # frequency
+            # model._initialize_biases(cf.to(device))
+            if plots:
+                plot_labels(labels, save_dir=save_dir)
+                if tb_writer:
+                    tb_writer.add_histogram('classes', c, 0)
+                if wandb:
+                    wandb.log({"Labels": [wandb.Image(str(x), caption=x.name) for x in save_dir.glob('*labels*.png')]})
+
+            # Anchors
+            # if not opt.noautoanchor:
+            #     check_anchors(dataset, model=model, thr=hyp['anchor_t'], imgsz=imgsz)
+
+    # Model parameters
+    hyp['cls'] *= nc / 80.  # scale coco-tuned hyp['cls'] to current dataset
+    model.nc = nc  # attach number of classes to model
+    model.hyp = hyp  # attach hyperparameters to model
+    model.gr = 1.0  # iou loss ratio (obj_loss = 1.0 or iou)
+    model.class_weights = labels_to_class_weights(dataset.labels, nc).to(device)  # attach class weights
+    model.names = names
+
+    # Start training
+    t0 = time.time()
+    nw = max(round(hyp['warmup_epochs'] * nb), 1000)  # number of warmup iterations, max(3 epochs, 1k iterations)
+    # nw = min(nw, (epochs - start_epoch) / 2 * nb)  # limit warmup to < 1/2 of training
+    maps = np.zeros(nc)  # mAP per class
+    results = (0, 0, 0, 0, 0, 0, 0)  # P, R, mAP@.5, mAP@.5-.95, val_loss(box, obj, cls)
+    scheduler.last_epoch = start_epoch - 1  # do not move
+    scaler = amp.GradScaler(enabled=cuda)
+    logger.info('Image sizes %g train, %g test\n'
+                'Using %g dataloader workers\nLogging results to %s\n'
+                'Starting training for %g epochs...' % (imgsz, imgsz_test, dataloader.num_workers, save_dir, epochs))
+    
+    torch.save(model, wdir / 'init.pt')
+    
+    for epoch in range(start_epoch, epochs):  # epoch ------------------------------------------------------------------
+        model.train()
+
+        # Update image weights (optional)
+        if opt.image_weights:
+            # Generate indices
+            if rank in [-1, 0]:
+                cw = model.class_weights.cpu().numpy() * (1 - maps) ** 2  # class weights
+                iw = labels_to_image_weights(dataset.labels, nc=nc, class_weights=cw)  # image weights
+                dataset.indices = random.choices(range(dataset.n), weights=iw, k=dataset.n)  # rand weighted idx
+            # Broadcast if DDP
+            if rank != -1:
+                indices = (torch.tensor(dataset.indices) if rank == 0 else torch.zeros(dataset.n)).int()
+                dist.broadcast(indices, 0)
+                if rank != 0:
+                    dataset.indices = indices.cpu().numpy()
+
+        # Update mosaic border
+        # b = int(random.uniform(0.25 * imgsz, 0.75 * imgsz + gs) // gs * gs)
+        # dataset.mosaic_border = [b - imgsz, -b]  # height, width borders
+
+        mloss = torch.zeros(4, device=device)  # mean losses
+        if rank != -1:
+            dataloader.sampler.set_epoch(epoch)
+        pbar = enumerate(dataloader)
+        logger.info(('\n' + '%10s' * 8) % ('Epoch', 'gpu_mem', 'box', 'obj', 'cls', 'total', 'targets', 'img_size'))
+        if rank in [-1, 0]:
+            pbar = tqdm(pbar, total=nb)  # progress bar
+        optimizer.zero_grad()
+        for i, (imgs, targets, paths, _) in pbar:  # batch -------------------------------------------------------------
+            ni = i + nb * epoch  # number integrated batches (since train start)
+            imgs = imgs.to(device, non_blocking=True).float() / 255.0  # uint8 to float32, 0-255 to 0.0-1.0
+
+            # Warmup
+            if ni <= nw:
+                xi = [0, nw]  # x interp
+                # model.gr = np.interp(ni, xi, [0.0, 1.0])  # iou loss ratio (obj_loss = 1.0 or iou)
+                accumulate = max(1, np.interp(ni, xi, [1, nbs / total_batch_size]).round())
+                for j, x in enumerate(optimizer.param_groups):
+                    # bias lr falls from 0.1 to lr0, all other lrs rise from 0.0 to lr0
+                    x['lr'] = np.interp(ni, xi, [hyp['warmup_bias_lr'] if j == 2 else 0.0, x['initial_lr'] * lf(epoch)])
+                    if 'momentum' in x:
+                        x['momentum'] = np.interp(ni, xi, [hyp['warmup_momentum'], hyp['momentum']])
+
+            # Multi-scale
+            if opt.multi_scale:
+                sz = random.randrange(imgsz * 0.5, imgsz * 1.5 + gs) // gs * gs  # size
+                sf = sz / max(imgs.shape[2:])  # scale factor
+                if sf != 1:
+                    ns = [math.ceil(x * sf / gs) * gs for x in imgs.shape[2:]]  # new shape (stretched to gs-multiple)
+                    imgs = F.interpolate(imgs, size=ns, mode='bilinear', align_corners=False)
+
+            # Forward
+            with amp.autocast(enabled=cuda):
+                pred = model(imgs)  # forward
+                loss, loss_items = compute_loss(pred, targets.to(device), model)  # loss scaled by batch_size
+                if rank != -1:
+                    loss *= opt.world_size  # gradient averaged between devices in DDP mode
+
+            # Backward
+            scaler.scale(loss).backward()
+
+            # Optimize
+            if ni % accumulate == 0:
+                scaler.step(optimizer)  # optimizer.step
+                scaler.update()
+                optimizer.zero_grad()
+                if ema:
+                    ema.update(model)
+
+            # Print
+            if rank in [-1, 0]:
+                mloss = (mloss * i + loss_items) / (i + 1)  # update mean losses
+                mem = '%.3gG' % (torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0)  # (GB)
+                s = ('%10s' * 2 + '%10.4g' * 6) % (
+                    '%g/%g' % (epoch, epochs - 1), mem, *mloss, targets.shape[0], imgs.shape[-1])
+                pbar.set_description(s)
+
+                # Plot
+                if plots and ni < 3:
+                    f = save_dir / f'train_batch{ni}.jpg'  # filename
+                    plot_images(images=imgs, targets=targets, paths=paths, fname=f)
+                    # if tb_writer:
+                    #     tb_writer.add_image(f, result, dataformats='HWC', global_step=epoch)
+                    #     tb_writer.add_graph(model, imgs)  # add model to tensorboard
+                elif plots and ni == 3 and wandb:
+                    wandb.log({"Mosaics": [wandb.Image(str(x), caption=x.name) for x in save_dir.glob('train*.jpg')]})
+
+            # end batch ------------------------------------------------------------------------------------------------
+        # end epoch ----------------------------------------------------------------------------------------------------
+
+        # Scheduler
+        lr = [x['lr'] for x in optimizer.param_groups]  # for tensorboard
+        scheduler.step()
+
+        # DDP process 0 or single-GPU
+        if rank in [-1, 0]:
+            # mAP
+            if ema:
+                ema.update_attr(model)
+            final_epoch = epoch + 1 == epochs
+            if not opt.notest or final_epoch:  # Calculate mAP
+                if epoch >= 3:
+                    results, maps, times = test.test(opt.data,
+                                                 batch_size=batch_size*2,
+                                                 imgsz=imgsz_test,
+                                                 model=ema.ema.module if hasattr(ema.ema, 'module') else ema.ema,
+                                                 single_cls=opt.single_cls,
+                                                 dataloader=testloader,
+                                                 save_dir=save_dir,
+                                                 plots=plots and final_epoch,
+                                                 log_imgs=opt.log_imgs if wandb else 0)
+
+            # Write
+            with open(results_file, 'a') as f:
+                f.write(s + '%10.4g' * 7 % results + '\n')  # P, R, mAP@.5, mAP@.5-.95, val_loss(box, obj, cls)
+            if len(opt.name) and opt.bucket:
+                os.system('gsutil cp %s gs://%s/results/results%s.txt' % (results_file, opt.bucket, opt.name))
+
+            # Log
+            tags = ['train/box_loss', 'train/obj_loss', 'train/cls_loss',  # train loss
+                    'metrics/precision', 'metrics/recall', 'metrics/mAP_0.5', 'metrics/mAP_0.5:0.95',
+                    'val/box_loss', 'val/obj_loss', 'val/cls_loss',  # val loss
+                    'x/lr0', 'x/lr1', 'x/lr2']  # params
+            for x, tag in zip(list(mloss[:-1]) + list(results) + lr, tags):
+                if tb_writer:
+                    tb_writer.add_scalar(tag, x, epoch)  # tensorboard
+                if wandb:
+                    wandb.log({tag: x})  # W&B
+
+            # Update best mAP
+            fi = fitness(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            fi_p = fitness_p(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            fi_r = fitness_r(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            fi_ap50 = fitness_ap50(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            fi_ap = fitness_ap(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            if (fi_p > 0.0) or (fi_r > 0.0):
+                fi_f = fitness_f(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            else:
+                fi_f = 0.0
+            if fi > best_fitness:
+                best_fitness = fi
+            if fi_p > best_fitness_p:
+                best_fitness_p = fi_p
+            if fi_r > best_fitness_r:
+                best_fitness_r = fi_r
+            if fi_ap50 > best_fitness_ap50:
+                best_fitness_ap50 = fi_ap50
+            if fi_ap > best_fitness_ap:
+                best_fitness_ap = fi_ap
+            if fi_f > best_fitness_f:
+                best_fitness_f = fi_f
+
+            # Save model
+            save = (not opt.nosave) or (final_epoch and not opt.evolve)
+            if save:
+                with open(results_file, 'r') as f:  # create checkpoint
+                    ckpt = {'epoch': epoch,
+                            'best_fitness': best_fitness,
+                            'best_fitness_p': best_fitness_p,
+                            'best_fitness_r': best_fitness_r,
+                            'best_fitness_ap50': best_fitness_ap50,
+                            'best_fitness_ap': best_fitness_ap,
+                            'best_fitness_f': best_fitness_f,
+                            'training_results': f.read(),
+                            'model': ema.ema.module.state_dict() if hasattr(ema, 'module') else ema.ema.state_dict(),
+                            'optimizer': None if final_epoch else optimizer.state_dict(),
+                            'wandb_id': wandb_run.id if wandb else None}
+
+                # Save last, best and delete
+                torch.save(ckpt, last)
+                if best_fitness == fi:
+                    torch.save(ckpt, best)
+                if (best_fitness == fi) and (epoch >= 200):
+                    torch.save(ckpt, wdir / 'best_{:03d}.pt'.format(epoch))
+                if best_fitness == fi:
+                    torch.save(ckpt, wdir / 'best_overall.pt')
+                if best_fitness_p == fi_p:
+                    torch.save(ckpt, wdir / 'best_p.pt')
+                if best_fitness_r == fi_r:
+                    torch.save(ckpt, wdir / 'best_r.pt')
+                if best_fitness_ap50 == fi_ap50:
+                    torch.save(ckpt, wdir / 'best_ap50.pt')
+                if best_fitness_ap == fi_ap:
+                    torch.save(ckpt, wdir / 'best_ap.pt')
+                if best_fitness_f == fi_f:
+                    torch.save(ckpt, wdir / 'best_f.pt')
+                if epoch == 0:
+                    torch.save(ckpt, wdir / 'epoch_{:03d}.pt'.format(epoch))
+                if ((epoch+1) % 25) == 0:
+                    torch.save(ckpt, wdir / 'epoch_{:03d}.pt'.format(epoch))
+                if epoch >= (epochs-5):
+                    torch.save(ckpt, wdir / 'last_{:03d}.pt'.format(epoch))
+                elif epoch >= 420: 
+                    torch.save(ckpt, wdir / 'last_{:03d}.pt'.format(epoch))
+                del ckpt
+        # end epoch ----------------------------------------------------------------------------------------------------
+    # end training
+
+    if rank in [-1, 0]:
+        # Strip optimizers
+        n = opt.name if opt.name.isnumeric() else ''
+        fresults, flast, fbest = save_dir / f'results{n}.txt', wdir / f'last{n}.pt', wdir / f'best{n}.pt'
+        for f1, f2 in zip([wdir / 'last.pt', wdir / 'best.pt', results_file], [flast, fbest, fresults]):
+            if f1.exists():
+                os.rename(f1, f2)  # rename
+                if str(f2).endswith('.pt'):  # is *.pt
+                    strip_optimizer(f2)  # strip optimizer
+                    os.system('gsutil cp %s gs://%s/weights' % (f2, opt.bucket)) if opt.bucket else None  # upload
+        # Finish
+        if plots:
+            plot_results(save_dir=save_dir)  # save as results.png
+            if wandb:
+                wandb.log({"Results": [wandb.Image(str(save_dir / x), caption=x) for x in
+                                       ['results.png', 'precision-recall_curve.png']]})
+        logger.info('%g epochs completed in %.3f hours.\n' % (epoch - start_epoch + 1, (time.time() - t0) / 3600))
+    else:
+        dist.destroy_process_group()
+
+    wandb.run.finish() if wandb and wandb.run else None
+    torch.cuda.empty_cache()
+    return results
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--weights', type=str, default='yolor_p6.pt', help='initial weights path')
+    parser.add_argument('--cfg', type=str, default='', help='model.yaml path')
+    parser.add_argument('--data', type=str, default='data/coco.yaml', help='data.yaml path')
+    parser.add_argument('--hyp', type=str, default='data/hyp.scratch.1280.yaml', help='hyperparameters path')
+    parser.add_argument('--epochs', type=int, default=300)
+    parser.add_argument('--batch-size', type=int, default=8, help='total batch size for all GPUs')
+    parser.add_argument('--img-size', nargs='+', type=int, default=[1280, 1280], help='[train, test] image sizes')
+    parser.add_argument('--rect', action='store_true', help='rectangular training')
+    parser.add_argument('--resume', nargs='?', const=True, default=False, help='resume most recent training')
+    parser.add_argument('--nosave', action='store_true', help='only save final checkpoint')
+    parser.add_argument('--notest', action='store_true', help='only test final epoch')
+    parser.add_argument('--noautoanchor', action='store_true', help='disable autoanchor check')
+    parser.add_argument('--evolve', action='store_true', help='evolve hyperparameters')
+    parser.add_argument('--bucket', type=str, default='', help='gsutil bucket')
+    parser.add_argument('--cache-images', action='store_true', help='cache images for faster training')
+    parser.add_argument('--image-weights', action='store_true', help='use weighted image selection for training')
+    parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+    parser.add_argument('--multi-scale', action='store_true', help='vary img-size +/- 50%%')
+    parser.add_argument('--single-cls', action='store_true', help='train as single-class dataset')
+    parser.add_argument('--adam', action='store_true', help='use torch.optim.Adam() optimizer')
+    parser.add_argument('--sync-bn', action='store_true', help='use SyncBatchNorm, only available in DDP mode')
+    parser.add_argument('--local_rank', type=int, default=-1, help='DDP parameter, do not modify')
+    parser.add_argument('--log-imgs', type=int, default=16, help='number of images for W&B logging, max 100')
+    parser.add_argument('--workers', type=int, default=8, help='maximum number of dataloader workers')
+    parser.add_argument('--project', default='runs/train', help='save to project/name')
+    parser.add_argument('--name', default='exp', help='save to project/name')
+    parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+    opt = parser.parse_args()
+
+    # Set DDP variables
+    opt.total_batch_size = opt.batch_size
+    opt.world_size = int(os.environ['WORLD_SIZE']) if 'WORLD_SIZE' in os.environ else 1
+    opt.global_rank = int(os.environ['RANK']) if 'RANK' in os.environ else -1
+    set_logging(opt.global_rank)
+    if opt.global_rank in [-1, 0]:
+        check_git_status()
+
+    # Resume
+    if opt.resume:  # resume an interrupted run
+        ckpt = opt.resume if isinstance(opt.resume, str) else get_latest_run()  # specified or most recent path
+        assert os.path.isfile(ckpt), 'ERROR: --resume checkpoint does not exist'
+        with open(Path(ckpt).parent.parent / 'opt.yaml') as f:
+            opt = argparse.Namespace(**yaml.load(f, Loader=yaml.FullLoader))  # replace
+        opt.cfg, opt.weights, opt.resume = '', ckpt, True
+        logger.info('Resuming training from %s' % ckpt)
+    else:
+        # opt.hyp = opt.hyp or ('hyp.finetune.yaml' if opt.weights else 'hyp.scratch.yaml')
+        opt.data, opt.cfg, opt.hyp = check_file(opt.data), check_file(opt.cfg), check_file(opt.hyp)  # check files
+        assert len(opt.cfg) or len(opt.weights), 'either --cfg or --weights must be specified'
+        opt.img_size.extend([opt.img_size[-1]] * (2 - len(opt.img_size)))  # extend to 2 sizes (train, test)
+        opt.name = 'evolve' if opt.evolve else opt.name
+        opt.save_dir = increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok | opt.evolve)  # increment run
+
+    # DDP mode
+    device = select_device(opt.device, batch_size=opt.batch_size)
+    if opt.local_rank != -1:
+        assert torch.cuda.device_count() > opt.local_rank
+        torch.cuda.set_device(opt.local_rank)
+        device = torch.device('cuda', opt.local_rank)
+        dist.init_process_group(backend='nccl', init_method='env://')  # distributed backend
+        assert opt.batch_size % opt.world_size == 0, '--batch-size must be multiple of CUDA device count'
+        opt.batch_size = opt.total_batch_size // opt.world_size
+
+    # Hyperparameters
+    with open(opt.hyp) as f:
+        hyp = yaml.load(f, Loader=yaml.FullLoader)  # load hyps
+        if 'box' not in hyp:
+            warn('Compatibility: %s missing "box" which was renamed from "giou" in %s' %
+                 (opt.hyp, 'https://github.com/ultralytics/yolov5/pull/1120'))
+            hyp['box'] = hyp.pop('giou')
+
+    # Train
+    logger.info(opt)
+    if not opt.evolve:
+        tb_writer = None  # init loggers
+        if opt.global_rank in [-1, 0]:
+            logger.info(f'Start Tensorboard with "tensorboard --logdir {opt.project}", view at http://localhost:6006/')
+            tb_writer = SummaryWriter(opt.save_dir)  # Tensorboard
+        train(hyp, opt, device, tb_writer, wandb)
+
+    # Evolve hyperparameters (optional)
+    else:
+        # Hyperparameter evolution metadata (mutation scale 0-1, lower_limit, upper_limit)
+        meta = {'lr0': (1, 1e-5, 1e-1),  # initial learning rate (SGD=1E-2, Adam=1E-3)
+                'lrf': (1, 0.01, 1.0),  # final OneCycleLR learning rate (lr0 * lrf)
+                'momentum': (0.3, 0.6, 0.98),  # SGD momentum/Adam beta1
+                'weight_decay': (1, 0.0, 0.001),  # optimizer weight decay
+                'warmup_epochs': (1, 0.0, 5.0),  # warmup epochs (fractions ok)
+                'warmup_momentum': (1, 0.0, 0.95),  # warmup initial momentum
+                'warmup_bias_lr': (1, 0.0, 0.2),  # warmup initial bias lr
+                'box': (1, 0.02, 0.2),  # box loss gain
+                'cls': (1, 0.2, 4.0),  # cls loss gain
+                'cls_pw': (1, 0.5, 2.0),  # cls BCELoss positive_weight
+                'obj': (1, 0.2, 4.0),  # obj loss gain (scale with pixels)
+                'obj_pw': (1, 0.5, 2.0),  # obj BCELoss positive_weight
+                'iou_t': (0, 0.1, 0.7),  # IoU training threshold
+                'anchor_t': (1, 2.0, 8.0),  # anchor-multiple threshold
+                'anchors': (2, 2.0, 10.0),  # anchors per output grid (0 to ignore)
+                'fl_gamma': (0, 0.0, 2.0),  # focal loss gamma (efficientDet default gamma=1.5)
+                'hsv_h': (1, 0.0, 0.1),  # image HSV-Hue augmentation (fraction)
+                'hsv_s': (1, 0.0, 0.9),  # image HSV-Saturation augmentation (fraction)
+                'hsv_v': (1, 0.0, 0.9),  # image HSV-Value augmentation (fraction)
+                'degrees': (1, 0.0, 45.0),  # image rotation (+/- deg)
+                'translate': (1, 0.0, 0.9),  # image translation (+/- fraction)
+                'scale': (1, 0.0, 0.9),  # image scale (+/- gain)
+                'shear': (1, 0.0, 10.0),  # image shear (+/- deg)
+                'perspective': (0, 0.0, 0.001),  # image perspective (+/- fraction), range 0-0.001
+                'flipud': (1, 0.0, 1.0),  # image flip up-down (probability)
+                'fliplr': (0, 0.0, 1.0),  # image flip left-right (probability)
+                'mosaic': (1, 0.0, 1.0),  # image mixup (probability)
+                'mixup': (1, 0.0, 1.0)}  # image mixup (probability)
+
+        assert opt.local_rank == -1, 'DDP mode not implemented for --evolve'
+        opt.notest, opt.nosave = True, True  # only test/save final epoch
+        # ei = [isinstance(x, (int, float)) for x in hyp.values()]  # evolvable indices
+        yaml_file = Path(opt.save_dir) / 'hyp_evolved.yaml'  # save best result here
+        if opt.bucket:
+            os.system('gsutil cp gs://%s/evolve.txt .' % opt.bucket)  # download evolve.txt if exists
+
+        for _ in range(300):  # generations to evolve
+            if Path('evolve.txt').exists():  # if evolve.txt exists: select best hyps and mutate
+                # Select parent(s)
+                parent = 'single'  # parent selection method: 'single' or 'weighted'
+                x = np.loadtxt('evolve.txt', ndmin=2)
+                n = min(5, len(x))  # number of previous results to consider
+                x = x[np.argsort(-fitness(x))][:n]  # top n mutations
+                w = fitness(x) - fitness(x).min()  # weights
+                if parent == 'single' or len(x) == 1:
+                    # x = x[random.randint(0, n - 1)]  # random selection
+                    x = x[random.choices(range(n), weights=w)[0]]  # weighted selection
+                elif parent == 'weighted':
+                    x = (x * w.reshape(n, 1)).sum(0) / w.sum()  # weighted combination
+
+                # Mutate
+                mp, s = 0.8, 0.2  # mutation probability, sigma
+                npr = np.random
+                npr.seed(int(time.time()))
+                g = np.array([x[0] for x in meta.values()])  # gains 0-1
+                ng = len(meta)
+                v = np.ones(ng)
+                while all(v == 1):  # mutate until a change occurs (prevent duplicates)
+                    v = (g * (npr.random(ng) < mp) * npr.randn(ng) * npr.random() * s + 1).clip(0.3, 3.0)
+                for i, k in enumerate(hyp.keys()):  # plt.hist(v.ravel(), 300)
+                    hyp[k] = float(x[i + 7] * v[i])  # mutate
+
+            # Constrain to limits
+            for k, v in meta.items():
+                hyp[k] = max(hyp[k], v[1])  # lower limit
+                hyp[k] = min(hyp[k], v[2])  # upper limit
+                hyp[k] = round(hyp[k], 5)  # significant digits
+
+            # Train mutation
+            results = train(hyp.copy(), opt, device, wandb=wandb)
+
+            # Write mutation results
+            print_mutation(hyp.copy(), results, yaml_file, opt.bucket)
+
+        # Plot results
+        plot_evolution(yaml_file)
+        print(f'Hyperparameter evolution complete. Best results saved as: {yaml_file}\n'
+              f'Command to train a new model with these hyperparameters: $ python train.py --hyp {yaml_file}')

utils/__init__.py

@@ -0,0 +1 @@
+