├── LICENSE ├── README.md ├── checkpoints └── log.txt ├── datasets ├── images.txt ├── voc2007.txt ├── voc2007test.txt └── voc2012.txt ├── eval_voc.py ├── imgs ├── person.jpg └── person_result.jpg ├── models ├── __init__.py ├── __pycache__ │ ├── __init__.cpython-37.pyc │ └── resnet_yolo.cpython-37.pyc ├── resnet_yolo.py ├── vgg_yolo.py └── yoloLoss.py ├── predict.py ├── requirements.txt ├── train.py └── utils ├── __init__.py ├── dataset.py ├── piplist2equal.py └── xml2txt.py /LICENSE: -------------------------------------------------------------------------------- 1 | GNU GENERAL PUBLIC LICENSE 2 | Version 3, 29 June 2007 3 | 4 | Copyright (C) 2007 Free Software Foundation, Inc. 5 | Everyone is permitted to copy and distribute verbatim copies 6 | of this license document, but changing it is not allowed. 7 | 8 | Preamble 9 | 10 | The GNU General Public License is a free, copyleft license for 11 | software and other kinds of works. 12 | 13 | The licenses for most software and other practical works are designed 14 | to take away your freedom to share and change the works. By contrast, 15 | the GNU General Public License is intended to guarantee your freedom to 16 | share and change all versions of a program--to make sure it remains free 17 | software for all its users. We, the Free Software Foundation, use the 18 | GNU General Public License for most of our software; it applies also to 19 | any other work released this way by its authors. You can apply it to 20 | your programs, too. 21 | 22 | When we speak of free software, we are referring to freedom, not 23 | price. Our General Public Licenses are designed to make sure that you 24 | have the freedom to distribute copies of free software (and charge for 25 | them if you wish), that you receive source code or can get it if you 26 | want it, that you can change the software or use pieces of it in new 27 | free programs, and that you know you can do these things. 28 | 29 | To protect your rights, we need to prevent others from denying you 30 | these rights or asking you to surrender the rights. Therefore, you have 31 | certain responsibilities if you distribute copies of the software, or if 32 | you modify it: responsibilities to respect the freedom of others. 33 | 34 | For example, if you distribute copies of such a program, whether 35 | gratis or for a fee, you must pass on to the recipients the same 36 | freedoms that you received. You must make sure that they, too, receive 37 | or can get the source code. And you must show them these terms so they 38 | know their rights. 39 | 40 | Developers that use the GNU GPL protect your rights with two steps: 41 | (1) assert copyright on the software, and (2) offer you this License 42 | giving you legal permission to copy, distribute and/or modify it. 43 | 44 | For the developers' and authors' protection, the GPL clearly explains 45 | that there is no warranty for this free software. For both users' and 46 | authors' sake, the GPL requires that modified versions be marked as 47 | changed, so that their problems will not be attributed erroneously to 48 | authors of previous versions. 49 | 50 | Some devices are designed to deny users access to install or run 51 | modified versions of the software inside them, although the manufacturer 52 | can do so. This is fundamentally incompatible with the aim of 53 | protecting users' freedom to change the software. The systematic 54 | pattern of such abuse occurs in the area of products for individuals to 55 | use, which is precisely where it is most unacceptable. Therefore, we 56 | have designed this version of the GPL to prohibit the practice for those 57 | products. If such problems arise substantially in other domains, we 58 | stand ready to extend this provision to those domains in future versions 59 | of the GPL, as needed to protect the freedom of users. 60 | 61 | Finally, every program is threatened constantly by software patents. 62 | States should not allow patents to restrict development and use of 63 | software on general-purpose computers, but in those that do, we wish to 64 | avoid the special danger that patents applied to a free program could 65 | make it effectively proprietary. To prevent this, the GPL assures that 66 | patents cannot be used to render the program non-free. 67 | 68 | The precise terms and conditions for copying, distribution and 69 | modification follow. 70 | 71 | TERMS AND CONDITIONS 72 | 73 | 0. Definitions. 74 | 75 | "This License" refers to version 3 of the GNU General Public License. 76 | 77 | "Copyright" also means copyright-like laws that apply to other kinds of 78 | works, such as semiconductor masks. 79 | 80 | "The Program" refers to any copyrightable work licensed under this 81 | License. Each licensee is addressed as "you". "Licensees" and 82 | "recipients" may be individuals or organizations. 83 | 84 | To "modify" a work means to copy from or adapt all or part of the work 85 | in a fashion requiring copyright permission, other than the making of an 86 | exact copy. The resulting work is called a "modified version" of the 87 | earlier work or a work "based on" the earlier work. 88 | 89 | A "covered work" means either the unmodified Program or a work based 90 | on the Program. 91 | 92 | To "propagate" a work means to do anything with it that, without 93 | permission, would make you directly or secondarily liable for 94 | infringement under applicable copyright law, except executing it on a 95 | computer or modifying a private copy. Propagation includes copying, 96 | distribution (with or without modification), making available to the 97 | public, and in some countries other activities as well. 98 | 99 | To "convey" a work means any kind of propagation that enables other 100 | parties to make or receive copies. Mere interaction with a user through 101 | a computer network, with no transfer of a copy, is not conveying. 102 | 103 | An interactive user interface displays "Appropriate Legal Notices" 104 | to the extent that it includes a convenient and prominently visible 105 | feature that (1) displays an appropriate copyright notice, and (2) 106 | tells the user that there is no warranty for the work (except to the 107 | extent that warranties are provided), that licensees may convey the 108 | work under this License, and how to view a copy of this License. If 109 | the interface presents a list of user commands or options, such as a 110 | menu, a prominent item in the list meets this criterion. 111 | 112 | 1. Source Code. 113 | 114 | The "source code" for a work means the preferred form of the work 115 | for making modifications to it. "Object code" means any non-source 116 | form of a work. 117 | 118 | A "Standard Interface" means an interface that either is an official 119 | standard defined by a recognized standards body, or, in the case of 120 | interfaces specified for a particular programming language, one that 121 | is widely used among developers working in that language. 122 | 123 | The "System Libraries" of an executable work include anything, other 124 | than the work as a whole, that (a) is included in the normal form of 125 | packaging a Major Component, but which is not part of that Major 126 | Component, and (b) serves only to enable use of the work with that 127 | Major Component, or to implement a Standard Interface for which an 128 | implementation is available to the public in source code form. A 129 | "Major Component", in this context, means a major essential component 130 | (kernel, window system, and so on) of the specific operating system 131 | (if any) on which the executable work runs, or a compiler used to 132 | produce the work, or an object code interpreter used to run it. 133 | 134 | The "Corresponding Source" for a work in object code form means all 135 | the source code needed to generate, install, and (for an executable 136 | work) run the object code and to modify the work, including scripts to 137 | control those activities. However, it does not include the work's 138 | System Libraries, or general-purpose tools or generally available free 139 | programs which are used unmodified in performing those activities but 140 | which are not part of the work. For example, Corresponding Source 141 | includes interface definition files associated with source files for 142 | the work, and the source code for shared libraries and dynamically 143 | linked subprograms that the work is specifically designed to require, 144 | such as by intimate data communication or control flow between those 145 | subprograms and other parts of the work. 146 | 147 | The Corresponding Source need not include anything that users 148 | can regenerate automatically from other parts of the Corresponding 149 | Source. 150 | 151 | The Corresponding Source for a work in source code form is that 152 | same work. 153 | 154 | 2. Basic Permissions. 155 | 156 | All rights granted under this License are granted for the term of 157 | copyright on the Program, and are irrevocable provided the stated 158 | conditions are met. This License explicitly affirms your unlimited 159 | permission to run the unmodified Program. The output from running a 160 | covered work is covered by this License only if the output, given its 161 | content, constitutes a covered work. This License acknowledges your 162 | rights of fair use or other equivalent, as provided by copyright law. 163 | 164 | You may make, run and propagate covered works that you do not 165 | convey, without conditions so long as your license otherwise remains 166 | in force. You may convey covered works to others for the sole purpose 167 | of having them make modifications exclusively for you, or provide you 168 | with facilities for running those works, provided that you comply with 169 | the terms of this License in conveying all material for which you do 170 | not control copyright. Those thus making or running the covered works 171 | for you must do so exclusively on your behalf, under your direction 172 | and control, on terms that prohibit them from making any copies of 173 | your copyrighted material outside their relationship with you. 174 | 175 | Conveying under any other circumstances is permitted solely under 176 | the conditions stated below. Sublicensing is not allowed; section 10 177 | makes it unnecessary. 178 | 179 | 3. Protecting Users' Legal Rights From Anti-Circumvention Law. 180 | 181 | No covered work shall be deemed part of an effective technological 182 | measure under any applicable law fulfilling obligations under article 183 | 11 of the WIPO copyright treaty adopted on 20 December 1996, or 184 | similar laws prohibiting or restricting circumvention of such 185 | measures. 186 | 187 | When you convey a covered work, you waive any legal power to forbid 188 | circumvention of technological measures to the extent such circumvention 189 | is effected by exercising rights under this License with respect to 190 | the covered work, and you disclaim any intention to limit operation or 191 | modification of the work as a means of enforcing, against the work's 192 | users, your or third parties' legal rights to forbid circumvention of 193 | technological measures. 194 | 195 | 4. Conveying Verbatim Copies. 196 | 197 | You may convey verbatim copies of the Program's source code as you 198 | receive it, in any medium, provided that you conspicuously and 199 | appropriately publish on each copy an appropriate copyright notice; 200 | keep intact all notices stating that this License and any 201 | non-permissive terms added in accord with section 7 apply to the code; 202 | keep intact all notices of the absence of any warranty; and give all 203 | recipients a copy of this License along with the Program. 204 | 205 | You may charge any price or no price for each copy that you convey, 206 | and you may offer support or warranty protection for a fee. 207 | 208 | 5. Conveying Modified Source Versions. 209 | 210 | You may convey a work based on the Program, or the modifications to 211 | produce it from the Program, in the form of source code under the 212 | terms of section 4, provided that you also meet all of these conditions: 213 | 214 | a) The work must carry prominent notices stating that you modified 215 | it, and giving a relevant date. 216 | 217 | b) The work must carry prominent notices stating that it is 218 | released under this License and any conditions added under section 219 | 7. This requirement modifies the requirement in section 4 to 220 | "keep intact all notices". 221 | 222 | c) You must license the entire work, as a whole, under this 223 | License to anyone who comes into possession of a copy. This 224 | License will therefore apply, along with any applicable section 7 225 | additional terms, to the whole of the work, and all its parts, 226 | regardless of how they are packaged. This License gives no 227 | permission to license the work in any other way, but it does not 228 | invalidate such permission if you have separately received it. 229 | 230 | d) If the work has interactive user interfaces, each must display 231 | Appropriate Legal Notices; however, if the Program has interactive 232 | interfaces that do not display Appropriate Legal Notices, your 233 | work need not make them do so. 234 | 235 | A compilation of a covered work with other separate and independent 236 | works, which are not by their nature extensions of the covered work, 237 | and which are not combined with it such as to form a larger program, 238 | in or on a volume of a storage or distribution medium, is called an 239 | "aggregate" if the compilation and its resulting copyright are not 240 | used to limit the access or legal rights of the compilation's users 241 | beyond what the individual works permit. Inclusion of a covered work 242 | in an aggregate does not cause this License to apply to the other 243 | parts of the aggregate. 244 | 245 | 6. Conveying Non-Source Forms. 246 | 247 | You may convey a covered work in object code form under the terms 248 | of sections 4 and 5, provided that you also convey the 249 | machine-readable Corresponding Source under the terms of this License, 250 | in one of these ways: 251 | 252 | a) Convey the object code in, or embodied in, a physical product 253 | (including a physical distribution medium), accompanied by the 254 | Corresponding Source fixed on a durable physical medium 255 | customarily used for software interchange. 256 | 257 | b) Convey the object code in, or embodied in, a physical product 258 | (including a physical distribution medium), accompanied by a 259 | written offer, valid for at least three years and valid for as 260 | long as you offer spare parts or customer support for that product 261 | model, to give anyone who possesses the object code either (1) a 262 | copy of the Corresponding Source for all the software in the 263 | product that is covered by this License, on a durable physical 264 | medium customarily used for software interchange, for a price no 265 | more than your reasonable cost of physically performing this 266 | conveying of source, or (2) access to copy the 267 | Corresponding Source from a network server at no charge. 268 | 269 | c) Convey individual copies of the object code with a copy of the 270 | written offer to provide the Corresponding Source. This 271 | alternative is allowed only occasionally and noncommercially, and 272 | only if you received the object code with such an offer, in accord 273 | with subsection 6b. 274 | 275 | d) Convey the object code by offering access from a designated 276 | place (gratis or for a charge), and offer equivalent access to the 277 | Corresponding Source in the same way through the same place at no 278 | further charge. You need not require recipients to copy the 279 | Corresponding Source along with the object code. If the place to 280 | copy the object code is a network server, the Corresponding Source 281 | may be on a different server (operated by you or a third party) 282 | that supports equivalent copying facilities, provided you maintain 283 | clear directions next to the object code saying where to find the 284 | Corresponding Source. Regardless of what server hosts the 285 | Corresponding Source, you remain obligated to ensure that it is 286 | available for as long as needed to satisfy these requirements. 287 | 288 | e) Convey the object code using peer-to-peer transmission, provided 289 | you inform other peers where the object code and Corresponding 290 | Source of the work are being offered to the general public at no 291 | charge under subsection 6d. 292 | 293 | A separable portion of the object code, whose source code is excluded 294 | from the Corresponding Source as a System Library, need not be 295 | included in conveying the object code work. 296 | 297 | A "User Product" is either (1) a "consumer product", which means any 298 | tangible personal property which is normally used for personal, family, 299 | or household purposes, or (2) anything designed or sold for incorporation 300 | into a dwelling. In determining whether a product is a consumer product, 301 | doubtful cases shall be resolved in favor of coverage. For a particular 302 | product received by a particular user, "normally used" refers to a 303 | typical or common use of that class of product, regardless of the status 304 | of the particular user or of the way in which the particular user 305 | actually uses, or expects or is expected to use, the product. A product 306 | is a consumer product regardless of whether the product has substantial 307 | commercial, industrial or non-consumer uses, unless such uses represent 308 | the only significant mode of use of the product. 309 | 310 | "Installation Information" for a User Product means any methods, 311 | procedures, authorization keys, or other information required to install 312 | and execute modified versions of a covered work in that User Product from 313 | a modified version of its Corresponding Source. The information must 314 | suffice to ensure that the continued functioning of the modified object 315 | code is in no case prevented or interfered with solely because 316 | modification has been made. 317 | 318 | If you convey an object code work under this section in, or with, or 319 | specifically for use in, a User Product, and the conveying occurs as 320 | part of a transaction in which the right of possession and use of the 321 | User Product is transferred to the recipient in perpetuity or for a 322 | fixed term (regardless of how the transaction is characterized), the 323 | Corresponding Source conveyed under this section must be accompanied 324 | by the Installation Information. But this requirement does not apply 325 | if neither you nor any third party retains the ability to install 326 | modified object code on the User Product (for example, the work has 327 | been installed in ROM). 328 | 329 | The requirement to provide Installation Information does not include a 330 | requirement to continue to provide support service, warranty, or updates 331 | for a work that has been modified or installed by the recipient, or for 332 | the User Product in which it has been modified or installed. Access to a 333 | network may be denied when the modification itself materially and 334 | adversely affects the operation of the network or violates the rules and 335 | protocols for communication across the network. 336 | 337 | Corresponding Source conveyed, and Installation Information provided, 338 | in accord with this section must be in a format that is publicly 339 | documented (and with an implementation available to the public in 340 | source code form), and must require no special password or key for 341 | unpacking, reading or copying. 342 | 343 | 7. Additional Terms. 344 | 345 | "Additional permissions" are terms that supplement the terms of this 346 | License by making exceptions from one or more of its conditions. 347 | Additional permissions that are applicable to the entire Program shall 348 | be treated as though they were included in this License, to the extent 349 | that they are valid under applicable law. If additional permissions 350 | apply only to part of the Program, that part may be used separately 351 | under those permissions, but the entire Program remains governed by 352 | this License without regard to the additional permissions. 353 | 354 | When you convey a copy of a covered work, you may at your option 355 | remove any additional permissions from that copy, or from any part of 356 | it. (Additional permissions may be written to require their own 357 | removal in certain cases when you modify the work.) You may place 358 | additional permissions on material, added by you to a covered work, 359 | for which you have or can give appropriate copyright permission. 360 | 361 | Notwithstanding any other provision of this License, for material you 362 | add to a covered work, you may (if authorized by the copyright holders of 363 | that material) supplement the terms of this License with terms: 364 | 365 | a) Disclaiming warranty or limiting liability differently from the 366 | terms of sections 15 and 16 of this License; or 367 | 368 | b) Requiring preservation of specified reasonable legal notices or 369 | author attributions in that material or in the Appropriate Legal 370 | Notices displayed by works containing it; or 371 | 372 | c) Prohibiting misrepresentation of the origin of that material, or 373 | requiring that modified versions of such material be marked in 374 | reasonable ways as different from the original version; or 375 | 376 | d) Limiting the use for publicity purposes of names of licensors or 377 | authors of the material; or 378 | 379 | e) Declining to grant rights under trademark law for use of some 380 | trade names, trademarks, or service marks; or 381 | 382 | f) Requiring indemnification of licensors and authors of that 383 | material by anyone who conveys the material (or modified versions of 384 | it) with contractual assumptions of liability to the recipient, for 385 | any liability that these contractual assumptions directly impose on 386 | those licensors and authors. 387 | 388 | All other non-permissive additional terms are considered "further 389 | restrictions" within the meaning of section 10. If the Program as you 390 | received it, or any part of it, contains a notice stating that it is 391 | governed by this License along with a term that is a further 392 | restriction, you may remove that term. If a license document contains 393 | a further restriction but permits relicensing or conveying under this 394 | License, you may add to a covered work material governed by the terms 395 | of that license document, provided that the further restriction does 396 | not survive such relicensing or conveying. 397 | 398 | If you add terms to a covered work in accord with this section, you 399 | must place, in the relevant source files, a statement of the 400 | additional terms that apply to those files, or a notice indicating 401 | where to find the applicable terms. 402 | 403 | Additional terms, permissive or non-permissive, may be stated in the 404 | form of a separately written license, or stated as exceptions; 405 | the above requirements apply either way. 406 | 407 | 8. Termination. 408 | 409 | You may not propagate or modify a covered work except as expressly 410 | provided under this License. Any attempt otherwise to propagate or 411 | modify it is void, and will automatically terminate your rights under 412 | this License (including any patent licenses granted under the third 413 | paragraph of section 11). 414 | 415 | However, if you cease all violation of this License, then your 416 | license from a particular copyright holder is reinstated (a) 417 | provisionally, unless and until the copyright holder explicitly and 418 | finally terminates your license, and (b) permanently, if the copyright 419 | holder fails to notify you of the violation by some reasonable means 420 | prior to 60 days after the cessation. 421 | 422 | Moreover, your license from a particular copyright holder is 423 | reinstated permanently if the copyright holder notifies you of the 424 | violation by some reasonable means, this is the first time you have 425 | received notice of violation of this License (for any work) from that 426 | copyright holder, and you cure the violation prior to 30 days after 427 | your receipt of the notice. 428 | 429 | Termination of your rights under this section does not terminate the 430 | licenses of parties who have received copies or rights from you under 431 | this License. If your rights have been terminated and not permanently 432 | reinstated, you do not qualify to receive new licenses for the same 433 | material under section 10. 434 | 435 | 9. Acceptance Not Required for Having Copies. 436 | 437 | You are not required to accept this License in order to receive or 438 | run a copy of the Program. Ancillary propagation of a covered work 439 | occurring solely as a consequence of using peer-to-peer transmission 440 | to receive a copy likewise does not require acceptance. However, 441 | nothing other than this License grants you permission to propagate or 442 | modify any covered work. These actions infringe copyright if you do 443 | not accept this License. Therefore, by modifying or propagating a 444 | covered work, you indicate your acceptance of this License to do so. 445 | 446 | 10. Automatic Licensing of Downstream Recipients. 447 | 448 | Each time you convey a covered work, the recipient automatically 449 | receives a license from the original licensors, to run, modify and 450 | propagate that work, subject to this License. You are not responsible 451 | for enforcing compliance by third parties with this License. 452 | 453 | An "entity transaction" is a transaction transferring control of an 454 | organization, or substantially all assets of one, or subdividing an 455 | organization, or merging organizations. If propagation of a covered 456 | work results from an entity transaction, each party to that 457 | transaction who receives a copy of the work also receives whatever 458 | licenses to the work the party's predecessor in interest had or could 459 | give under the previous paragraph, plus a right to possession of the 460 | Corresponding Source of the work from the predecessor in interest, if 461 | the predecessor has it or can get it with reasonable efforts. 462 | 463 | You may not impose any further restrictions on the exercise of the 464 | rights granted or affirmed under this License. For example, you may 465 | not impose a license fee, royalty, or other charge for exercise of 466 | rights granted under this License, and you may not initiate litigation 467 | (including a cross-claim or counterclaim in a lawsuit) alleging that 468 | any patent claim is infringed by making, using, selling, offering for 469 | sale, or importing the Program or any portion of it. 470 | 471 | 11. Patents. 472 | 473 | A "contributor" is a copyright holder who authorizes use under this 474 | License of the Program or a work on which the Program is based. The 475 | work thus licensed is called the contributor's "contributor version". 476 | 477 | A contributor's "essential patent claims" are all patent claims 478 | owned or controlled by the contributor, whether already acquired or 479 | hereafter acquired, that would be infringed by some manner, permitted 480 | by this License, of making, using, or selling its contributor version, 481 | but do not include claims that would be infringed only as a 482 | consequence of further modification of the contributor version. For 483 | purposes of this definition, "control" includes the right to grant 484 | patent sublicenses in a manner consistent with the requirements of 485 | this License. 486 | 487 | Each contributor grants you a non-exclusive, worldwide, royalty-free 488 | patent license under the contributor's essential patent claims, to 489 | make, use, sell, offer for sale, import and otherwise run, modify and 490 | propagate the contents of its contributor version. 491 | 492 | In the following three paragraphs, a "patent license" is any express 493 | agreement or commitment, however denominated, not to enforce a patent 494 | (such as an express permission to practice a patent or covenant not to 495 | sue for patent infringement). To "grant" such a patent license to a 496 | party means to make such an agreement or commitment not to enforce a 497 | patent against the party. 498 | 499 | If you convey a covered work, knowingly relying on a patent license, 500 | and the Corresponding Source of the work is not available for anyone 501 | to copy, free of charge and under the terms of this License, through a 502 | publicly available network server or other readily accessible means, 503 | then you must either (1) cause the Corresponding Source to be so 504 | available, or (2) arrange to deprive yourself of the benefit of the 505 | patent license for this particular work, or (3) arrange, in a manner 506 | consistent with the requirements of this License, to extend the patent 507 | license to downstream recipients. "Knowingly relying" means you have 508 | actual knowledge that, but for the patent license, your conveying the 509 | covered work in a country, or your recipient's use of the covered work 510 | in a country, would infringe one or more identifiable patents in that 511 | country that you have reason to believe are valid. 512 | 513 | If, pursuant to or in connection with a single transaction or 514 | arrangement, you convey, or propagate by procuring conveyance of, a 515 | covered work, and grant a patent license to some of the parties 516 | receiving the covered work authorizing them to use, propagate, modify 517 | or convey a specific copy of the covered work, then the patent license 518 | you grant is automatically extended to all recipients of the covered 519 | work and works based on it. 520 | 521 | A patent license is "discriminatory" if it does not include within 522 | the scope of its coverage, prohibits the exercise of, or is 523 | conditioned on the non-exercise of one or more of the rights that are 524 | specifically granted under this License. You may not convey a covered 525 | work if you are a party to an arrangement with a third party that is 526 | in the business of distributing software, under which you make payment 527 | to the third party based on the extent of your activity of conveying 528 | the work, and under which the third party grants, to any of the 529 | parties who would receive the covered work from you, a discriminatory 530 | patent license (a) in connection with copies of the covered work 531 | conveyed by you (or copies made from those copies), or (b) primarily 532 | for and in connection with specific products or compilations that 533 | contain the covered work, unless you entered into that arrangement, 534 | or that patent license was granted, prior to 28 March 2007. 535 | 536 | Nothing in this License shall be construed as excluding or limiting 537 | any implied license or other defenses to infringement that may 538 | otherwise be available to you under applicable patent law. 539 | 540 | 12. No Surrender of Others' Freedom. 541 | 542 | If conditions are imposed on you (whether by court order, agreement or 543 | otherwise) that contradict the conditions of this License, they do not 544 | excuse you from the conditions of this License. If you cannot convey a 545 | covered work so as to satisfy simultaneously your obligations under this 546 | License and any other pertinent obligations, then as a consequence you may 547 | not convey it at all. For example, if you agree to terms that obligate you 548 | to collect a royalty for further conveying from those to whom you convey 549 | the Program, the only way you could satisfy both those terms and this 550 | License would be to refrain entirely from conveying the Program. 551 | 552 | 13. Use with the GNU Affero General Public License. 553 | 554 | Notwithstanding any other provision of this License, you have 555 | permission to link or combine any covered work with a work licensed 556 | under version 3 of the GNU Affero General Public License into a single 557 | combined work, and to convey the resulting work. The terms of this 558 | License will continue to apply to the part which is the covered work, 559 | but the special requirements of the GNU Affero General Public License, 560 | section 13, concerning interaction through a network will apply to the 561 | combination as such. 562 | 563 | 14. Revised Versions of this License. 564 | 565 | The Free Software Foundation may publish revised and/or new versions of 566 | the GNU General Public License from time to time. Such new versions will 567 | be similar in spirit to the present version, but may differ in detail to 568 | address new problems or concerns. 569 | 570 | Each version is given a distinguishing version number. If the 571 | Program specifies that a certain numbered version of the GNU General 572 | Public License "or any later version" applies to it, you have the 573 | option of following the terms and conditions either of that numbered 574 | version or of any later version published by the Free Software 575 | Foundation. If the Program does not specify a version number of the 576 | GNU General Public License, you may choose any version ever published 577 | by the Free Software Foundation. 578 | 579 | If the Program specifies that a proxy can decide which future 580 | versions of the GNU General Public License can be used, that proxy's 581 | public statement of acceptance of a version permanently authorizes you 582 | to choose that version for the Program. 583 | 584 | Later license versions may give you additional or different 585 | permissions. However, no additional obligations are imposed on any 586 | author or copyright holder as a result of your choosing to follow a 587 | later version. 588 | 589 | 15. Disclaimer of Warranty. 590 | 591 | THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY 592 | APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT 593 | HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY 594 | OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, 595 | THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 596 | PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM 597 | IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF 598 | ALL NECESSARY SERVICING, REPAIR OR CORRECTION. 599 | 600 | 16. Limitation of Liability. 601 | 602 | IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING 603 | WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS 604 | THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY 605 | GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE 606 | USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF 607 | DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD 608 | PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), 609 | EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF 610 | SUCH DAMAGES. 611 | 612 | 17. Interpretation of Sections 15 and 16. 613 | 614 | If the disclaimer of warranty and limitation of liability provided 615 | above cannot be given local legal effect according to their terms, 616 | reviewing courts shall apply local law that most closely approximates 617 | an absolute waiver of all civil liability in connection with the 618 | Program, unless a warranty or assumption of liability accompanies a 619 | copy of the Program in return for a fee. 620 | 621 | END OF TERMS AND CONDITIONS 622 | 623 | How to Apply These Terms to Your New Programs 624 | 625 | If you develop a new program, and you want it to be of the greatest 626 | possible use to the public, the best way to achieve this is to make it 627 | free software which everyone can redistribute and change under these terms. 628 | 629 | To do so, attach the following notices to the program. It is safest 630 | to attach them to the start of each source file to most effectively 631 | state the exclusion of warranty; and each file should have at least 632 | the "copyright" line and a pointer to where the full notice is found. 633 | 634 | 635 | Copyright (C) 636 | 637 | This program is free software: you can redistribute it and/or modify 638 | it under the terms of the GNU General Public License as published by 639 | the Free Software Foundation, either version 3 of the License, or 640 | (at your option) any later version. 641 | 642 | This program is distributed in the hope that it will be useful, 643 | but WITHOUT ANY WARRANTY; without even the implied warranty of 644 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 645 | GNU General Public License for more details. 646 | 647 | You should have received a copy of the GNU General Public License 648 | along with this program. If not, see . 649 | 650 | Also add information on how to contact you by electronic and paper mail. 651 | 652 | If the program does terminal interaction, make it output a short 653 | notice like this when it starts in an interactive mode: 654 | 655 | Copyright (C) 656 | This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'. 657 | This is free software, and you are welcome to redistribute it 658 | under certain conditions; type `show c' for details. 659 | 660 | The hypothetical commands `show w' and `show c' should show the appropriate 661 | parts of the General Public License. Of course, your program's commands 662 | might be different; for a GUI interface, you would use an "about box". 663 | 664 | You should also get your employer (if you work as a programmer) or school, 665 | if any, to sign a "copyright disclaimer" for the program, if necessary. 666 | For more information on this, and how to apply and follow the GNU GPL, see 667 | . 668 | 669 | The GNU General Public License does not permit incorporating your program 670 | into proprietary programs. If your program is a subroutine library, you 671 | may consider it more useful to permit linking proprietary applications with 672 | the library. If this is what you want to do, use the GNU Lesser General 673 | Public License instead of this License. But first, please read 674 | . 675 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # yolo v1_pytorch 2 | ## 项目目录 3 | ```text 4 | ｜--checkpoints 日志和权重 5 | |--log.txt 6 | |--best.pth 7 | ｜--datasets 数据集 8 | |--images.txt 9 | |--voc2007.txt 10 | |--voc2012.txt 11 | |--voc2007test.txt 12 | |--images/ 13 | |--XXXX.jpg 14 | |--... 15 | ｜--imgs 测试图片 16 | |--person.jpg 17 | ｜--models 模型 18 | |--resnet_yolo.py 19 | |--vgg_yolo.py 20 | |--yoloLoss.py 21 | ｜--utils 工具 22 | |--dataset.py 23 | |--piplist2equal.py 24 | |--xml2txt.py 25 | ｜--train.py 26 | ｜--predict.py 27 | ｜--eval_voc.py 28 | ｜--requirements.txt 环境 29 | 30 | 31 | ``` 32 | 33 | ## 1. 环境准备 34 | 详细内容参考[requirements.txt](requirements.txt) 35 | 36 | ## 2. 数据集准备 37 | **下载数据集** 38 | 链接: https://pan.baidu.com/s/1hturxvztlt_ePnZt3TTzWQ 密码: 6qgn 39 | 40 | **解压数据集** 41 | 1、将voc2007和voc2012的所有图片放到`datasets/images`目录下。 42 | 43 | 2、然后使用`utils/xml2txt.pyt`将xml文件转成txt格式。此步骤结果已经保存，可省略。 44 | 45 | ## 3.训练 46 | ```shell 47 | python train.py 48 | ``` 49 | ## 4.预测 50 | ```text 51 | python predict.py 52 | ``` 53 | ## 5.计算mAP 54 | ```text 55 | python eval_voc.py 56 | ``` 57 | 58 | ![](imgs/person.jpg)![](imgs/person_result.jpg) 59 | 60 | ## 附 61 | 对于初学者或硬件条件限制问题，提供训练后的权重.pth文件。 62 | 63 | 【获取方式】 64 | 关注公众号 【OAOA】，回复【0813】获取百度盘链接。 65 | 66 | ## 参考 67 | 原项目地址：https://github.com/FelixFu520/yolov1 -------------------------------------------------------------------------------- /checkpoints/log.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/isbrycee/yolov1_pytorch/ecb0ef944ec38fa5b887263d61654f714ca07108/checkpoints/log.txt -------------------------------------------------------------------------------- /eval_voc.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | """ 3 | @Time : 2020/08/12 18:30 4 | @Author : Bryce 5 | @File : eval_voc.py 6 | @Noice : 7 | @Modificattion : 8 | @Author : 9 | @Time : 10 | @Detail : 11 | """ 12 | import os 13 | os.environ["CUDA_VISIBLE_DEVICES"] = "0" 14 | import numpy as np 15 | 16 | 17 | VOC_CLASSES = ( # always index 0 18 | 'aeroplane', 'bicycle', 'bird', 'boat', 19 | 'bottle', 'bus', 'car', 'cat', 'chair', 20 | 'cow', 'diningtable', 'dog', 'horse', 21 | 'motorbike', 'person', 'pottedplant', 22 | 'sheep', 'sofa', 'train', 'tvmonitor') 23 | Color = [ 24 | [0, 0, 0], 25 | [128, 0, 0], 26 | [0, 128, 0], 27 | [128, 128, 0], 28 | [0, 0, 128], 29 | [128, 0, 128], 30 | [0, 128, 128], 31 | [128, 128, 128], 32 | [64, 0, 0], 33 | [192, 0, 0], 34 | [64, 128, 0], 35 | [192, 128, 0], 36 | [64, 0, 128], 37 | [192, 0, 128], 38 | [64, 128, 128], 39 | [192, 128, 128], 40 | [0, 64, 0], 41 | [128, 64, 0], 42 | [0, 192, 0], 43 | [128, 192, 0], 44 | [0, 64, 128] 45 | ] 46 | 47 | 48 | def voc_ap(rec, prec, use_07_metric=False): 49 | if use_07_metric: 50 | # 11 point metric 51 | ap = 0. 52 | for t in np.arange(0., 1.1, 0.1): 53 | if np.sum(rec >= t) == 0: 54 | p = 0 55 | else: 56 | p = np.max(prec[rec >= t]) 57 | ap = ap + p/11. 58 | else: 59 | # correct ap caculation 60 | mrec = np.concatenate(([0.], rec, [1.])) 61 | mpre = np.concatenate(([0.], prec, [0.])) 62 | 63 | for i in range(mpre.size - 1, 0, -1): 64 | mpre[i-1] = np.maximum(mpre[i-1], mpre[i]) 65 | 66 | i = np.where(mrec[1:] != mrec[:-1])[0] 67 | 68 | ap = np.sum((mrec[i + 1] - mrec[i]) * mpre[i + 1]) 69 | 70 | return ap 71 | 72 | 73 | def voc_eval(preds, target, VOC_CLASSES=VOC_CLASSES, threshold=0.5, use_07_metric=False,): 74 | ''' 75 | preds {'cat':[[image_id,confidence,x1,y1,x2,y2],...],'dog':[[],...]} 76 | target {(image_id,class):[[],]} 77 | ''' 78 | aps = [] 79 | for i,class_ in enumerate(VOC_CLASSES): 80 | pred = preds[class_] # [[image_id,confidence,x1,y1,x2,y2],...] 81 | if len(pred) == 0: # 如果这个类别一个都没有检测到的异常情况 82 | ap = -1 83 | print('---class {} ap {}---'.format(class_,ap)) 84 | aps += [ap] 85 | break 86 | #print(pred) 87 | image_ids = [x[0] for x in pred] 88 | confidence = np.array([float(x[1]) for x in pred]) 89 | BB = np.array([x[2:] for x in pred]) 90 | # sort by confidence 91 | sorted_ind = np.argsort(-confidence) 92 | sorted_scores = np.sort(-confidence) 93 | BB = BB[sorted_ind, :] 94 | image_ids = [image_ids[x] for x in sorted_ind] 95 | 96 | # go down dets and mark TPs and FPs 97 | npos = 0. 98 | for (key1,key2) in target: 99 | if key2 == class_: 100 | npos += len(target[(key1,key2)]) #统计这个类别的正样本，在这里统计才不会遗漏 101 | nd = len(image_ids) 102 | tp = np.zeros(nd) 103 | fp = np.zeros(nd) 104 | for d,image_id in enumerate(image_ids): 105 | bb = BB[d] #预测框 106 | if (image_id,class_) in target: 107 | BBGT = target[(image_id,class_)] #[[],] 108 | for bbgt in BBGT: 109 | # compute overlaps 110 | # intersection 111 | ixmin = np.maximum(bbgt[0], bb[0]) 112 | iymin = np.maximum(bbgt[1], bb[1]) 113 | ixmax = np.minimum(bbgt[2], bb[2]) 114 | iymax = np.minimum(bbgt[3], bb[3]) 115 | iw = np.maximum(ixmax - ixmin + 1., 0.) 116 | ih = np.maximum(iymax - iymin + 1., 0.) 117 | inters = iw * ih 118 | 119 | union = (bb[2]-bb[0]+1.)*(bb[3]-bb[1]+1.) + (bbgt[2]-bbgt[0]+1.)*(bbgt[3]-bbgt[1]+1.) - inters 120 | if union == 0: 121 | print(bb,bbgt) 122 | 123 | overlaps = inters/union 124 | if overlaps > threshold: 125 | tp[d] = 1 126 | BBGT.remove(bbgt) #这个框已经匹配到了，不能再匹配 127 | if len(BBGT) == 0: 128 | del target[(image_id,class_)] #删除没有box的键值 129 | break 130 | fp[d] = 1-tp[d] 131 | else: 132 | fp[d] = 1 133 | fp = np.cumsum(fp) 134 | tp = np.cumsum(tp) 135 | rec = tp/float(npos) 136 | prec = tp/np.maximum(tp + fp, np.finfo(np.float64).eps) 137 | #print(rec,prec) 138 | ap = voc_ap(rec, prec, use_07_metric) 139 | print('---class {} ap {}---'.format(class_,ap)) 140 | aps += [ap] 141 | print('---map {}---'.format(np.mean(aps))) 142 | 143 | 144 | def test_eval(): 145 | preds = {'cat':[['image01',0.9,20,20,40,40],['image01',0.8,20,20,50,50],['image02',0.8,30,30,50,50]],'dog':[['image01',0.78,60,60,90,90]]} 146 | target = {('image01','cat'):[[20,20,41,41]],('image01','dog'):[[60,60,91,91]],('image02','cat'):[[30,30,51,51]]} 147 | voc_eval(preds,target,VOC_CLASSES=['cat','dog']) 148 | 149 | 150 | if __name__ == '__main__': 151 | #test_eval() 152 | from predict import * 153 | from collections import defaultdict 154 | from tqdm import tqdm 155 | 156 | target = defaultdict(list) 157 | preds = defaultdict(list) 158 | image_list = [] #image path list 159 | 160 | f = open('datasets/voc2007test.txt') 161 | lines = f.readlines() 162 | file_list = [] 163 | for line in lines: 164 | splited = line.strip().split() 165 | file_list.append(splited) 166 | f.close() 167 | print('---prepare target---') 168 | for index,image_file in enumerate(file_list): 169 | image_id = image_file[0] 170 | 171 | image_list.append(image_id) 172 | num_obj = (len(image_file) - 1) // 5 173 | for i in range(num_obj): 174 | x1 = int(image_file[1+5*i]) 175 | y1 = int(image_file[2+5*i]) 176 | x2 = int(image_file[3+5*i]) 177 | y2 = int(image_file[4+5*i]) 178 | c = int(image_file[5+5*i]) 179 | class_name = VOC_CLASSES[c] 180 | target[(image_id,class_name)].append([x1,y1,x2,y2]) 181 | # 182 | #start test 183 | # 184 | print('---start test---') 185 | # model = vgg16_bn(pretrained=False) 186 | model = resnet50() 187 | # model.classifier = nn.Sequential( 188 | # nn.Linear(512 * 7 * 7, 4096), 189 | # nn.ReLU(True), 190 | # nn.Dropout(), 191 | # #nn.Linear(4096, 4096), 192 | # #nn.ReLU(True), 193 | # #nn.Dropout(), 194 | # nn.Linear(4096, 1470), 195 | # ) 196 | model.load_state_dict(torch.load('checkpoints/best.pth')) 197 | model.eval() 198 | model.cuda() 199 | count = 0 200 | for image_path in tqdm(image_list): 201 | result = predict_gpu(model,image_path,root_path='datasets/images/') #result[[left_up,right_bottom,class_name,image_path],] 202 | for (x1,y1),(x2,y2),class_name,image_id,prob in result: #image_id is actually image_path 203 | preds[class_name].append([image_id,prob,x1,y1,x2,y2]) 204 | # print(image_path) 205 | # image = cv2.imread('/home/xzh/data/VOCdevkit/VOC2012/allimgs/'+image_path) 206 | # for left_up,right_bottom,class_name,_,prob in result: 207 | # color = Color[VOC_CLASSES.index(class_name)] 208 | # cv2.rectangle(image,left_up,right_bottom,color,2) 209 | # label = class_name+str(round(prob,2)) 210 | # text_size, baseline = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX, 0.4, 1) 211 | # p1 = (left_up[0], left_up[1]- text_size[1]) 212 | # cv2.rectangle(image, (p1[0] - 2//2, p1[1] - 2 - baseline), (p1[0] + text_size[0], p1[1] + text_size[1]), color, -1) 213 | # cv2.putText(image, label, (p1[0], p1[1] + baseline), cv2.FONT_HERSHEY_SIMPLEX, 0.4, (255,255,255), 1, 8) 214 | 215 | # cv2.imwrite('testimg/'+image_path,image) 216 | # count += 1 217 | # if count == 100: 218 | # break 219 | 220 | print('---start evaluate---') 221 | voc_eval(preds,target,VOC_CLASSES=VOC_CLASSES) -------------------------------------------------------------------------------- /imgs/person.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/isbrycee/yolov1_pytorch/ecb0ef944ec38fa5b887263d61654f714ca07108/imgs/person.jpg -------------------------------------------------------------------------------- /imgs/person_result.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/isbrycee/yolov1_pytorch/ecb0ef944ec38fa5b887263d61654f714ca07108/imgs/person_result.jpg -------------------------------------------------------------------------------- /models/__init__.py: -------------------------------------------------------------------------------- 1 | # !/usr/bin/python 2 | # -*- coding:utf-8 -*- 3 | # author: Felix Fu -------------------------------------------------------------------------------- /models/__pycache__/__init__.cpython-37.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/isbrycee/yolov1_pytorch/ecb0ef944ec38fa5b887263d61654f714ca07108/models/__pycache__/__init__.cpython-37.pyc -------------------------------------------------------------------------------- /models/__pycache__/resnet_yolo.cpython-37.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/isbrycee/yolov1_pytorch/ecb0ef944ec38fa5b887263d61654f714ca07108/models/__pycache__/resnet_yolo.cpython-37.pyc -------------------------------------------------------------------------------- /models/resnet_yolo.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | """ 3 | @Time : 2020/08/12 18:30 4 | @Author : Bryce 5 | @File : resnet_yolo.py 6 | @Noice : 7 | @Modificattion : 8 | @Author : 9 | @Time : 10 | @Detail : 11 | """ 12 | 13 | import torch.nn as nn 14 | import math 15 | import torch.utils.model_zoo as model_zoo 16 | import torch.nn.functional as F 17 | import torch 18 | 19 | 20 | __all__ = ['ResNet', 'resnet18', 'resnet34', 'resnet50', 'resnet101', 'resnet152'] 21 | 22 | 23 | model_urls = { 24 | 'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth', 25 | 'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth', 26 | 'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth', 27 | 'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth', 28 | 'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth', 29 | } 30 | 31 | 32 | def conv3x3(in_planes, out_planes, stride=1): 33 | """3x3 convolution with padding 34 | :param in_planes: 35 | :param out_planes: 36 | :param stride: 37 | :return: 38 | """ 39 | return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, 40 | padding=1, bias=False) 41 | 42 | 43 | class BasicBlock(nn.Module): 44 | expansion = 1 45 | 46 | def __init__(self, inplanes, planes, stride=1, downsample=None): 47 | super(BasicBlock, self).__init__() 48 | self.conv1 = conv3x3(inplanes, planes, stride) 49 | self.bn1 = nn.BatchNorm2d(planes) 50 | self.relu = nn.ReLU(inplace=True) 51 | self.conv2 = conv3x3(planes, planes) 52 | self.bn2 = nn.BatchNorm2d(planes) 53 | self.downsample = downsample 54 | self.stride = stride 55 | 56 | def forward(self, x): 57 | residual = x 58 | 59 | out = self.conv1(x) 60 | out = self.bn1(out) 61 | out = self.relu(out) 62 | 63 | out = self.conv2(out) 64 | out = self.bn2(out) 65 | 66 | if self.downsample is not None: 67 | residual = self.downsample(x) 68 | 69 | out += residual 70 | out = self.relu(out) 71 | 72 | return out 73 | 74 | 75 | class Bottleneck(nn.Module): 76 | expansion = 4 77 | 78 | def __init__(self, inplanes, planes, stride=1, downsample=None): 79 | super(Bottleneck, self).__init__() 80 | self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False) 81 | self.bn1 = nn.BatchNorm2d(planes) 82 | self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, 83 | padding=1, bias=False) 84 | self.bn2 = nn.BatchNorm2d(planes) 85 | self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False) 86 | self.bn3 = nn.BatchNorm2d(planes * 4) 87 | self.relu = nn.ReLU(inplace=True) 88 | self.downsample = downsample 89 | self.stride = stride 90 | 91 | def forward(self, x): 92 | residual = x 93 | 94 | out = self.conv1(x) 95 | out = self.bn1(out) 96 | out = self.relu(out) 97 | 98 | out = self.conv2(out) 99 | out = self.bn2(out) 100 | out = self.relu(out) 101 | 102 | out = self.conv3(out) 103 | out = self.bn3(out) 104 | 105 | if self.downsample is not None: 106 | residual = self.downsample(x) 107 | 108 | out += residual 109 | out = self.relu(out) 110 | 111 | return out 112 | 113 | 114 | class detnet_bottleneck(nn.Module): 115 | # no expansion 116 | # dilation = 2 117 | # type B use 1x1 conv 118 | expansion = 1 119 | 120 | def __init__(self, in_planes, planes, stride=1, block_type='A'): 121 | super(detnet_bottleneck, self).__init__() 122 | self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=1, bias=False) 123 | self.bn1 = nn.BatchNorm2d(planes) 124 | self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, padding=2, bias=False, dilation=2) 125 | self.bn2 = nn.BatchNorm2d(planes) 126 | self.conv3 = nn.Conv2d(planes, self.expansion*planes, kernel_size=1, bias=False) 127 | self.bn3 = nn.BatchNorm2d(self.expansion*planes) 128 | 129 | self.downsample = nn.Sequential() 130 | if stride != 1 or in_planes != self.expansion*planes or block_type == 'B': 131 | self.downsample = nn.Sequential( 132 | nn.Conv2d(in_planes, self.expansion*planes, kernel_size=1, stride=stride, bias=False), 133 | nn.BatchNorm2d(self.expansion*planes) 134 | ) 135 | 136 | def forward(self, x): 137 | out = F.relu(self.bn1(self.conv1(x))) 138 | out = F.relu(self.bn2(self.conv2(out))) 139 | out = self.bn3(self.conv3(out)) 140 | out += self.downsample(x) 141 | out = F.relu(out) 142 | return out 143 | 144 | 145 | class ResNet(nn.Module): 146 | 147 | def __init__(self, block, layers, num_classes=1470): 148 | self.inplanes = 64 149 | super(ResNet, self).__init__() 150 | self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3, 151 | bias=False) 152 | self.bn1 = nn.BatchNorm2d(64) 153 | self.relu = nn.ReLU(inplace=True) 154 | self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1) 155 | self.layer1 = self._make_layer(block, 64, layers[0]) 156 | self.layer2 = self._make_layer(block, 128, layers[1], stride=2) 157 | self.layer3 = self._make_layer(block, 256, layers[2], stride=2) 158 | self.layer4 = self._make_layer(block, 512, layers[3], stride=2) 159 | self.layer5 = self._make_detnet_layer(in_channels=2048) 160 | # self.avgpool = nn.AvgPool2d(14) #fit 448 input size 161 | # self.fc = nn.Linear(512 * block.expansion, num_classes) 162 | self.conv_end = nn.Conv2d(256, 30, kernel_size=3, stride=1, padding=1, bias=False) 163 | self.bn_end = nn.BatchNorm2d(30) 164 | for m in self.modules(): # 遍历模型 165 | if isinstance(m, nn.Conv2d): # isinstance：m类型判断 若当前组件为 conv 166 | n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels 167 | m.weight.data.normal_(0, math.sqrt(2. / n)) # 正太分布初始化 168 | elif isinstance(m, nn.BatchNorm2d): # 若为batchnorm 169 | m.weight.data.fill_(1) # weight为1 170 | m.bias.data.zero_() # bias为0 171 | 172 | def _make_layer(self, block, planes, blocks, stride=1): 173 | downsample = None 174 | if stride != 1 or self.inplanes != planes * block.expansion: 175 | # 步长为2时，即第一次进入layer时，增加downsample层。 176 | # 或者inplans（输入通道数） 不等于 block.expansion倍的planes = (输出通道数） 177 | downsample = nn.Sequential( 178 | nn.Conv2d(self.inplanes, planes * block.expansion, 179 | kernel_size=1, stride=stride, bias=False), 180 | nn.BatchNorm2d(planes * block.expansion), 181 | ) 182 | 183 | layers = [] 184 | layers.append(block(self.inplanes, planes, stride, downsample)) 185 | self.inplanes = planes * block.expansion 186 | for i in range(1, blocks): 187 | layers.append(block(self.inplanes, planes)) 188 | 189 | return nn.Sequential(*layers) 190 | 191 | def _make_detnet_layer(self, in_channels): 192 | layers = [] 193 | layers.append(detnet_bottleneck(in_planes=in_channels, planes=256, block_type='B')) 194 | layers.append(detnet_bottleneck(in_planes=256, planes=256, block_type='A')) 195 | layers.append(detnet_bottleneck(in_planes=256, planes=256, block_type='A')) 196 | return nn.Sequential(*layers) 197 | 198 | def forward(self, x): 199 | x = self.conv1(x) 200 | x = self.bn1(x) 201 | x = self.relu(x) 202 | x = self.maxpool(x) 203 | 204 | x = self.layer1(x) 205 | x = self.layer2(x) 206 | x = self.layer3(x) 207 | x = self.layer4(x) 208 | x = self.layer5(x) 209 | # x = self.avgpool(x) 210 | # x = x.view(x.size(0), -1) 211 | # x = self.fc(x) 212 | x = self.conv_end(x) 213 | x = self.bn_end(x) 214 | # x = F.sigmoid(x) # 归一化到0-1 215 | x = torch.sigmoid(x) 216 | # x = x.view(-1,7,7,30) 217 | x = x.permute(0, 2, 3, 1) # (-1,7,7,30) 218 | 219 | return x 220 | 221 | 222 | def resnet18(pretrained=False, **kwargs): 223 | """Constructs a ResNet-18 model. 224 | 225 | Args: 226 | pretrained (bool): If True, returns a model pre-trained on ImageNet 227 | """ 228 | model = ResNet(BasicBlock, [2, 2, 2, 2], **kwargs) 229 | if pretrained: 230 | model.load_state_dict(model_zoo.load_url(model_urls['resnet18'])) 231 | return model 232 | 233 | 234 | def resnet34(pretrained=False, **kwargs): 235 | """Constructs a ResNet-34 model. 236 | 237 | Args: 238 | pretrained (bool): If True, returns a model pre-trained on ImageNet 239 | """ 240 | model = ResNet(BasicBlock, [3, 4, 6, 3], **kwargs) 241 | if pretrained: 242 | model.load_state_dict(model_zoo.load_url(model_urls['resnet34'])) 243 | return model 244 | 245 | 246 | def resnet50(pretrained=False, **kwargs): 247 | """Constructs a ResNet-50 model. 248 | 249 | Args: 250 | pretrained (bool): If True, returns a model pre-trained on ImageNet 251 | """ 252 | model = ResNet(Bottleneck, [3, 4, 6, 3], **kwargs) 253 | if pretrained: 254 | model.load_state_dict(model_zoo.load_url(model_urls['resnet50'])) 255 | return model 256 | 257 | 258 | def resnet101(pretrained=False, **kwargs): 259 | """Constructs a ResNet-101 model. 260 | 261 | Args: 262 | pretrained (bool): If True, returns a model pre-trained on ImageNet 263 | """ 264 | model = ResNet(Bottleneck, [3, 4, 23, 3], **kwargs) 265 | if pretrained: 266 | model.load_state_dict(model_zoo.load_url(model_urls['resnet101'])) 267 | return model 268 | 269 | 270 | def resnet152(pretrained=False, **kwargs): 271 | """Constructs a ResNet-152 model. 272 | 273 | Args: 274 | pretrained (bool): If True, returns a model pre-trained on ImageNet 275 | """ 276 | model = ResNet(Bottleneck, [3, 8, 36, 3], **kwargs) 277 | if pretrained: 278 | model.load_state_dict(model_zoo.load_url(model_urls['resnet152'])) 279 | return model 280 | 281 | 282 | if __name__ == "__main__": 283 | from torchsummary import summary 284 | net = resnet50().cuda() 285 | summary(net, (3, 418, 418)) 286 | # print(net) -------------------------------------------------------------------------------- /models/vgg_yolo.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | """ 3 | @Time : 2020/08/12 18:30 4 | @Author : Bryce 5 | @File : vgg_yolo.py 6 | @Noice : 7 | @Modificattion : 8 | @Author : 9 | @Time : 10 | @Detail : 11 | """ 12 | import torch.nn as nn 13 | import torch.utils.model_zoo as model_zoo 14 | import math 15 | import torch 16 | import torch.nn.functional as F 17 | 18 | 19 | __all__ = [ 20 | 'VGG', 'vgg11', 'vgg11_bn', 'vgg13', 'vgg13_bn', 'vgg16', 'vgg16_bn', 21 | 'vgg19_bn', 'vgg19', 22 | ] 23 | 24 | 25 | model_urls = { 26 | 'vgg11': 'https://download.pytorch.org/models/vgg11-bbd30ac9.pth', 27 | 'vgg13': 'https://download.pytorch.org/models/vgg13-c768596a.pth', 28 | 'vgg16': 'https://download.pytorch.org/models/vgg16-397923af.pth', 29 | 'vgg19': 'https://download.pytorch.org/models/vgg19-dcbb9e9d.pth', 30 | 'vgg11_bn': 'https://download.pytorch.org/models/vgg11_bn-6002323d.pth', 31 | 'vgg13_bn': 'https://download.pytorch.org/models/vgg13_bn-abd245e5.pth', 32 | 'vgg16_bn': 'https://download.pytorch.org/models/vgg16_bn-6c64b313.pth', 33 | 'vgg19_bn': 'https://download.pytorch.org/models/vgg19_bn-c79401a0.pth', 34 | } 35 | 36 | 37 | class VGG(nn.Module): 38 | 39 | def __init__(self, features, num_classes=1000, image_size=448): 40 | super(VGG, self).__init__() 41 | self.features = features 42 | self.image_size = image_size 43 | # self.classifier = nn.Sequential( 44 | # nn.Linear(512 * 7 * 7, 4096), 45 | # nn.ReLU(True), 46 | # nn.Dropout(), 47 | # nn.Linear(4096, 4096), 48 | # nn.ReLU(True), 49 | # nn.Dropout(), 50 | # nn.Linear(4096, num_classes), 51 | # ) 52 | # if self.image_size == 448: 53 | # self.extra_conv1 = conv_bn_relu(512,512) 54 | # self.extra_conv2 = conv_bn_relu(512,512) 55 | # self.downsample = nn.MaxPool2d(kernel_size=2, stride=2) 56 | self.classifier = nn.Sequential( 57 | nn.Linear(512 * 7 * 7, 4096), 58 | nn.ReLU(True), 59 | nn.Dropout(), 60 | nn.Linear(4096, 1470), 61 | ) 62 | self._initialize_weights() 63 | 64 | def forward(self, x): 65 | x = self.features(x) 66 | # if self.image_size == 448: 67 | # x = self.extra_conv1(x) 68 | # x = self.extra_conv2(x) 69 | # x = self.downsample(x) 70 | x = x.view(x.size(0), -1) 71 | x = self.classifier(x) 72 | # x = F.sigmoid(x) #归一化到0-1 73 | x = torch.sigmoid(x) 74 | x = x.view(-1,7,7,30) 75 | return x 76 | 77 | def _initialize_weights(self): 78 | for m in self.modules(): 79 | if isinstance(m, nn.Conv2d): 80 | n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels 81 | m.weight.data.normal_(0, math.sqrt(2. / n)) 82 | if m.bias is not None: 83 | m.bias.data.zero_() 84 | elif isinstance(m, nn.BatchNorm2d): 85 | m.weight.data.fill_(1) 86 | m.bias.data.zero_() 87 | elif isinstance(m, nn.Linear): 88 | m.weight.data.normal_(0, 0.01) 89 | m.bias.data.zero_() 90 | 91 | 92 | def make_layers(cfg, batch_norm=False): 93 | layers = [] 94 | in_channels = 3 95 | s = 1 96 | first_flag=True 97 | for v in cfg: 98 | s=1 99 | if (v==64 and first_flag): 100 | s=2 101 | first_flag=False 102 | if v == 'M': 103 | layers += [nn.MaxPool2d(kernel_size=2, stride=2)] 104 | else: 105 | conv2d = nn.Conv2d(in_channels, v, kernel_size=3, stride=s, padding=1) 106 | if batch_norm: 107 | layers += [conv2d, nn.BatchNorm2d(v), nn.ReLU(inplace=True)] 108 | else: 109 | layers += [conv2d, nn.ReLU(inplace=True)] 110 | in_channels = v 111 | return nn.Sequential(*layers) 112 | 113 | def conv_bn_relu(in_channels,out_channels,kernel_size=3,stride=2,padding=1): 114 | return nn.Sequential( 115 | nn.Conv2d(in_channels,out_channels,kernel_size=kernel_size,padding=padding,stride=stride), 116 | nn.BatchNorm2d(out_channels), 117 | nn.ReLU(True) 118 | ) 119 | 120 | 121 | cfg = { 122 | 'A': [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'], 123 | 'B': [64, 64, 'M', 128, 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'], 124 | 'D': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512, 'M', 512, 512, 512, 'M'], 125 | 'E': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 256, 'M', 512, 512, 512, 512, 'M', 512, 512, 512, 512, 'M'], 126 | } 127 | 128 | 129 | def vgg11(pretrained=False, **kwargs): 130 | """VGG 11-layer model (configuration "A") 131 | 132 | Args: 133 | pretrained (bool): If True, returns a model pre-trained on ImageNet 134 | """ 135 | model = VGG(make_layers(cfg['A']), **kwargs) 136 | if pretrained: 137 | model.load_state_dict(model_zoo.load_url(model_urls['vgg11'])) 138 | return model 139 | 140 | 141 | def vgg11_bn(pretrained=False, **kwargs): 142 | """VGG 11-layer model (configuration "A") with batch normalization 143 | 144 | Args: 145 | pretrained (bool): If True, returns a model pre-trained on ImageNet 146 | """ 147 | model = VGG(make_layers(cfg['A'], batch_norm=True), **kwargs) 148 | if pretrained: 149 | model.load_state_dict(model_zoo.load_url(model_urls['vgg11_bn'])) 150 | return model 151 | 152 | 153 | def vgg13(pretrained=False, **kwargs): 154 | """VGG 13-layer model (configuration "B") 155 | 156 | Args: 157 | pretrained (bool): If True, returns a model pre-trained on ImageNet 158 | """ 159 | model = VGG(make_layers(cfg['B']), **kwargs) 160 | if pretrained: 161 | model.load_state_dict(model_zoo.load_url(model_urls['vgg13'])) 162 | return model 163 | 164 | 165 | def vgg13_bn(pretrained=False, **kwargs): 166 | """VGG 13-layer model (configuration "B") with batch normalization 167 | 168 | Args: 169 | pretrained (bool): If True, returns a model pre-trained on ImageNet 170 | """ 171 | model = VGG(make_layers(cfg['B'], batch_norm=True), **kwargs) 172 | if pretrained: 173 | model.load_state_dict(model_zoo.load_url(model_urls['vgg13_bn'])) 174 | return model 175 | 176 | 177 | def vgg16(pretrained=False, **kwargs): 178 | """VGG 16-layer model (configuration "D") 179 | 180 | Args: 181 | pretrained (bool): If True, returns a model pre-trained on ImageNet 182 | """ 183 | model = VGG(make_layers(cfg['D']), **kwargs) 184 | if pretrained: 185 | model.load_state_dict(model_zoo.load_url(model_urls['vgg16'])) 186 | return model 187 | 188 | 189 | def vgg16_bn(pretrained=False, **kwargs): 190 | """VGG 16-layer model (configuration "D") with batch normalization 191 | 192 | Args: 193 | pretrained (bool): If True, returns a model pre-trained on ImageNet 194 | """ 195 | model = VGG(make_layers(cfg['D'], batch_norm=True), **kwargs) 196 | if pretrained: 197 | model.load_state_dict(model_zoo.load_url(model_urls['vgg16_bn'])) 198 | return model 199 | 200 | 201 | def vgg19(pretrained=False, **kwargs): 202 | """VGG 19-layer model (configuration "E") 203 | 204 | Args: 205 | pretrained (bool): If True, returns a model pre-trained on ImageNet 206 | """ 207 | model = VGG(make_layers(cfg['E']), **kwargs) 208 | if pretrained: 209 | model.load_state_dict(model_zoo.load_url(model_urls['vgg19'])) 210 | return model 211 | 212 | 213 | def vgg19_bn(pretrained=False, **kwargs): 214 | """VGG 19-layer model (configuration 'E') with batch normalization 215 | 216 | Args: 217 | pretrained (bool): If True, returns a model pre-trained on ImageNet 218 | """ 219 | model = VGG(make_layers(cfg['E'], batch_norm=True), **kwargs) 220 | if pretrained: 221 | model.load_state_dict(model_zoo.load_url(model_urls['vgg19_bn'])) 222 | return model 223 | 224 | 225 | def test(): 226 | import torch 227 | from torch.autograd import Variable 228 | model = vgg16() 229 | model.classifier = nn.Sequential( 230 | nn.Linear(512 * 7 * 7, 4096), 231 | nn.ReLU(True), 232 | nn.Dropout(), 233 | nn.Linear(4096, 4096), 234 | nn.ReLU(True), 235 | nn.Dropout(), 236 | nn.Linear(4096, 1470), 237 | ) 238 | print(model.classifier[6]) 239 | #print(model) 240 | img = torch.rand(2,3,224,224) 241 | img = Variable(img) 242 | output = model(img) 243 | print(output.size()) 244 | 245 | 246 | if __name__ == '__main__': 247 | test() -------------------------------------------------------------------------------- /models/yoloLoss.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | """ 3 | @Time : 2020/08/12 18:30 4 | @Author : FelixFu / Bryce 5 | @File : yoloLoss.py 6 | @Noice : 7 | @Modificattion : 8 | @Detail : a little dufficult in builting yoloLoss funcion 9 | """ 10 | 11 | import torch 12 | import torch.nn as nn 13 | import torch.nn.functional as F 14 | 15 | 16 | class yoloLoss(nn.Module): 17 | def __init__(self, S, B, l_coord, l_noobj): 18 | super(yoloLoss, self).__init__() 19 | self.S = S 20 | self.B = B 21 | self.l_coord = l_coord 22 | self.l_noobj = l_noobj 23 | 24 | def compute_iou(self, box1, box2): 25 | """Compute the intersection over union of two set of boxes, each box is [x1,y1,x2,y2]. 26 | Args: 27 | box1: (tensor) bounding boxes, sized [N,4]. 28 | box2: (tensor) bounding boxes, sized [M,4]. 29 | Return: 30 | (tensor) iou, sized [N,M]. 31 | """ 32 | # 首先计算两个box左上角点坐标的最大值和右下角坐标的最小值，然后计算交集面积，最后把交集面积除以对应的并集面积 33 | N = box1.size(0) 34 | M = box2.size(0) 35 | 36 | lt = torch.max( # 左上角的点 37 | box1[:, :2].unsqueeze(1).expand(N, M, 2), # [N,2] -> [N,1,2] -> [N,M,2] 38 | box2[:, :2].unsqueeze(0).expand(N, M, 2), # [M,2] -> [1,M,2] -> [N,M,2] 39 | ) 40 | 41 | rb = torch.min( # 右下角的点 42 | box1[:, 2:].unsqueeze(1).expand(N, M, 2), # [N,2] -> [N,1,2] -> [N,M,2] 43 | box2[:, 2:].unsqueeze(0).expand(N, M, 2), # [M,2] -> [1,M,2] -> [N,M,2] 44 | ) 45 | 46 | wh = rb - lt # [N,M,2] 47 | wh[wh < 0] = 0 # clip at 指两个box没有重叠区域 48 | inter = wh[:, :, 0] * wh[:, :, 1] # [N,M] 49 | 50 | area1 = (box1[:, 2]-box1[:, 0]) * (box1[:, 3]-box1[:, 1]) # [N,] 51 | area2 = (box2[:, 2]-box2[:, 0]) * (box2[:, 3]-box2[:, 1]) # [M,] 52 | area1 = area1.unsqueeze(1).expand_as(inter) # [N,] -> [N,1] -> [N,M] 53 | area2 = area2.unsqueeze(0).expand_as(inter) # [M,] -> [1,M] -> [N,M] 54 | 55 | iou = inter / (area1 + area2 - inter) 56 | return iou 57 | 58 | def forward(self, pred_tensor, target_tensor): 59 | """ 60 | pred_tensor: (tensor) size(batchsize,S,S,Bx5+20=30) [x,y,w,h,c] 61 | target_tensor: (tensor) size(batchsize,S,S,30) 62 | """ 63 | N = pred_tensor.size()[0] 64 | # 具有目标标签的索引(bs, 7, 7, 30)中7*7方格中的哪个方格包含目标 65 | coo_mask = target_tensor[:, :, :, 4] > 0 # coo_mask.shape = (bs, 7, 7) 66 | noo_mask = target_tensor[:, :, :, 4] == 0 # 不具有目标的标签索引 67 | # 得到含物体的坐标等信息(coo_mask扩充到与target_tensor一样形状, 沿最后一维扩充) 68 | coo_mask = coo_mask.unsqueeze(-1).expand_as(target_tensor) 69 | noo_mask = noo_mask.unsqueeze(-1).expand_as(target_tensor) 70 | 71 | # coo_pred：tensor[, 30](所有batch数据都压缩在一起) 72 | coo_pred = pred_tensor[coo_mask].view(-1, 30) 73 | box_pred = coo_pred[:, :10].contiguous().view(-1, 5) # box[x1,y1,w1,h1,c1], [x2,y2,w2,h2,c2] 74 | class_pred = coo_pred[:, 10:] 75 | 76 | coo_target = target_tensor[coo_mask].view(-1, 30) 77 | box_target = coo_target[:, :10].contiguous().view(-1, 5) 78 | class_target = coo_target[:, 10:] 79 | 80 | # compute not contain obj loss 81 | noo_pred = pred_tensor[noo_mask].view(-1, 30) 82 | noo_target = target_tensor[noo_mask].view(-1, 30) 83 | 84 | noo_pred_mask = torch.cuda.ByteTensor(noo_pred.size()).bool() 85 | noo_pred_mask.zero_() 86 | noo_pred_mask[:, 4] = 1 87 | noo_pred_mask[:, 9] = 1 88 | noo_pred_c = noo_pred[noo_pred_mask] # noo pred只需要计算 c 的损失 size[-1,2] 89 | noo_target_c = noo_target[noo_pred_mask] 90 | nooobj_loss = F.mse_loss(noo_pred_c, noo_target_c, size_average=False) 91 | 92 | # compute contain obj loss 93 | coo_response_mask = torch.cuda.ByteTensor(box_target.size()).bool() 94 | coo_response_mask.zero_() 95 | coo_not_response_mask = torch.cuda.ByteTensor(box_target.size()).bool() 96 | coo_not_response_mask.zero_() 97 | box_target_iou = torch.zeros(box_target.size()).cuda() 98 | for i in range(0, box_target.size()[0], 2): # choose the best iou box 99 | box1 = box_pred[i:i+2] # 获取当前格点预测的b个box 100 | box1_xyxy = torch.FloatTensor(box1.size()) 101 | # (x,y,w,h) 102 | box1_xyxy[:, :2] = box1[:, :2]/14. - 0.5 * box1[:, 2:4] 103 | box1_xyxy[:, 2:4] = box1[:, :2]/14. + 0.5 * box1[:, 2:4] 104 | box2 = box_target[i].view(-1, 5) 105 | box2_xyxy = torch.FloatTensor(box2.size()) 106 | box2_xyxy[:, :2] = box2[:, :2]/14. - 0.5*box2[:, 2:4] 107 | box2_xyxy[:, 2:4] = box2[:, :2]/14. + 0.5*box2[:, 2:4] 108 | iou = self.compute_iou(box1_xyxy[:, :4], box2_xyxy[:, :4]) # [2,1] 109 | max_iou, max_index = iou.max(0) 110 | max_index = max_index.data.cuda() 111 | 112 | coo_response_mask[i+max_index] = 1 113 | coo_not_response_mask[i+1-max_index] = 1 114 | 115 | ##### 116 | # we want the confidence score to equal the 117 | # intersection over union (IOU) between the predicted box 118 | # and the ground truth 119 | ##### 120 | # iou value 作为box包含目标的confidence(赋值在向量的第五个位置) 121 | box_target_iou[i+max_index, torch.LongTensor([4]).cuda()] = (max_iou).data.cuda() 122 | box_target_iou = box_target_iou.cuda() 123 | # 1.response loss 124 | box_pred_response = box_pred[coo_response_mask].view(-1, 5) 125 | box_target_response_iou = box_target_iou[coo_response_mask].view(-1, 5) 126 | box_target_response = box_target[coo_response_mask].view(-1, 5) 127 | contain_loss = F.mse_loss(box_pred_response[:, 4], box_target_response_iou[:, 4], size_average=False) 128 | loc_loss = F.mse_loss(box_pred_response[:, :2], box_target_response[:, :2], size_average=False) + F.mse_loss(torch.sqrt(box_pred_response[:, 2:4]), torch.sqrt(box_target_response[:, 2:4]), size_average=False) 129 | 130 | # 2.not response loss 131 | box_pred_not_response = box_pred[coo_not_response_mask].view(-1, 5) 132 | box_target_not_response = box_target[coo_not_response_mask].view(-1, 5) 133 | box_target_not_response[:, 4] = 0 134 | # not_contain_loss = F.mse_loss(box_pred_response[:,4],box_target_response[:,4],size_average=False) 135 | 136 | # I believe this bug is simply a typo 137 | not_contain_loss = F.mse_loss(box_pred_not_response[:, 4], box_target_not_response[:, 4], size_average=False) 138 | 139 | # 3.class loss 140 | class_loss = F.mse_loss(class_pred, class_target, size_average=False) 141 | 142 | return (self.l_coord*loc_loss + self.B*contain_loss + not_contain_loss + self.l_noobj*nooobj_loss + class_loss)/N 143 | 144 | 145 | 146 | 147 | -------------------------------------------------------------------------------- /predict.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | """ 3 | @Time : 2020/08/12 18:30 4 | @Author : Bryce 5 | @File : train.py 6 | @Noice : 7 | @Modificattion : 8 | @Author : 9 | @Time : 10 | @Detail : 11 | """ 12 | import torch 13 | 14 | from models.resnet_yolo import resnet50 15 | import torchvision.transforms as transforms 16 | import cv2 17 | import numpy as np 18 | import operator 19 | 20 | VOC_CLASSES = ( # always index 0 21 | 'aeroplane', 'bicycle', 'bird', 'boat', 22 | 'bottle', 'bus', 'car', 'cat', 'chair', 23 | 'cow', 'diningtable', 'dog', 'horse', 24 | 'motorbike', 'person', 'pottedplant', 25 | 'sheep', 'sofa', 'train', 'tvmonitor') 26 | 27 | # 数据集20个类别的所对应的20种颜色 28 | Color = [ 29 | [0, 0, 0], 30 | [128, 0, 0], 31 | [0, 128, 0], 32 | [128, 128, 0], 33 | [0, 0, 128], 34 | [128, 0, 128], 35 | [0, 128, 128], 36 | [128, 128, 128], 37 | [64, 0, 0], 38 | [192, 0, 0], 39 | [64, 128, 0], 40 | [192, 128, 0], 41 | [64, 0, 128], 42 | [192, 0, 128], 43 | [64, 128, 128], 44 | [192, 128, 128], 45 | [0, 64, 0], 46 | [128, 64, 0], 47 | [0, 192, 0], 48 | [128, 192, 0], 49 | [0, 64, 128] 50 | ] 51 | 52 | # 对于网络输出预测 改为再图片上画出框及score 53 | def decoder(pred): 54 | """ 55 | pred (tensor) torch.Size([1, 14, 14, 30]) 56 | return (tensor) box[[x1,y1,x2,y2]] label[...] 57 | """ 58 | grid_num = 14 59 | boxes = [] 60 | cls_indexs = [] 61 | probs = [] 62 | cell_size = 1./grid_num 63 | pred = pred.data # torch.Size([1, 14, 14, 30]) 64 | pred = pred.squeeze(0) # torch.Size([14, 14, 30]) 65 | # 0 1 2 3 4 5 6 7 8 9 66 | # [中心坐标,长宽,置信度,中心坐标,长宽,置信度, 20个类别] x 7x7 67 | contain1 = pred[:, :, 4].unsqueeze(2) # torch.Size([14, 14, 1]) 68 | contain2 = pred[:, :, 9].unsqueeze(2) # torch.Size([14, 14, 1]) 69 | contain = torch.cat((contain1, contain2), 2) # torch.Size([14, 14, 2]) 70 | 71 | mask1 = contain > 0.1 # 大于阈值, torch.Size([14, 14, 2]) content: tensor([False, False]) 72 | mask2 = (contain == contain.max()) # we always select the best contain_prob what ever it>0.9 73 | mask = (mask1+mask2).gt(0) 74 | 75 | # min_score,min_index = torch.min(contain, 2) # 每个cell只选最大概率的那个预测框 76 | for i in range(grid_num): 77 | for j in range(grid_num): 78 | for b in range(2): 79 | # index = min_index[i,j] 80 | # mask[i,j,index] = 0 81 | if mask[i, j, b] == 1: 82 | box = pred[i, j, b*5:b*5+4] 83 | contain_prob = torch.FloatTensor([pred[i, j, b*5+4]]) 84 | xy = torch.FloatTensor([j, i])*cell_size # cell左上角 up left of cell 85 | box[:2] = box[:2]*cell_size + xy # return cxcy relative to image 86 | box_xy = torch.FloatTensor(box.size()) # 转换成xy形式 convert[cx,cy,w,h] to [x1,y1,x2,y2] 87 | box_xy[:2] = box[:2] - 0.5*box[2:] 88 | box_xy[2:] = box[:2] + 0.5*box[2:] 89 | max_prob, cls_index = torch.max(pred[i, j, 10:], 0) 90 | if float((contain_prob*max_prob)[0]) > 0.1: 91 | boxes.append(box_xy.view(1, 4)) 92 | cls_indexs.append(cls_index.item()) 93 | probs.append(contain_prob*max_prob) 94 | if len(boxes) == 0: 95 | boxes = torch.zeros((1, 4)) 96 | probs = torch.zeros(1) 97 | cls_indexs = torch.zeros(1) 98 | else: 99 | boxes = torch.cat(boxes, 0) # (n,4) 100 | # print(type(probs)) 101 | # print(len(probs)) 102 | # print(probs) 103 | probs = torch.cat(probs, 0) # (n,) 104 | # print(probs) 105 | # print(type(cls_indexs)) 106 | # print(len(cls_indexs)) 107 | # print(cls_indexs) 108 | cls_indexs = torch.IntTensor(cls_indexs) # (n,) 109 | keep = nms(boxes, probs) 110 | # print("keep:", keep) 111 | 112 | a = boxes[keep] 113 | b = cls_indexs[keep] 114 | c = probs[keep] 115 | return a, b, c 116 | 117 | 118 | def nms(bboxes, scores, threshold=0.5): 119 | ''' 120 | bboxes(tensor) [N,4] 121 | scores(tensor) [N,] 122 | ''' 123 | x1 = bboxes[:, 0] 124 | y1 = bboxes[:, 1] 125 | x2 = bboxes[:, 2] 126 | y2 = bboxes[:, 3] 127 | areas = (x2-x1) * (y2-y1) 128 | # print(scores) # tensor([0.1006, 0.2381, 0.1185, 0.5342, 0.2892, 0.3521, 0.6027]) 129 | _, order = scores.sort(0, descending=True) # 降序排列score 130 | keep = [] 131 | # print("order:", order) # order: tensor([6, 3, 5, 4, 1, 2, 0]) 132 | # print("order.numel:", order.numel()) # 7 133 | while order.numel() > 0: # torch.numel()返回张量元素个数 134 | if order.numel() == 1: # 保留框只剩一个 135 | # print("end1") 136 | # print(type(order)) 137 | # print(order) 138 | i = order 139 | keep.append(i) 140 | break 141 | # print("len:", order.size()) 142 | # print(keep) 143 | i = order[0] # i=6，对应得分中最大的框 保留scores最大的那个框box[i] 144 | keep.append(i) 145 | 146 | # 计算box[i]与其余各框的IOU 147 | xx1 = x1[order[1:]].clamp(min=x1[i]) # [N-1,] 148 | yy1 = y1[order[1:]].clamp(min=y1[i]) 149 | xx2 = x2[order[1:]].clamp(max=x2[i]) 150 | yy2 = y2[order[1:]].clamp(max=y2[i]) 151 | 152 | w = (xx2-xx1).clamp(min=0) 153 | h = (yy2-yy1).clamp(min=0) 154 | inter = w*h # [N-1,] 155 | ovr = inter / (areas[i] + areas[order[1:]] - inter) 156 | ids = (ovr <= threshold).nonzero(as_tuple=False).squeeze() # 注意此时idx为[N-1,] 而order为[N,] 157 | if ids.numel() == 0: 158 | # print("end2") 159 | break 160 | order = order[ids+1] # 修补索引之间的差值 161 | # print(keep) 162 | return torch.LongTensor(keep) 163 | # return keep 164 | 165 | 166 | # start predict one image 167 | def predict_gpu(model, image_name, root_path=''): 168 | result = [] 169 | image = cv2.imread(root_path+image_name) 170 | # print(root_path , image_name) 171 | h, w, _ = image.shape 172 | img = cv2.resize(image, (448, 448)) 173 | img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) 174 | mean = (123, 117, 104) # RGB 175 | img = img - np.array(mean, dtype=np.float32) 176 | 177 | transform = transforms.Compose([transforms.ToTensor(), ]) 178 | img = transform(img) # torch.Size([3, 448, 448]) 179 | img = img[None, :, :, :] # img: torch.Size([1, 3, 448, 448]) 180 | img = img.cuda() 181 | 182 | pred = model(img) # 1x14x14x30 183 | pred = pred.cpu() 184 | boxes, cls_indexs, probs = decoder(pred) 185 | 186 | for i, box in enumerate(boxes): 187 | x1 = int(box[0]*w) 188 | x2 = int(box[2]*w) 189 | y1 = int(box[1]*h) 190 | y2 = int(box[3]*h) 191 | cls_index = cls_indexs[i] 192 | cls_index = int(cls_index) # convert LongTensor to int 193 | prob = probs[i] 194 | prob = float(prob) 195 | result.append([(x1, y1), (x2, y2), VOC_CLASSES[cls_index], image_name, prob]) 196 | return result 197 | 198 | 199 | if __name__ == '__main__': 200 | model = resnet50() 201 | print('load model...') 202 | model.load_state_dict(torch.load('checkpoints/best.pth')) 203 | model.eval() 204 | model.cuda() 205 | image_name = 'imgs/001526.jpg' 206 | image = cv2.imread(image_name) 207 | print('predicting...') 208 | result = predict_gpu(model, image_name) 209 | 210 | for left_up, right_bottom, class_name, _, prob in result: 211 | color = Color[VOC_CLASSES.index(class_name)] 212 | cv2.rectangle(image, left_up, right_bottom, color, 2) 213 | label = class_name+str(round(prob, 2)) 214 | text_size, baseline = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX, 0.4, 1) 215 | p1 = (left_up[0], left_up[1] - text_size[1]) 216 | cv2.rectangle(image, (p1[0] - 2//2, p1[1] - 2 - baseline), (p1[0] + text_size[0], p1[1] + text_size[1]), color, -1) 217 | cv2.putText(image, label, (p1[0], p1[1] + baseline), cv2.FONT_HERSHEY_SIMPLEX, 0.4, (255, 255, 255), 1, 8) 218 | 219 | cv2.imwrite('imgs/001526_result.jpg', image) 220 | img = cv2.imread('imgs/001526_result.jpg') 221 | cv2.imshow('img', img) 222 | cv2.waitKey(0) 223 | 224 | 225 | 226 | -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | absl-py==0.9.0 2 | adabound==0.0.5 3 | appdirs==1.4.4 4 | asn1crypto==0.24.0 5 | attrs==17.4.0 6 | Automat==0.6.0 7 | backcall==0.2.0 8 | bleach==3.1.5 9 | blinker==1.4 10 | cachetools==4.1.1 11 | certifi==2018.1.18 12 | cffi==1.14.0 13 | chardet==3.0.4 14 | click==6.7 15 | cloud-init==19.4 16 | cloudpickle==1.4.1 17 | colorama==0.3.7 18 | command-not-found==0.3 19 | configobj==5.0.6 20 | constantly==15.1.0 21 | cryptography==2.1.4 22 | cycler==0.10.0 23 | Cython==0.29.17 24 | decorator==4.4.2 25 | defusedxml==0.6.0 26 | distro-info==0.18ubuntu0.18.04.1 27 | entrypoints==0.3 28 | fire==0.3.1 29 | Flask==1.1.2 30 | future==0.18.2 31 | google-auth==1.18.0 32 | google-auth-oauthlib==0.4.1 33 | grpcio==1.30.0 34 | grpcio-tools==1.30.0 35 | horovod==0.19.2 36 | httplib2==0.9.2 37 | hyperlink==17.3.1 38 | idna==2.6 39 | imageio==2.9.0 40 | importlib-metadata==1.7.0 41 | incremental==16.10.1 42 | ipykernel==5.3.0 43 | ipython==7.15.0 44 | ipython-genutils==0.2.0 45 | ipywidgets==7.5.1 46 | itsdangerous==1.1.0 47 | jedi==0.17.0 48 | Jinja2==2.11.2 49 | joblib==0.16.0 50 | jsonpatch==1.16 51 | jsonpointer==1.10 52 | jsonschema==2.6.0 53 | jupyter==1.0.0 54 | jupyter-client==6.1.3 55 | jupyter-console==6.1.0 56 | jupyter-core==4.6.3 57 | keras2onnx==1.7.0 58 | keyring==10.6.0 59 | keyrings.alt==3.0 60 | kiwisolver==1.2.0 61 | language-selector==0.1 62 | Mako==1.1.3 63 | Markdown==3.2.2 64 | MarkupSafe==1.0 65 | matplotlib==3.2.1 66 | mistune==0.8.4 67 | mlxtend==0.17.2 68 | nbconvert==5.6.1 69 | nbformat==5.0.7 70 | netifaces==0.10.4 71 | netron==4.3.5 72 | networkx==2.4 73 | notebook==6.0.3 74 | numpy==1.18.4 75 | oauthlib==3.1.0 76 | onnx==1.7.0 77 | onnxconverter-common==1.7.0 78 | onnxmltools==1.7.0 79 | onnxruntime==1.4.0 80 | onnxruntime-gpu==1.4.0 81 | opencv-python==4.3.0.36 82 | ort-gpu-nightly==1.4.0.dev202007171 83 | packaging==20.4 84 | PAM==0.4.2 85 | pandas==1.0.5 86 | pandocfilters==1.4.2 87 | parso==0.7.0 88 | pexpect==4.8.0 89 | pickleshare==0.7.5 90 | Pillow==7.1.2 91 | pip==20.1 92 | prometheus-client==0.8.0 93 | prompt-toolkit==3.0.5 94 | protobuf==3.12.2 95 | psutil==5.7.0 96 | ptyprocess==0.6.0 97 | pyasn1==0.4.2 98 | pyasn1-modules==0.2.1 99 | pycocotools==2.0.0 100 | pycparser==2.20 101 | pycrypto==2.6.1 102 | pycuda==2019.1.2 103 | Pygments==2.6.1 104 | pygobject==3.26.1 105 | PyJWT==1.5.3 106 | pyOpenSSL==17.5.0 107 | pyparsing==2.4.7 108 | pyserial==3.4 109 | python-apt==1.6.5+ubuntu0.2 110 | python-dateutil==2.8.1 111 | python-debian==0.1.32 112 | pytools==2020.2 113 | pytz==2020.1 114 | PyWavelets==1.1.1 115 | pyxdg==0.25 116 | PyYAML==3.12 117 | pyzmq==19.0.1 118 | qtconsole==4.7.4 119 | QtPy==1.9.0 120 | requests==2.24.0 121 | requests-oauthlib==1.3.0 122 | requests-unixsocket==0.1.5 123 | rsa==4.6 124 | scikit-image==0.17.2 125 | scikit-learn==0.23.1 126 | scipy==1.5.1 127 | SecretStorage==2.3.1 128 | Send2Trash==1.5.0 129 | service-identity==16.0.0 130 | setuptools==49.1.2 131 | six==1.11.0 132 | skl2onnx==1.7.0 133 | sklearn==0.0 134 | ssh-import-id==5.7 135 | systemd-python==234 136 | tensorboard==2.2.2 137 | tensorboard-plugin-wit==1.7.0 138 | tensorrt==7.0.0.11 139 | termcolor==1.1.0 140 | terminado==0.8.3 141 | terminaltables==3.1.0 142 | testpath==0.4.4 143 | threadpoolctl==2.1.0 144 | tifffile==2020.7.17 145 | torch==1.5.0 146 | torch2trt==0.1.0 147 | torchfile==0.1.0 148 | torchsummary==1.5.1 149 | torchvision==0.6.0 150 | tornado==6.0.4 151 | tqdm==4.46.0 152 | traitlets==4.3.3 153 | Twisted==17.9.0 154 | typing-extensions==3.7.4.2 155 | ufw==0.36 156 | unattended-upgrades==0.1 157 | urllib3==1.22 158 | visdom==0.1.8.9 159 | wcwidth==0.2.4 160 | webencodings==0.5.1 161 | websocket-client==0.57.0 162 | Werkzeug==1.0.1 163 | wget==3.2 164 | wheel==0.30.0 165 | widgetsnbextension==3.5.1 166 | zipp==3.1.0 167 | zope.interface==4.3.2 168 | -------------------------------------------------------------------------------- /train.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | """ 3 | @Time : 2020/08/12 18:30 4 | @Author : Bryce 5 | @File : train.py 6 | @Noice : 7 | @Modificattion : 8 | @Author : 9 | @Time : 10 | @Detail : 11 | """ 12 | import warnings 13 | import os 14 | import numpy as np 15 | 16 | import torch 17 | from torch.utils.data import DataLoader 18 | import torchvision.transforms as transforms 19 | from torchvision import models 20 | 21 | from models.vgg_yolo import vgg16_bn 22 | from models.resnet_yolo import resnet50 23 | from models.yoloLoss import yoloLoss 24 | from utils.dataset import yoloDataset 25 | 26 | warnings.filterwarnings('ignore') 27 | # 设置GPU ID 28 | os.environ["CUDA_VISIBLE_DEVICES"] = "0" 29 | 30 | # 判断GPU是否可用 31 | use_gpu = torch.cuda.is_available() 32 | 33 | # 数据文件 34 | file_root = 'datasets' 35 | 36 | # 超参数 37 | learning_rate = 0.001 38 | num_epochs = 100 39 | batch_size = 24 40 | 41 | # checkpoints 42 | resume = True 43 | 44 | # ---------------------数据读取--------------------- 45 | train_dataset = yoloDataset(root=file_root, list_file='images.txt', train=True, 46 | transform=[transforms.ToTensor()]) 47 | # train_dataset = yoloDataset(root=file_root, list_file=['voc12_trainval.txt','voc07_trainval.txt'], 48 | # train=True,transform = [transforms.ToTensor()] ) 49 | train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True, num_workers=4) 50 | 51 | test_dataset = yoloDataset(root=file_root, list_file='voc2007test.txt', train=False, 52 | transform=[transforms.ToTensor()]) 53 | test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False, num_workers=4) 54 | print('the train dataset has %d images' % (len(train_dataset))) 55 | print('the test dataset has %d images' % (len(test_dataset))) 56 | print('the batch_size is %d' % batch_size) 57 | 58 | 59 | # ---------------------网络选择--------------------- 60 | use_resnet = True 61 | if use_resnet: 62 | net = resnet50() 63 | else: 64 | net = vgg16_bn() 65 | 66 | if resume: 67 | print("loading weight from checkpoints/best.pth") 68 | net.load_state_dict(torch.load('checkpoints/best.pth')) 69 | else: 70 | print('loading pre-trined model ......') 71 | if use_resnet: 72 | resnet = models.resnet50(pretrained=True) 73 | new_state_dict = resnet.state_dict() 74 | dd = net.state_dict() 75 | for k in new_state_dict.keys(): 76 | print(k) 77 | if k in dd.keys() and not k.startswith('fc'): 78 | # print('yes') 79 | dd[k] = new_state_dict[k] 80 | net.load_state_dict(dd) 81 | else: 82 | vgg = models.vgg16_bn(pretrained=True) 83 | new_state_dict = vgg.state_dict() 84 | dd = net.state_dict() 85 | for k in new_state_dict.keys(): 86 | print(k) 87 | if k in dd.keys() and k.startswith('features'): 88 | print('yes') 89 | dd[k] = new_state_dict[k] 90 | net.load_state_dict(dd) 91 | 92 | if use_gpu: 93 | print('this computer has gpu %d and current is %s' % (torch.cuda.device_count(), 94 | torch.cuda.current_device())) 95 | net.cuda() 96 | 97 | 98 | # ---------------------损失函数--------------------- 99 | criterion = yoloLoss(7, 2, 5, 0.5) 100 | 101 | # ---------------------优化器---------------------- 102 | 103 | # different learning rate 104 | params = [] 105 | params_dict = dict(net.named_parameters()) 106 | for key, value in params_dict.items(): 107 | if key.startswith('features'): 108 | params += [{'params': [value], 'lr':learning_rate*1}] 109 | else: 110 | params += [{'params': [value], 'lr':learning_rate}] 111 | optimizer = torch.optim.SGD(params, lr=learning_rate, momentum=0.9, weight_decay=5e-4) 112 | # optimizer = torch.optim.Adam(net.parameters(),lr=learning_rate,weight_decay=1e-4) 113 | 114 | 115 | # ---------------------训练--------------------- 116 | logfile = open('checkpoints/log.txt', 'w') 117 | num_iter = 0 118 | best_test_loss = np.inf 119 | 120 | for epoch in range(num_epochs): 121 | # train 122 | net.train() 123 | if epoch == 30: 124 | learning_rate = 0.0001 125 | if epoch == 40: 126 | learning_rate = 0.00001 127 | for param_group in optimizer.param_groups: 128 | param_group['lr'] = learning_rate 129 | 130 | print('\n\nStarting epoch %d / %d' % (epoch + 1, num_epochs)) 131 | print('Learning Rate for this epoch: {}'.format(learning_rate)) 132 | 133 | total_loss = 0. 134 | 135 | for i, (images, target) in enumerate(train_loader): 136 | if use_gpu: 137 | images, target = images.cuda(), target.cuda() 138 | 139 | pred = net(images) 140 | loss = criterion(pred, target) 141 | total_loss += loss.data.item() 142 | 143 | optimizer.zero_grad() 144 | loss.backward() 145 | optimizer.step() 146 | if (i+1) % 5 == 0: 147 | print('Epoch [%d/%d], Iter [%d/%d] Loss: %.4f, average_loss: %.4f' 148 | % (epoch+1, num_epochs, i+1, len(train_loader), loss.item(), total_loss / (i+1))) 149 | num_iter += 1 150 | 151 | # validation 152 | validation_loss = 0.0 153 | net.eval() 154 | for i, (images, target) in enumerate(test_loader): 155 | if use_gpu: 156 | images, target = images.cuda(), target.cuda() 157 | 158 | pred = net(images) 159 | loss = criterion(pred, target) 160 | validation_loss += loss.item() 161 | validation_loss /= len(test_loader) 162 | 163 | if best_test_loss > validation_loss: 164 | best_test_loss = validation_loss 165 | print('get best test loss %.5f' % best_test_loss) 166 | torch.save(net.state_dict(), 'checkpoints/best.pth') 167 | logfile.writelines(str(epoch) + '\t' + str(validation_loss) + '\n') 168 | logfile.flush() 169 | 170 | 171 | -------------------------------------------------------------------------------- /utils/__init__.py: -------------------------------------------------------------------------------- 1 | # !/usr/bin/python 2 | # -*- coding:utf-8 -*- 3 | # author: Bryce -------------------------------------------------------------------------------- /utils/dataset.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | """ 3 | @Time : 2020/08/12 18:30 4 | @Author : Bryce 5 | @File : dataset.py 6 | @Noice : 7 | @Modificattion : txt描述文件 image_name.jpg x y w h c x y w h c 这样就是说一张图片中有两个目标 8 | @Author : 9 | @Time : 10 | @Detail : 11 | """ 12 | 13 | import os 14 | import os.path 15 | 16 | import random 17 | import numpy as np 18 | 19 | import torch 20 | import torch.utils.data as data 21 | import torchvision.transforms as transforms 22 | 23 | import cv2 24 | import matplotlib.pyplot as plt 25 | 26 | 27 | class yoloDataset(data.Dataset): 28 | image_size = 448 29 | # train_dataset = yoloDataset(root=datasets/, 30 | # list_file=['voc2012.txt','voc2007.txt'], 31 | # train=True,transform = [transforms.ToTensor()] ) 32 | 33 | def __init__(self, root, list_file, train, transform): 34 | self.root = root # 数据集根目录 35 | self.train = train # 是否为训练 36 | self.transform = transform # 转换 37 | self.fnames = [] # 文件名s [001.jpg, 002.jpg] 38 | self.boxes = [] # boxes [ [box], [[x1,y1,x2,y2], ...], ... ] 39 | self.labels = [] # labels [ [1], [2], ... ] 40 | self.mean = (123, 117, 104) # RGB 41 | self.num_samples = 0 # 样本总数 42 | 43 | if isinstance(list_file, list): 44 | # Cat multiple list files together. 45 | # This is especially useful for voc07/voc12 combination. 46 | tmp_file = os.path.join(root, 'images.txt') 47 | list_file = [os.path.join(root, list_file[0]), os.path.join(root, list_file[1])] 48 | os.system('cat %s > %s' % (' '.join(list_file), tmp_file)) 49 | list_file = tmp_file 50 | else: 51 | list_file = os.path.join(root, list_file) 52 | 53 | # 处理标签 54 | with open(list_file) as f: 55 | lines = f.readlines() 56 | for line in lines: 57 | splited = line.strip().split() # ['005246.jpg', '84', '48', '493', '387', '2'] 坐标 + 类型（labels） 58 | self.fnames.append(splited[0]) 59 | num_boxes = (len(splited) - 1) // 5 60 | box = [] 61 | label = [] 62 | for i in range(num_boxes): 63 | x = float(splited[1+5*i]) 64 | y = float(splited[2+5*i]) 65 | x2 = float(splited[3+5*i]) 66 | y2 = float(splited[4+5*i]) 67 | c = splited[5+5*i] 68 | box.append([x, y, x2, y2]) 69 | label.append(int(c)+1) 70 | self.boxes.append(torch.Tensor(box)) 71 | self.labels.append(torch.LongTensor(label)) 72 | self.num_samples = len(self.boxes) # 数据集中包含所有Ground truth个数 73 | 74 | def __getitem__(self, idx): 75 | fname = self.fnames[idx] 76 | img = cv2.imread(os.path.join(self.root, "images", fname)) 77 | boxes = self.boxes[idx].clone() 78 | labels = self.labels[idx].clone() 79 | 80 | # 数据增强 81 | # if self.train: 82 | # img = self.random_bright(img) 83 | # img, boxes = self.random_flip(img, boxes) 84 | # img, boxes = self.randomScale(img, boxes) 85 | # img = self.randomBlur(img) 86 | # img = self.RandomBrightness(img) 87 | # img = self.RandomHue(img) 88 | # img = self.RandomSaturation(img) 89 | # img, boxes, labels = self.randomShift(img, boxes, labels) 90 | # img, boxes, labels = self.randomCrop(img, boxes, labels) 91 | 92 | # # debug 93 | # box_show = boxes.numpy().reshape(-1) 94 | # print(box_show) 95 | # img_show = self.BGR2RGB(img) 96 | # pt1 = (int(box_show[0]), int(box_show[1])) 97 | # pt2 = (int(box_show[2]), int(box_show[3])) 98 | # cv2.rectangle(img_show, pt1=pt1, pt2=pt2, color=(0, 255, 0), thickness=1) 99 | # print(type(img_show)) 100 | # plt.figure() 101 | # plt.imshow(img_show) 102 | # plt.show() 103 | # plt.savefig("a.png") 104 | # #debug 105 | h, w, _ = img.shape # 不管通道数 _ 106 | boxes /= torch.Tensor([w, h, w, h]).expand_as(boxes) # 一张图片中框的坐标归一化，即转换为对于0,0点的(0,1)范围内的表述 107 | img = self.BGR2RGB(img) # because pytorch pretrained model use RGB 108 | img = self.subMean(img, self.mean) # 减去均值 109 | img = cv2.resize(img, (self.image_size, self.image_size)) 110 | target = self.encoder(boxes, labels) # 7x7x30 111 | for t in self.transform: 112 | img = t(img) 113 | 114 | return img, target 115 | 116 | def __len__(self): 117 | return self.num_samples 118 | 119 | def encoder(self, boxes, labels): 120 | ''' 121 | boxes (tensor) [[x1,y1,x2,y2],[]] 122 | labels (tensor) [...] 123 | return 7x7x30 124 | ''' 125 | grid_num = 14 # 论文中设为7 126 | target = torch.zeros((grid_num, grid_num, 30)) 127 | cell_size = 1./grid_num # 之前已经把目标框归一化，故这里用1. 作为除数 128 | wh = boxes[:, 2:]-boxes[:, :2] # 宽高 129 | cxcy = (boxes[:, 2:]+boxes[:, :2])/2 # 中心点 130 | for i in range(cxcy.size()[0]): # 对于数据集中的每个框 这里cxcy.size() == num_samples 131 | cxcy_sample = cxcy[i] 132 | ij = (cxcy_sample/cell_size).ceil()-1 # ij 是一个list, 表示目标中心点cxcy在归一化后的图片中所处的x y 方向的第几个网格 133 | # 0 1 2 3 4 5 6 7 8 9 134 | # [中心坐标,长宽,置信度,中心坐标,长宽,置信度, 20个类别] x 7x7 135 | target[int(ij[1]), int(ij[0]), 4] = 1 # 第一个框的置信度 136 | target[int(ij[1]), int(ij[0]), 9] = 1 # 第二个框的置信度 137 | target[int(ij[1]), int(ij[0]), int(labels[i])+9] = 1 138 | xy = ij*cell_size # 匹配到划分后的子网格的左上角相对坐标 139 | delta_xy = (cxcy_sample -xy)/cell_size # delta_xy对于目标中心点落入的子网格，目标中心坐标相对于子网格左上点的位置比例 140 | target[int(ij[1]), int(ij[0]), 2:4] = wh[i] # 坐标w,h代表了预测边界框的width、height相对于整幅图像width,height的比例，范围为(0,1) 141 | target[int(ij[1]), int(ij[0]), :2] = delta_xy 142 | target[int(ij[1]), int(ij[0]), 7:9] = wh[i] 143 | target[int(ij[1]), int(ij[0]), 5:7] = delta_xy 144 | return target 145 | 146 | def BGR2RGB(self, img): 147 | return cv2.cvtColor(img, cv2.COLOR_BGR2RGB) 148 | 149 | def BGR2HSV(self,img): 150 | return cv2.cvtColor(img,cv2.COLOR_BGR2HSV) 151 | 152 | def HSV2BGR(self,img): 153 | return cv2.cvtColor(img,cv2.COLOR_HSV2BGR) 154 | 155 | def RandomBrightness(self,bgr): 156 | if random.random() < 0.5: 157 | hsv = self.BGR2HSV(bgr) 158 | h,s,v = cv2.split(hsv) 159 | adjust = random.choice([0.5,1.5]) 160 | v = v*adjust 161 | v = np.clip(v, 0, 255).astype(hsv.dtype) 162 | hsv = cv2.merge((h,s,v)) 163 | bgr = self.HSV2BGR(hsv) 164 | return bgr 165 | 166 | def RandomSaturation(self,bgr): 167 | if random.random() < 0.5: 168 | hsv = self.BGR2HSV(bgr) 169 | h,s,v = cv2.split(hsv) 170 | adjust = random.choice([0.5,1.5]) 171 | s = s*adjust 172 | s = np.clip(s, 0, 255).astype(hsv.dtype) 173 | hsv = cv2.merge((h,s,v)) 174 | bgr = self.HSV2BGR(hsv) 175 | return bgr 176 | 177 | def RandomHue(self,bgr): 178 | if random.random() < 0.5: 179 | hsv = self.BGR2HSV(bgr) 180 | h,s,v = cv2.split(hsv) 181 | adjust = random.choice([0.5,1.5]) 182 | h = h*adjust 183 | h = np.clip(h, 0, 255).astype(hsv.dtype) 184 | hsv = cv2.merge((h,s,v)) 185 | bgr = self.HSV2BGR(hsv) 186 | return bgr 187 | 188 | def randomBlur(self,bgr): 189 | if random.random()<0.5: 190 | bgr = cv2.blur(bgr,(5,5)) 191 | return bgr 192 | 193 | def randomShift(self,bgr,boxes,labels): 194 | #平移变换 195 | center = (boxes[:,2:]+boxes[:,:2])/2 196 | if random.random() <0.5: 197 | height,width,c = bgr.shape 198 | after_shfit_image = np.zeros((height,width,c),dtype=bgr.dtype) 199 | after_shfit_image[:,:,:] = (104,117,123) #bgr 200 | shift_x = random.uniform(-width*0.2,width*0.2) 201 | shift_y = random.uniform(-height*0.2,height*0.2) 202 | #print(bgr.shape,shift_x,shift_y) 203 | #原图像的平移 204 | if shift_x>=0 and shift_y>=0: 205 | after_shfit_image[int(shift_y):,int(shift_x):,:] = bgr[:height-int(shift_y),:width-int(shift_x),:] 206 | elif shift_x>=0 and shift_y<0: 207 | after_shfit_image[:height+int(shift_y),int(shift_x):,:] = bgr[-int(shift_y):,:width-int(shift_x),:] 208 | elif shift_x <0 and shift_y >=0: 209 | after_shfit_image[int(shift_y):,:width+int(shift_x),:] = bgr[:height-int(shift_y),-int(shift_x):,:] 210 | elif shift_x<0 and shift_y<0: 211 | after_shfit_image[:height+int(shift_y),:width+int(shift_x),:] = bgr[-int(shift_y):,-int(shift_x):,:] 212 | 213 | shift_xy = torch.FloatTensor([[int(shift_x),int(shift_y)]]).expand_as(center) 214 | center = center + shift_xy 215 | mask1 = (center[:,0] >0) & (center[:,0] < width) 216 | mask2 = (center[:,1] >0) & (center[:,1] < height) 217 | mask = (mask1 & mask2).view(-1,1) 218 | boxes_in = boxes[mask.expand_as(boxes)].view(-1,4) 219 | if len(boxes_in) == 0: 220 | return bgr,boxes,labels 221 | box_shift = torch.FloatTensor([[int(shift_x),int(shift_y),int(shift_x),int(shift_y)]]).expand_as(boxes_in) 222 | boxes_in = boxes_in+box_shift 223 | labels_in = labels[mask.view(-1)] 224 | return after_shfit_image,boxes_in,labels_in 225 | return bgr,boxes,labels 226 | 227 | def randomScale(self,bgr,boxes): 228 | #固定住高度，以0.8-1.2伸缩宽度，做图像形变 229 | if random.random() < 0.5: 230 | scale = random.uniform(0.8,1.2) 231 | height,width,c = bgr.shape 232 | bgr = cv2.resize(bgr,(int(width*scale),height)) 233 | scale_tensor = torch.FloatTensor([[scale,1,scale,1]]).expand_as(boxes) 234 | boxes = boxes * scale_tensor 235 | return bgr,boxes 236 | return bgr,boxes 237 | 238 | def randomCrop(self,bgr,boxes,labels): 239 | if random.random() < 0.5: 240 | center = (boxes[:,2:]+boxes[:,:2])/2 241 | height,width,c = bgr.shape 242 | h = random.uniform(0.6*height,height) 243 | w = random.uniform(0.6*width,width) 244 | x = random.uniform(0,width-w) 245 | y = random.uniform(0,height-h) 246 | x,y,h,w = int(x),int(y),int(h),int(w) 247 | 248 | center = center - torch.FloatTensor([[x,y]]).expand_as(center) 249 | mask1 = (center[:,0]>0) & (center[:,0]0) & (center[:,1] 0.3: 288 | im = im * alpha + random.randrange(-delta, delta) 289 | im = im.clip(min=0, max=255).astype(np.uint8) 290 | return im 291 | 292 | 293 | if __name__ == '__main__': 294 | from torch.utils.data import DataLoader 295 | import torchvision.transforms as transforms 296 | file_root = "../datasets" 297 | # train_dataset = yoloDataset(root=file_root, list_file=['voc2012.txt', 'voc2007.txt'], 298 | # train=True, transform=[transforms.ToTensor()]) 299 | train_dataset = yoloDataset(root=file_root, list_file='images.txt', 300 | train=True, transform=[transforms.ToTensor()]) 301 | train_loader = DataLoader(train_dataset, batch_size=1, shuffle=True, num_workers=0) 302 | train_iter = iter(train_loader) 303 | for i in range(1): 304 | img, target = next(train_iter) 305 | print(img.shape, target.shape) 306 | print(train_dataset.num_samples) 307 | 308 | 309 | -------------------------------------------------------------------------------- /utils/piplist2equal.py: -------------------------------------------------------------------------------- 1 | # !/usr/bin/python 2 | # -*- coding:utf-8 -*- 3 | # author: Bryce 4 | 5 | import os 6 | 7 | 8 | reqs = [] 9 | with open("requirements.txt", mode='r') as f_old: 10 | lines = f_old.readlines() 11 | lines = lines[2:] 12 | for line in lines: 13 | line = line.split() 14 | temp = line[0] + "==" + line[1] 15 | reqs.append(temp) 16 | 17 | with open("requirements.txt", mode='w') as f: 18 | for line in reqs: 19 | f.write(line + "\n") 20 | -------------------------------------------------------------------------------- /utils/xml2txt.py: -------------------------------------------------------------------------------- 1 | import xml.etree.ElementTree as ET 2 | import os 3 | 4 | VOC_CLASSES = ( # always index 0 5 | 'aeroplane', 'bicycle', 'bird', 'boat', 6 | 'bottle', 'bus', 'car', 'cat', 'chair', 7 | 'cow', 'diningtable', 'dog', 'horse', 8 | 'motorbike', 'person', 'pottedplant', 9 | 'sheep', 'sofa', 'train', 'tvmonitor') 10 | 11 | 12 | def parse_rec(filename): 13 | """ Parse a PASCAL VOC xml file """ 14 | tree = ET.parse(filename) 15 | objects = [] 16 | for obj in tree.findall('object'): 17 | obj_struct = {} 18 | difficult = int(obj.find('difficult').text) 19 | if difficult == 1: # 剔除困难标签的图片 20 | # print(filename) 21 | continue 22 | obj_struct['name'] = obj.find('name').text 23 | bbox = obj.find('bndbox') 24 | obj_struct['bbox'] = [int(float(bbox.find('xmin').text)), 25 | int(float(bbox.find('ymin').text)), 26 | int(float(bbox.find('xmax').text)), 27 | int(float(bbox.find('ymax').text))] 28 | objects.append(obj_struct) 29 | 30 | return objects 31 | 32 | 33 | txt_file = open('voc2007test.txt', 'w') 34 | test_file = open('voc07testimg.txt', 'r') 35 | lines = test_file.readlines() 36 | lines = [x[:-1] for x in lines] 37 | print(lines) 38 | 39 | Annotations = 'path/to/VOC2007/Annotations/' 40 | xml_files = os.listdir(Annotations) 41 | 42 | count = 0 43 | for xml_file in xml_files: 44 | count += 1 45 | if xml_file.split('.')[0] not in lines: 46 | # print(xml_file.split('.')[0]) 47 | continue 48 | image_path = xml_file.split('.')[0] + '.jpg' 49 | results = parse_rec(Annotations + xml_file) 50 | if len(results) == 0: 51 | print(xml_file) 52 | continue 53 | txt_file.write(image_path) 54 | # num_obj = len(results) 55 | # txt_file.write(str(num_obj)+' ') 56 | for result in results: 57 | class_name = result['name'] 58 | bbox = result['bbox'] 59 | class_name = VOC_CLASSES.index(class_name) 60 | txt_file.write(' '+str(bbox[0])+' '+str(bbox[1])+' '+str(bbox[2])+' '+str(bbox[3])+' '+str(class_name)) 61 | txt_file.write('\n') 62 | 63 | txt_file.close() --------------------------------------------------------------------------------