├── README.md
├── data
    ├── 1.JPEG
    └── 2.JPEG
├── imagenet.lua
├── images
    ├── 1.png
    ├── 10.png
    ├── 11.png
    ├── 12.png
    ├── 13.png
    ├── 14.png
    ├── 15.png
    ├── 16.png
    ├── 17.png
    ├── 18.png
    ├── 19.png
    ├── 2.png
    ├── 20.png
    ├── 3.png
    ├── 4.png
    ├── 5.png
    ├── 6.png
    ├── 7.png
    ├── 8.png
    └── 9.png
├── transforms.lua
└── visual.lua


/README.md:
--------------------------------------------------------------------------------
 1 | # visualResNet_torch
 2 | 
 3 | Curious about where is the ResNet looking at? This repo visualizes the class specific saliency map or discriminative location. 
 4 | It is Torch re-implementation of [Learning Deep Features for Discriminative Localization, Bolei Zhou et. al.](http://arxiv.org/abs/1512.04150), 
 5 | without modifying the network or re-training. (MatConvNet re-implementation can be found at this [repo](https://github.com/zhanghang1989/visualResNet).) We look directly through the ResNet to see the world.
 6 | 
 7 | 
 8 | ### Get Started
 9 | 
10 | - The code relies on [Torch](https://github.com/torch/torch7), which should be downloaded and built before running the experiments. Download the code:
11 | ```bash
12 | git clone https://github.com/zhanghang1989/visualResNet_torch.git
13 | ```
14 | 
15 | - Download the models from [facebook git](https://github.com/facebook/fb.resnet.torch/tree/master/pretrained) 
16 | 
17 | - Run the progrem:
18 | ```bash
19 | th visual.lua resnet-50.t7 data/1.JPEG data/2.JPEG 
20 | ```
21 | 
22 | - Visualize all the images in a folder
23 | ```bash
24 | th visual.lua resnet-50.t7 data/*
25 | ```
26 | 
27 | ### Examples
28 | <div style="text-align:center">
29 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/1.png" width="420" />
30 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/2.png" width="420" />
31 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/3.png" width="420" />
32 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/4.png" width="420" />
33 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/5.png" width="420" />
34 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/6.png" width="420" />
35 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/7.png" width="420" />
36 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/8.png" width="420" />
37 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/9.png" width="420" />
38 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/10.png" width="420" />
39 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/11.png" width="420" />
40 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/12.png" width="420" />
41 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/13.png" width="420" />
42 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/14.png" width="420" />
43 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/15.png" width="420" />
44 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/16.png" width="420" />
45 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/17.png" width="420" />
46 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/18.png" width="420" />
47 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/19.png" width="420" />
48 | <img src ="https://github.com/zhanghang1989/visualResNet_torch/raw/master/images/20.png" width="420" />
49 | </div>
50 | 
51 | 


--------------------------------------------------------------------------------
/data/1.JPEG:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/data/1.JPEG


--------------------------------------------------------------------------------
/data/2.JPEG:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/data/2.JPEG


--------------------------------------------------------------------------------
/imagenet.lua:
--------------------------------------------------------------------------------
   1 | return{
   2 |  'tench, Tinca tinca',
   3 |  'goldfish, Carassius auratus',
   4 |  'great white shark, white shark, man-eater, man-eating shark, Carcharodon carcharias',
   5 |  'tiger shark, Galeocerdo cuvieri',
   6 |  'hammerhead, hammerhead shark',
   7 |  'electric ray, crampfish, numbfish, torpedo',
   8 |  'stingray',
   9 |  'cock',
  10 |  'hen',
  11 |  'ostrich, Struthio camelus',
  12 |  'brambling, Fringilla montifringilla',
  13 |  'goldfinch, Carduelis carduelis',
  14 |  'house finch, linnet, Carpodacus mexicanus',
  15 |  'junco, snowbird',
  16 |  'indigo bunting, indigo finch, indigo bird, Passerina cyanea',
  17 |  'robin, American robin, Turdus migratorius',
  18 |  'bulbul',
  19 |  'jay',
  20 |  'magpie',
  21 |  'chickadee',
  22 |  'water ouzel, dipper',
  23 |  'kite',
  24 |  'bald eagle, American eagle, Haliaeetus leucocephalus',
  25 |  'vulture',
  26 |  'great grey owl, great gray owl, Strix nebulosa',
  27 |  'European fire salamander, Salamandra salamandra',
  28 |  'common newt, Triturus vulgaris',
  29 |  'eft',
  30 |  'spotted salamander, Ambystoma maculatum',
  31 |  'axolotl, mud puppy, Ambystoma mexicanum',
  32 |  'bullfrog, Rana catesbeiana',
  33 |  'tree frog, tree-frog',
  34 |  'tailed frog, bell toad, ribbed toad, tailed toad, Ascaphus trui',
  35 |  'loggerhead, loggerhead turtle, Caretta caretta',
  36 |  'leatherback turtle, leatherback, leathery turtle, Dermochelys coriacea',
  37 |  'mud turtle',
  38 |  'terrapin',
  39 |  'box turtle, box tortoise',
  40 |  'banded gecko',
  41 |  'common iguana, iguana, Iguana iguana',
  42 |  'American chameleon, anole, Anolis carolinensis',
  43 |  'whiptail, whiptail lizard',
  44 |  'agama',
  45 |  'frilled lizard, Chlamydosaurus kingi',
  46 |  'alligator lizard',
  47 |  'Gila monster, Heloderma suspectum',
  48 |  'green lizard, Lacerta viridis',
  49 |  'African chameleon, Chamaeleo chamaeleon',
  50 |  'Komodo dragon, Komodo lizard, dragon lizard, giant lizard, Varanus komodoensis',
  51 |  'African crocodile, Nile crocodile, Crocodylus niloticus',
  52 |  'American alligator, Alligator mississipiensis',
  53 |  'triceratops',
  54 |  'thunder snake, worm snake, Carphophis amoenus',
  55 |  'ringneck snake, ring-necked snake, ring snake',
  56 |  'hognose snake, puff adder, sand viper',
  57 |  'green snake, grass snake',
  58 |  'king snake, kingsnake',
  59 |  'garter snake, grass snake',
  60 |  'water snake',
  61 |  'vine snake',
  62 |  'night snake, Hypsiglena torquata',
  63 |  'boa constrictor, Constrictor constrictor',
  64 |  'rock python, rock snake, Python sebae',
  65 |  'Indian cobra, Naja naja',
  66 |  'green mamba',
  67 |  'sea snake',
  68 |  'horned viper, cerastes, sand viper, horned asp, Cerastes cornutus',
  69 |  'diamondback, diamondback rattlesnake, Crotalus adamanteus',
  70 |  'sidewinder, horned rattlesnake, Crotalus cerastes',
  71 |  'trilobite',
  72 |  'harvestman, daddy longlegs, Phalangium opilio',
  73 |  'scorpion',
  74 |  'black and gold garden spider, Argiope aurantia',
  75 |  'barn spider, Araneus cavaticus',
  76 |  'garden spider, Aranea diademata',
  77 |  'black widow, Latrodectus mactans',
  78 |  'tarantula',
  79 |  'wolf spider, hunting spider',
  80 |  'tick',
  81 |  'centipede',
  82 |  'black grouse',
  83 |  'ptarmigan',
  84 |  'ruffed grouse, partridge, Bonasa umbellus',
  85 |  'prairie chicken, prairie grouse, prairie fowl',
  86 |  'peacock',
  87 |  'quail',
  88 |  'partridge',
  89 |  'African grey, African gray, Psittacus erithacus',
  90 |  'macaw',
  91 |  'sulphur-crested cockatoo, Kakatoe galerita, Cacatua galerita',
  92 |  'lorikeet',
  93 |  'coucal',
  94 |  'bee eater',
  95 |  'hornbill',
  96 |  'hummingbird',
  97 |  'jacamar',
  98 |  'toucan',
  99 |  'drake',
 100 |  'red-breasted merganser, Mergus serrator',
 101 |  'goose',
 102 |  'black swan, Cygnus atratus',
 103 |  'tusker',
 104 |  'echidna, spiny anteater, anteater',
 105 |  'platypus, duckbill, duckbilled platypus, duck-billed platypus, Ornithorhynchus anatinus',
 106 |  'wallaby, brush kangaroo',
 107 |  'koala, koala bear, kangaroo bear, native bear, Phascolarctos cinereus',
 108 |  'wombat',
 109 |  'jellyfish',
 110 |  'sea anemone, anemone',
 111 |  'brain coral',
 112 |  'flatworm, platyhelminth',
 113 |  'nematode, nematode worm, roundworm',
 114 |  'conch',
 115 |  'snail',
 116 |  'slug',
 117 |  'sea slug, nudibranch',
 118 |  'chiton, coat-of-mail shell, sea cradle, polyplacophore',
 119 |  'chambered nautilus, pearly nautilus, nautilus',
 120 |  'Dungeness crab, Cancer magister',
 121 |  'rock crab, Cancer irroratus',
 122 |  'fiddler crab',
 123 |  'king crab, Alaska crab, Alaskan king crab, Alaska king crab, Paralithodes camtschatica',
 124 |  'American lobster, Northern lobster, Maine lobster, Homarus americanus',
 125 |  'spiny lobster, langouste, rock lobster, crawfish, crayfish, sea crawfish',
 126 |  'crayfish, crawfish, crawdad, crawdaddy',
 127 |  'hermit crab',
 128 |  'isopod',
 129 |  'white stork, Ciconia ciconia',
 130 |  'black stork, Ciconia nigra',
 131 |  'spoonbill',
 132 |  'flamingo',
 133 |  'little blue heron, Egretta caerulea',
 134 |  'American egret, great white heron, Egretta albus',
 135 |  'bittern',
 136 |  'crane',
 137 |  'limpkin, Aramus pictus',
 138 |  'European gallinule, Porphyrio porphyrio',
 139 |  'American coot, marsh hen, mud hen, water hen, Fulica americana',
 140 |  'bustard',
 141 |  'ruddy turnstone, Arenaria interpres',
 142 |  'red-backed sandpiper, dunlin, Erolia alpina',
 143 |  'redshank, Tringa totanus',
 144 |  'dowitcher',
 145 |  'oystercatcher, oyster catcher',
 146 |  'pelican',
 147 |  'king penguin, Aptenodytes patagonica',
 148 |  'albatross, mollymawk',
 149 |  'grey whale, gray whale, devilfish, Eschrichtius gibbosus, Eschrichtius robustus',
 150 |  'killer whale, killer, orca, grampus, sea wolf, Orcinus orca',
 151 |  'dugong, Dugong dugon',
 152 |  'sea lion',
 153 |  'Chihuahua',
 154 |  'Japanese spaniel',
 155 |  'Maltese dog, Maltese terrier, Maltese',
 156 |  'Pekinese, Pekingese, Peke',
 157 |  'Shih-Tzu',
 158 |  'Blenheim spaniel',
 159 |  'papillon',
 160 |  'toy terrier',
 161 |  'Rhodesian ridgeback',
 162 |  'Afghan hound, Afghan',
 163 |  'basset, basset hound',
 164 |  'beagle',
 165 |  'bloodhound, sleuthhound',
 166 |  'bluetick',
 167 |  'black-and-tan coonhound',
 168 |  'Walker hound, Walker foxhound',
 169 |  'English foxhound',
 170 |  'redbone',
 171 |  'borzoi, Russian wolfhound',
 172 |  'Irish wolfhound',
 173 |  'Italian greyhound',
 174 |  'whippet',
 175 |  'Ibizan hound, Ibizan Podenco',
 176 |  'Norwegian elkhound, elkhound',
 177 |  'otterhound, otter hound',
 178 |  'Saluki, gazelle hound',
 179 |  'Scottish deerhound, deerhound',
 180 |  'Weimaraner',
 181 |  'Staffordshire bullterrier, Staffordshire bull terrier',
 182 |  'American Staffordshire terrier, Staffordshire terrier, American pit bull terrier, pit bull terrier',
 183 |  'Bedlington terrier',
 184 |  'Border terrier',
 185 |  'Kerry blue terrier',
 186 |  'Irish terrier',
 187 |  'Norfolk terrier',
 188 |  'Norwich terrier',
 189 |  'Yorkshire terrier',
 190 |  'wire-haired fox terrier',
 191 |  'Lakeland terrier',
 192 |  'Sealyham terrier, Sealyham',
 193 |  'Airedale, Airedale terrier',
 194 |  'cairn, cairn terrier',
 195 |  'Australian terrier',
 196 |  'Dandie Dinmont, Dandie Dinmont terrier',
 197 |  'Boston bull, Boston terrier',
 198 |  'miniature schnauzer',
 199 |  'giant schnauzer',
 200 |  'standard schnauzer',
 201 |  'Scotch terrier, Scottish terrier, Scottie',
 202 |  'Tibetan terrier, chrysanthemum dog',
 203 |  'silky terrier, Sydney silky',
 204 |  'soft-coated wheaten terrier',
 205 |  'West Highland white terrier',
 206 |  'Lhasa, Lhasa apso',
 207 |  'flat-coated retriever',
 208 |  'curly-coated retriever',
 209 |  'golden retriever',
 210 |  'Labrador retriever',
 211 |  'Chesapeake Bay retriever',
 212 |  'German short-haired pointer',
 213 |  'vizsla, Hungarian pointer',
 214 |  'English setter',
 215 |  'Irish setter, red setter',
 216 |  'Gordon setter',
 217 |  'Brittany spaniel',
 218 |  'clumber, clumber spaniel',
 219 |  'English springer, English springer spaniel',
 220 |  'Welsh springer spaniel',
 221 |  'cocker spaniel, English cocker spaniel, cocker',
 222 |  'Sussex spaniel',
 223 |  'Irish water spaniel',
 224 |  'kuvasz',
 225 |  'schipperke',
 226 |  'groenendael',
 227 |  'malinois',
 228 |  'briard',
 229 |  'kelpie',
 230 |  'komondor',
 231 |  'Old English sheepdog, bobtail',
 232 |  'Shetland sheepdog, Shetland sheep dog, Shetland',
 233 |  'collie',
 234 |  'Border collie',
 235 |  'Bouvier des Flandres, Bouviers des Flandres',
 236 |  'Rottweiler',
 237 |  'German shepherd, German shepherd dog, German police dog, alsatian',
 238 |  'Doberman, Doberman pinscher',
 239 |  'miniature pinscher',
 240 |  'Greater Swiss Mountain dog',
 241 |  'Bernese mountain dog',
 242 |  'Appenzeller',
 243 |  'EntleBucher',
 244 |  'boxer',
 245 |  'bull mastiff',
 246 |  'Tibetan mastiff',
 247 |  'French bulldog',
 248 |  'Great Dane',
 249 |  'Saint Bernard, St Bernard',
 250 |  'Eskimo dog, husky',
 251 |  'malamute, malemute, Alaskan malamute',
 252 |  'Siberian husky',
 253 |  'dalmatian, coach dog, carriage dog',
 254 |  'affenpinscher, monkey pinscher, monkey dog',
 255 |  'basenji',
 256 |  'pug, pug-dog',
 257 |  'Leonberg',
 258 |  'Newfoundland, Newfoundland dog',
 259 |  'Great Pyrenees',
 260 |  'Samoyed, Samoyede',
 261 |  'Pomeranian',
 262 |  'chow, chow chow',
 263 |  'keeshond',
 264 |  'Brabancon griffon',
 265 |  'Pembroke, Pembroke Welsh corgi',
 266 |  'Cardigan, Cardigan Welsh corgi',
 267 |  'toy poodle',
 268 |  'miniature poodle',
 269 |  'standard poodle',
 270 |  'Mexican hairless',
 271 |  'timber wolf, grey wolf, gray wolf, Canis lupus',
 272 |  'white wolf, Arctic wolf, Canis lupus tundrarum',
 273 |  'red wolf, maned wolf, Canis rufus, Canis niger',
 274 |  'coyote, prairie wolf, brush wolf, Canis latrans',
 275 |  'dingo, warrigal, warragal, Canis dingo',
 276 |  'dhole, Cuon alpinus',
 277 |  'African hunting dog, hyena dog, Cape hunting dog, Lycaon pictus',
 278 |  'hyena, hyaena',
 279 |  'red fox, Vulpes vulpes',
 280 |  'kit fox, Vulpes macrotis',
 281 |  'Arctic fox, white fox, Alopex lagopus',
 282 |  'grey fox, gray fox, Urocyon cinereoargenteus',
 283 |  'tabby, tabby cat',
 284 |  'tiger cat',
 285 |  'Persian cat',
 286 |  'Siamese cat, Siamese',
 287 |  'Egyptian cat',
 288 |  'cougar, puma, catamount, mountain lion, painter, panther, Felis concolor',
 289 |  'lynx, catamount',
 290 |  'leopard, Panthera pardus',
 291 |  'snow leopard, ounce, Panthera uncia',
 292 |  'jaguar, panther, Panthera onca, Felis onca',
 293 |  'lion, king of beasts, Panthera leo',
 294 |  'tiger, Panthera tigris',
 295 |  'cheetah, chetah, Acinonyx jubatus',
 296 |  'brown bear, bruin, Ursus arctos',
 297 |  'American black bear, black bear, Ursus americanus, Euarctos americanus',
 298 |  'ice bear, polar bear, Ursus Maritimus, Thalarctos maritimus',
 299 |  'sloth bear, Melursus ursinus, Ursus ursinus',
 300 |  'mongoose',
 301 |  'meerkat, mierkat',
 302 |  'tiger beetle',
 303 |  'ladybug, ladybeetle, lady beetle, ladybird, ladybird beetle',
 304 |  'ground beetle, carabid beetle',
 305 |  'long-horned beetle, longicorn, longicorn beetle',
 306 |  'leaf beetle, chrysomelid',
 307 |  'dung beetle',
 308 |  'rhinoceros beetle',
 309 |  'weevil',
 310 |  'fly',
 311 |  'bee',
 312 |  'ant, emmet, pismire',
 313 |  'grasshopper, hopper',
 314 |  'cricket',
 315 |  'walking stick, walkingstick, stick insect',
 316 |  'cockroach, roach',
 317 |  'mantis, mantid',
 318 |  'cicada, cicala',
 319 |  'leafhopper',
 320 |  'lacewing, lacewing fly',
 321 |  "dragonfly, darning needle, devil's darning needle, sewing needle, snake feeder, snake doctor, mosquito hawk, skeeter hawk",
 322 |  'damselfly',
 323 |  'admiral',
 324 |  'ringlet, ringlet butterfly',
 325 |  'monarch, monarch butterfly, milkweed butterfly, Danaus plexippus',
 326 |  'cabbage butterfly',
 327 |  'sulphur butterfly, sulfur butterfly',
 328 |  'lycaenid, lycaenid butterfly',
 329 |  'starfish, sea star',
 330 |  'sea urchin',
 331 |  'sea cucumber, holothurian',
 332 |  'wood rabbit, cottontail, cottontail rabbit',
 333 |  'hare',
 334 |  'Angora, Angora rabbit',
 335 |  'hamster',
 336 |  'porcupine, hedgehog',
 337 |  'fox squirrel, eastern fox squirrel, Sciurus niger',
 338 |  'marmot',
 339 |  'beaver',
 340 |  'guinea pig, Cavia cobaya',
 341 |  'sorrel',
 342 |  'zebra',
 343 |  'hog, pig, grunter, squealer, Sus scrofa',
 344 |  'wild boar, boar, Sus scrofa',
 345 |  'warthog',
 346 |  'hippopotamus, hippo, river horse, Hippopotamus amphibius',
 347 |  'ox',
 348 |  'water buffalo, water ox, Asiatic buffalo, Bubalus bubalis',
 349 |  'bison',
 350 |  'ram, tup',
 351 |  'bighorn, bighorn sheep, cimarron, Rocky Mountain bighorn, Rocky Mountain sheep, Ovis canadensis',
 352 |  'ibex, Capra ibex',
 353 |  'hartebeest',
 354 |  'impala, Aepyceros melampus',
 355 |  'gazelle',
 356 |  'Arabian camel, dromedary, Camelus dromedarius',
 357 |  'llama',
 358 |  'weasel',
 359 |  'mink',
 360 |  'polecat, fitch, foulmart, foumart, Mustela putorius',
 361 |  'black-footed ferret, ferret, Mustela nigripes',
 362 |  'otter',
 363 |  'skunk, polecat, wood pussy',
 364 |  'badger',
 365 |  'armadillo',
 366 |  'three-toed sloth, ai, Bradypus tridactylus',
 367 |  'orangutan, orang, orangutang, Pongo pygmaeus',
 368 |  'gorilla, Gorilla gorilla',
 369 |  'chimpanzee, chimp, Pan troglodytes',
 370 |  'gibbon, Hylobates lar',
 371 |  'siamang, Hylobates syndactylus, Symphalangus syndactylus',
 372 |  'guenon, guenon monkey',
 373 |  'patas, hussar monkey, Erythrocebus patas',
 374 |  'baboon',
 375 |  'macaque',
 376 |  'langur',
 377 |  'colobus, colobus monkey',
 378 |  'proboscis monkey, Nasalis larvatus',
 379 |  'marmoset',
 380 |  'capuchin, ringtail, Cebus capucinus',
 381 |  'howler monkey, howler',
 382 |  'titi, titi monkey',
 383 |  'spider monkey, Ateles geoffroyi',
 384 |  'squirrel monkey, Saimiri sciureus',
 385 |  'Madagascar cat, ring-tailed lemur, Lemur catta',
 386 |  'indri, indris, Indri indri, Indri brevicaudatus',
 387 |  'Indian elephant, Elephas maximus',
 388 |  'African elephant, Loxodonta africana',
 389 |  'lesser panda, red panda, panda, bear cat, cat bear, Ailurus fulgens',
 390 |  'giant panda, panda, panda bear, coon bear, Ailuropoda melanoleuca',
 391 |  'barracouta, snoek',
 392 |  'eel',
 393 |  'coho, cohoe, coho salmon, blue jack, silver salmon, Oncorhynchus kisutch',
 394 |  'rock beauty, Holocanthus tricolor',
 395 |  'anemone fish',
 396 |  'sturgeon',
 397 |  'gar, garfish, garpike, billfish, Lepisosteus osseus',
 398 |  'lionfish',
 399 |  'puffer, pufferfish, blowfish, globefish',
 400 |  'abacus',
 401 |  'abaya',
 402 |  "academic gown, academic robe, judge's robe",
 403 |  'accordion, piano accordion, squeeze box',
 404 |  'acoustic guitar',
 405 |  'aircraft carrier, carrier, flattop, attack aircraft carrier',
 406 |  'airliner',
 407 |  'airship, dirigible',
 408 |  'altar',
 409 |  'ambulance',
 410 |  'amphibian, amphibious vehicle',
 411 |  'analog clock',
 412 |  'apiary, bee house',
 413 |  'apron',
 414 |  'ashcan, trash can, garbage can, wastebin, ash bin, ash-bin, ashbin, dustbin, trash barrel, trash bin',
 415 |  'assault rifle, assault gun',
 416 |  'backpack, back pack, knapsack, packsack, rucksack, haversack',
 417 |  'bakery, bakeshop, bakehouse',
 418 |  'balance beam, beam',
 419 |  'balloon',
 420 |  'ballpoint, ballpoint pen, ballpen, Biro',
 421 |  'Band Aid',
 422 |  'banjo',
 423 |  'bannister, banister, balustrade, balusters, handrail',
 424 |  'barbell',
 425 |  'barber chair',
 426 |  'barbershop',
 427 |  'barn',
 428 |  'barometer',
 429 |  'barrel, cask',
 430 |  'barrow, garden cart, lawn cart, wheelbarrow',
 431 |  'baseball',
 432 |  'basketball',
 433 |  'bassinet',
 434 |  'bassoon',
 435 |  'bathing cap, swimming cap',
 436 |  'bath towel',
 437 |  'bathtub, bathing tub, bath, tub',
 438 |  'beach wagon, station wagon, wagon, estate car, beach waggon, station waggon, waggon',
 439 |  'beacon, lighthouse, beacon light, pharos',
 440 |  'beaker',
 441 |  'bearskin, busby, shako',
 442 |  'beer bottle',
 443 |  'beer glass',
 444 |  'bell cote, bell cot',
 445 |  'bib',
 446 |  'bicycle-built-for-two, tandem bicycle, tandem',
 447 |  'bikini, two-piece',
 448 |  'binder, ring-binder',
 449 |  'binoculars, field glasses, opera glasses',
 450 |  'birdhouse',
 451 |  'boathouse',
 452 |  'bobsled, bobsleigh, bob',
 453 |  'bolo tie, bolo, bola tie, bola',
 454 |  'bonnet, poke bonnet',
 455 |  'bookcase',
 456 |  'bookshop, bookstore, bookstall',
 457 |  'bottlecap',
 458 |  'bow',
 459 |  'bow tie, bow-tie, bowtie',
 460 |  'brass, memorial tablet, plaque',
 461 |  'brassiere, bra, bandeau',
 462 |  'breakwater, groin, groyne, mole, bulwark, seawall, jetty',
 463 |  'breastplate, aegis, egis',
 464 |  'broom',
 465 |  'bucket, pail',
 466 |  'buckle',
 467 |  'bulletproof vest',
 468 |  'bullet train, bullet',
 469 |  'butcher shop, meat market',
 470 |  'cab, hack, taxi, taxicab',
 471 |  'caldron, cauldron',
 472 |  'candle, taper, wax light',
 473 |  'cannon',
 474 |  'canoe',
 475 |  'can opener, tin opener',
 476 |  'cardigan',
 477 |  'car mirror',
 478 |  'carousel, carrousel, merry-go-round, roundabout, whirligig',
 479 |  "carpenter's kit, tool kit",
 480 |  'carton',
 481 |  'car wheel',
 482 |  'cash machine, cash dispenser, automated teller machine, automatic teller machine, automated teller, automatic teller, ATM',
 483 |  'cassette',
 484 |  'cassette player',
 485 |  'castle',
 486 |  'catamaran',
 487 |  'CD player',
 488 |  'cello, violoncello',
 489 |  'cellular telephone, cellular phone, cellphone, cell, mobile phone',
 490 |  'chain',
 491 |  'chainlink fence',
 492 |  'chain mail, ring mail, mail, chain armor, chain armour, ring armor, ring armour',
 493 |  'chain saw, chainsaw',
 494 |  'chest',
 495 |  'chiffonier, commode',
 496 |  'chime, bell, gong',
 497 |  'china cabinet, china closet',
 498 |  'Christmas stocking',
 499 |  'church, church building',
 500 |  'cinema, movie theater, movie theatre, movie house, picture palace',
 501 |  'cleaver, meat cleaver, chopper',
 502 |  'cliff dwelling',
 503 |  'cloak',
 504 |  'clog, geta, patten, sabot',
 505 |  'cocktail shaker',
 506 |  'coffee mug',
 507 |  'coffeepot',
 508 |  'coil, spiral, volute, whorl, helix',
 509 |  'combination lock',
 510 |  'computer keyboard, keypad',
 511 |  'confectionery, confectionary, candy store',
 512 |  'container ship, containership, container vessel',
 513 |  'convertible',
 514 |  'corkscrew, bottle screw',
 515 |  'cornet, horn, trumpet, trump',
 516 |  'cowboy boot',
 517 |  'cowboy hat, ten-gallon hat',
 518 |  'cradle',
 519 |  'crane',
 520 |  'crash helmet',
 521 |  'crate',
 522 |  'crib, cot',
 523 |  'Crock Pot',
 524 |  'croquet ball',
 525 |  'crutch',
 526 |  'cuirass',
 527 |  'dam, dike, dyke',
 528 |  'desk',
 529 |  'desktop computer',
 530 |  'dial telephone, dial phone',
 531 |  'diaper, nappy, napkin',
 532 |  'digital clock',
 533 |  'digital watch',
 534 |  'dining table, board',
 535 |  'dishrag, dishcloth',
 536 |  'dishwasher, dish washer, dishwashing machine',
 537 |  'disk brake, disc brake',
 538 |  'dock, dockage, docking facility',
 539 |  'dogsled, dog sled, dog sleigh',
 540 |  'dome',
 541 |  'doormat, welcome mat',
 542 |  'drilling platform, offshore rig',
 543 |  'drum, membranophone, tympan',
 544 |  'drumstick',
 545 |  'dumbbell',
 546 |  'Dutch oven',
 547 |  'electric fan, blower',
 548 |  'electric guitar',
 549 |  'electric locomotive',
 550 |  'entertainment center',
 551 |  'envelope',
 552 |  'espresso maker',
 553 |  'face powder',
 554 |  'feather boa, boa',
 555 |  'file, file cabinet, filing cabinet',
 556 |  'fireboat',
 557 |  'fire engine, fire truck',
 558 |  'fire screen, fireguard',
 559 |  'flagpole, flagstaff',
 560 |  'flute, transverse flute',
 561 |  'folding chair',
 562 |  'football helmet',
 563 |  'forklift',
 564 |  'fountain',
 565 |  'fountain pen',
 566 |  'four-poster',
 567 |  'freight car',
 568 |  'French horn, horn',
 569 |  'frying pan, frypan, skillet',
 570 |  'fur coat',
 571 |  'garbage truck, dustcart',
 572 |  'gasmask, respirator, gas helmet',
 573 |  'gas pump, gasoline pump, petrol pump, island dispenser',
 574 |  'goblet',
 575 |  'go-kart',
 576 |  'golf ball',
 577 |  'golfcart, golf cart',
 578 |  'gondola',
 579 |  'gong, tam-tam',
 580 |  'gown',
 581 |  'grand piano, grand',
 582 |  'greenhouse, nursery, glasshouse',
 583 |  'grille, radiator grille',
 584 |  'grocery store, grocery, food market, market',
 585 |  'guillotine',
 586 |  'hair slide',
 587 |  'hair spray',
 588 |  'half track',
 589 |  'hammer',
 590 |  'hamper',
 591 |  'hand blower, blow dryer, blow drier, hair dryer, hair drier',
 592 |  'hand-held computer, hand-held microcomputer',
 593 |  'handkerchief, hankie, hanky, hankey',
 594 |  'hard disc, hard disk, fixed disk',
 595 |  'harmonica, mouth organ, harp, mouth harp',
 596 |  'harp',
 597 |  'harvester, reaper',
 598 |  'hatchet',
 599 |  'holster',
 600 |  'home theater, home theatre',
 601 |  'honeycomb',
 602 |  'hook, claw',
 603 |  'hoopskirt, crinoline',
 604 |  'horizontal bar, high bar',
 605 |  'horse cart, horse-cart',
 606 |  'hourglass',
 607 |  'iPod',
 608 |  'iron, smoothing iron',
 609 |  "jack-o'-lantern",
 610 |  'jean, blue jean, denim',
 611 |  'jeep, landrover',
 612 |  'jersey, T-shirt, tee shirt',
 613 |  'jigsaw puzzle',
 614 |  'jinrikisha, ricksha, rickshaw',
 615 |  'joystick',
 616 |  'kimono',
 617 |  'knee pad',
 618 |  'knot',
 619 |  'lab coat, laboratory coat',
 620 |  'ladle',
 621 |  'lampshade, lamp shade',
 622 |  'laptop, laptop computer',
 623 |  'lawn mower, mower',
 624 |  'lens cap, lens cover',
 625 |  'letter opener, paper knife, paperknife',
 626 |  'library',
 627 |  'lifeboat',
 628 |  'lighter, light, igniter, ignitor',
 629 |  'limousine, limo',
 630 |  'liner, ocean liner',
 631 |  'lipstick, lip rouge',
 632 |  'Loafer',
 633 |  'lotion',
 634 |  'loudspeaker, speaker, speaker unit, loudspeaker system, speaker system',
 635 |  "loupe, jeweler's loupe",
 636 |  'lumbermill, sawmill',
 637 |  'magnetic compass',
 638 |  'mailbag, postbag',
 639 |  'mailbox, letter box',
 640 |  'maillot',
 641 |  'maillot, tank suit',
 642 |  'manhole cover',
 643 |  'maraca',
 644 |  'marimba, xylophone',
 645 |  'mask',
 646 |  'matchstick',
 647 |  'maypole',
 648 |  'maze, labyrinth',
 649 |  'measuring cup',
 650 |  'medicine chest, medicine cabinet',
 651 |  'megalith, megalithic structure',
 652 |  'microphone, mike',
 653 |  'microwave, microwave oven',
 654 |  'military uniform',
 655 |  'milk can',
 656 |  'minibus',
 657 |  'miniskirt, mini',
 658 |  'minivan',
 659 |  'missile',
 660 |  'mitten',
 661 |  'mixing bowl',
 662 |  'mobile home, manufactured home',
 663 |  'Model T',
 664 |  'modem',
 665 |  'monastery',
 666 |  'monitor',
 667 |  'moped',
 668 |  'mortar',
 669 |  'mortarboard',
 670 |  'mosque',
 671 |  'mosquito net',
 672 |  'motor scooter, scooter',
 673 |  'mountain bike, all-terrain bike, off-roader',
 674 |  'mountain tent',
 675 |  'mouse, computer mouse',
 676 |  'mousetrap',
 677 |  'moving van',
 678 |  'muzzle',
 679 |  'nail',
 680 |  'neck brace',
 681 |  'necklace',
 682 |  'nipple',
 683 |  'notebook, notebook computer',
 684 |  'obelisk',
 685 |  'oboe, hautboy, hautbois',
 686 |  'ocarina, sweet potato',
 687 |  'odometer, hodometer, mileometer, milometer',
 688 |  'oil filter',
 689 |  'organ, pipe organ',
 690 |  'oscilloscope, scope, cathode-ray oscilloscope, CRO',
 691 |  'overskirt',
 692 |  'oxcart',
 693 |  'oxygen mask',
 694 |  'packet',
 695 |  'paddle, boat paddle',
 696 |  'paddlewheel, paddle wheel',
 697 |  'padlock',
 698 |  'paintbrush',
 699 |  "pajama, pyjama, pj's, jammies",
 700 |  'palace',
 701 |  'panpipe, pandean pipe, syrinx',
 702 |  'paper towel',
 703 |  'parachute, chute',
 704 |  'parallel bars, bars',
 705 |  'park bench',
 706 |  'parking meter',
 707 |  'passenger car, coach, carriage',
 708 |  'patio, terrace',
 709 |  'pay-phone, pay-station',
 710 |  'pedestal, plinth, footstall',
 711 |  'pencil box, pencil case',
 712 |  'pencil sharpener',
 713 |  'perfume, essence',
 714 |  'Petri dish',
 715 |  'photocopier',
 716 |  'pick, plectrum, plectron',
 717 |  'pickelhaube',
 718 |  'picket fence, paling',
 719 |  'pickup, pickup truck',
 720 |  'pier',
 721 |  'piggy bank, penny bank',
 722 |  'pill bottle',
 723 |  'pillow',
 724 |  'ping-pong ball',
 725 |  'pinwheel',
 726 |  'pirate, pirate ship',
 727 |  'pitcher, ewer',
 728 |  "plane, carpenter's plane, woodworking plane",
 729 |  'planetarium',
 730 |  'plastic bag',
 731 |  'plate rack',
 732 |  'plow, plough',
 733 |  "plunger, plumber's helper",
 734 |  'Polaroid camera, Polaroid Land camera',
 735 |  'pole',
 736 |  'police van, police wagon, paddy wagon, patrol wagon, wagon, black Maria',
 737 |  'poncho',
 738 |  'pool table, billiard table, snooker table',
 739 |  'pop bottle, soda bottle',
 740 |  'pot, flowerpot',
 741 |  "potter's wheel",
 742 |  'power drill',
 743 |  'prayer rug, prayer mat',
 744 |  'printer',
 745 |  'prison, prison house',
 746 |  'projectile, missile',
 747 |  'projector',
 748 |  'puck, hockey puck',
 749 |  'punching bag, punch bag, punching ball, punchball',
 750 |  'purse',
 751 |  'quill, quill pen',
 752 |  'quilt, comforter, comfort, puff',
 753 |  'racer, race car, racing car',
 754 |  'racket, racquet',
 755 |  'radiator',
 756 |  'radio, wireless',
 757 |  'radio telescope, radio reflector',
 758 |  'rain barrel',
 759 |  'recreational vehicle, RV, R.V.',
 760 |  'reel',
 761 |  'reflex camera',
 762 |  'refrigerator, icebox',
 763 |  'remote control, remote',
 764 |  'restaurant, eating house, eating place, eatery',
 765 |  'revolver, six-gun, six-shooter',
 766 |  'rifle',
 767 |  'rocking chair, rocker',
 768 |  'rotisserie',
 769 |  'rubber eraser, rubber, pencil eraser',
 770 |  'rugby ball',
 771 |  'rule, ruler',
 772 |  'running shoe',
 773 |  'safe',
 774 |  'safety pin',
 775 |  'saltshaker, salt shaker',
 776 |  'sandal',
 777 |  'sarong',
 778 |  'sax, saxophone',
 779 |  'scabbard',
 780 |  'scale, weighing machine',
 781 |  'school bus',
 782 |  'schooner',
 783 |  'scoreboard',
 784 |  'screen, CRT screen',
 785 |  'screw',
 786 |  'screwdriver',
 787 |  'seat belt, seatbelt',
 788 |  'sewing machine',
 789 |  'shield, buckler',
 790 |  'shoe shop, shoe-shop, shoe store',
 791 |  'shoji',
 792 |  'shopping basket',
 793 |  'shopping cart',
 794 |  'shovel',
 795 |  'shower cap',
 796 |  'shower curtain',
 797 |  'ski',
 798 |  'ski mask',
 799 |  'sleeping bag',
 800 |  'slide rule, slipstick',
 801 |  'sliding door',
 802 |  'slot, one-armed bandit',
 803 |  'snorkel',
 804 |  'snowmobile',
 805 |  'snowplow, snowplough',
 806 |  'soap dispenser',
 807 |  'soccer ball',
 808 |  'sock',
 809 |  'solar dish, solar collector, solar furnace',
 810 |  'sombrero',
 811 |  'soup bowl',
 812 |  'space bar',
 813 |  'space heater',
 814 |  'space shuttle',
 815 |  'spatula',
 816 |  'speedboat',
 817 |  "spider web, spider's web",
 818 |  'spindle',
 819 |  'sports car, sport car',
 820 |  'spotlight, spot',
 821 |  'stage',
 822 |  'steam locomotive',
 823 |  'steel arch bridge',
 824 |  'steel drum',
 825 |  'stethoscope',
 826 |  'stole',
 827 |  'stone wall',
 828 |  'stopwatch, stop watch',
 829 |  'stove',
 830 |  'strainer',
 831 |  'streetcar, tram, tramcar, trolley, trolley car',
 832 |  'stretcher',
 833 |  'studio couch, day bed',
 834 |  'stupa, tope',
 835 |  'submarine, pigboat, sub, U-boat',
 836 |  'suit, suit of clothes',
 837 |  'sundial',
 838 |  'sunglass',
 839 |  'sunglasses, dark glasses, shades',
 840 |  'sunscreen, sunblock, sun blocker',
 841 |  'suspension bridge',
 842 |  'swab, swob, mop',
 843 |  'sweatshirt',
 844 |  'swimming trunks, bathing trunks',
 845 |  'swing',
 846 |  'switch, electric switch, electrical switch',
 847 |  'syringe',
 848 |  'table lamp',
 849 |  'tank, army tank, armored combat vehicle, armoured combat vehicle',
 850 |  'tape player',
 851 |  'teapot',
 852 |  'teddy, teddy bear',
 853 |  'television, television system',
 854 |  'tennis ball',
 855 |  'thatch, thatched roof',
 856 |  'theater curtain, theatre curtain',
 857 |  'thimble',
 858 |  'thresher, thrasher, threshing machine',
 859 |  'throne',
 860 |  'tile roof',
 861 |  'toaster',
 862 |  'tobacco shop, tobacconist shop, tobacconist',
 863 |  'toilet seat',
 864 |  'torch',
 865 |  'totem pole',
 866 |  'tow truck, tow car, wrecker',
 867 |  'toyshop',
 868 |  'tractor',
 869 |  'trailer truck, tractor trailer, trucking rig, rig, articulated lorry, semi',
 870 |  'tray',
 871 |  'trench coat',
 872 |  'tricycle, trike, velocipede',
 873 |  'trimaran',
 874 |  'tripod',
 875 |  'triumphal arch',
 876 |  'trolleybus, trolley coach, trackless trolley',
 877 |  'trombone',
 878 |  'tub, vat',
 879 |  'turnstile',
 880 |  'typewriter keyboard',
 881 |  'umbrella',
 882 |  'unicycle, monocycle',
 883 |  'upright, upright piano',
 884 |  'vacuum, vacuum cleaner',
 885 |  'vase',
 886 |  'vault',
 887 |  'velvet',
 888 |  'vending machine',
 889 |  'vestment',
 890 |  'viaduct',
 891 |  'violin, fiddle',
 892 |  'volleyball',
 893 |  'waffle iron',
 894 |  'wall clock',
 895 |  'wallet, billfold, notecase, pocketbook',
 896 |  'wardrobe, closet, press',
 897 |  'warplane, military plane',
 898 |  'washbasin, handbasin, washbowl, lavabo, wash-hand basin',
 899 |  'washer, automatic washer, washing machine',
 900 |  'water bottle',
 901 |  'water jug',
 902 |  'water tower',
 903 |  'whiskey jug',
 904 |  'whistle',
 905 |  'wig',
 906 |  'window screen',
 907 |  'window shade',
 908 |  'Windsor tie',
 909 |  'wine bottle',
 910 |  'wing',
 911 |  'wok',
 912 |  'wooden spoon',
 913 |  'wool, woolen, woollen',
 914 |  'worm fence, snake fence, snake-rail fence, Virginia fence',
 915 |  'wreck',
 916 |  'yawl',
 917 |  'yurt',
 918 |  'web site, website, internet site, site',
 919 |  'comic book',
 920 |  'crossword puzzle, crossword',
 921 |  'street sign',
 922 |  'traffic light, traffic signal, stoplight',
 923 |  'book jacket, dust cover, dust jacket, dust wrapper',
 924 |  'menu',
 925 |  'plate',
 926 |  'guacamole',
 927 |  'consomme',
 928 |  'hot pot, hotpot',
 929 |  'trifle',
 930 |  'ice cream, icecream',
 931 |  'ice lolly, lolly, lollipop, popsicle',
 932 |  'French loaf',
 933 |  'bagel, beigel',
 934 |  'pretzel',
 935 |  'cheeseburger',
 936 |  'hotdog, hot dog, red hot',
 937 |  'mashed potato',
 938 |  'head cabbage',
 939 |  'broccoli',
 940 |  'cauliflower',
 941 |  'zucchini, courgette',
 942 |  'spaghetti squash',
 943 |  'acorn squash',
 944 |  'butternut squash',
 945 |  'cucumber, cuke',
 946 |  'artichoke, globe artichoke',
 947 |  'bell pepper',
 948 |  'cardoon',
 949 |  'mushroom',
 950 |  'Granny Smith',
 951 |  'strawberry',
 952 |  'orange',
 953 |  'lemon',
 954 |  'fig',
 955 |  'pineapple, ananas',
 956 |  'banana',
 957 |  'jackfruit, jak, jack',
 958 |  'custard apple',
 959 |  'pomegranate',
 960 |  'hay',
 961 |  'carbonara',
 962 |  'chocolate sauce, chocolate syrup',
 963 |  'dough',
 964 |  'meat loaf, meatloaf',
 965 |  'pizza, pizza pie',
 966 |  'potpie',
 967 |  'burrito',
 968 |  'red wine',
 969 |  'espresso',
 970 |  'cup',
 971 |  'eggnog',
 972 |  'alp',
 973 |  'bubble',
 974 |  'cliff, drop, drop-off',
 975 |  'coral reef',
 976 |  'geyser',
 977 |  'lakeside, lakeshore',
 978 |  'promontory, headland, head, foreland',
 979 |  'sandbar, sand bar',
 980 |  'seashore, coast, seacoast, sea-coast',
 981 |  'valley, vale',
 982 |  'volcano',
 983 |  'ballplayer, baseball player',
 984 |  'groom, bridegroom',
 985 |  'scuba diver',
 986 |  'rapeseed',
 987 |  'daisy',
 988 |  "yellow lady's slipper, yellow lady-slipper, Cypripedium calceolus, Cypripedium parviflorum",
 989 |  'corn',
 990 |  'acorn',
 991 |  'hip, rose hip, rosehip',
 992 |  'buckeye, horse chestnut, conker',
 993 |  'coral fungus',
 994 |  'agaric',
 995 |  'gyromitra',
 996 |  'stinkhorn, carrion fungus',
 997 |  'earthstar',
 998 |  'hen-of-the-woods, hen of the woods, Polyporus frondosus, Grifola frondosa',
 999 |  'bolete',
1000 |  'ear, spike, capitulum',
1001 |  'toilet tissue, toilet paper, bathroom tissue'
1002 | }
1003 | 
1004 | 


--------------------------------------------------------------------------------
/images/1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/1.png


--------------------------------------------------------------------------------
/images/10.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/10.png


--------------------------------------------------------------------------------
/images/11.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/11.png


--------------------------------------------------------------------------------
/images/12.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/12.png


--------------------------------------------------------------------------------
/images/13.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/13.png


--------------------------------------------------------------------------------
/images/14.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/14.png


--------------------------------------------------------------------------------
/images/15.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/15.png


--------------------------------------------------------------------------------
/images/16.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/16.png


--------------------------------------------------------------------------------
/images/17.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/17.png


--------------------------------------------------------------------------------
/images/18.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/18.png


--------------------------------------------------------------------------------
/images/19.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/19.png


--------------------------------------------------------------------------------
/images/2.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/2.png


--------------------------------------------------------------------------------
/images/20.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/20.png


--------------------------------------------------------------------------------
/images/3.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/3.png


--------------------------------------------------------------------------------
/images/4.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/4.png


--------------------------------------------------------------------------------
/images/5.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/5.png


--------------------------------------------------------------------------------
/images/6.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/6.png


--------------------------------------------------------------------------------
/images/7.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/7.png


--------------------------------------------------------------------------------
/images/8.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/8.png


--------------------------------------------------------------------------------
/images/9.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/zhanghang1989/visualResNet_torch/5d4e021cc2b2030726bf2f10c2e8dd5475ca72f3/images/9.png


--------------------------------------------------------------------------------
/transforms.lua:
--------------------------------------------------------------------------------
  1 | --
  2 | --  Copyright (c) 2016, Facebook, Inc.
  3 | --  All rights reserved.
  4 | --
  5 | --  This source code is licensed under the BSD-style license found in the
  6 | --  LICENSE file in the root directory of this source tree. An additional grant
  7 | --  of patent rights can be found in the PATENTS file in the same directory.
  8 | --
  9 | --  Image transforms for data augmentation and input normalization
 10 | --
 11 | 
 12 | require 'image'
 13 | 
 14 | local M = {}
 15 | 
 16 | function M.Compose(transforms)
 17 |    return function(input)
 18 |       for _, transform in ipairs(transforms) do
 19 |          input = transform(input)
 20 |       end
 21 |       return input
 22 |    end
 23 | end
 24 | 
 25 | function M.ColorNormalize(meanstd)
 26 |    return function(img)
 27 |       img = img:clone()
 28 |       for i=1,3 do
 29 |          img[i]:add(-meanstd.mean[i])
 30 |          img[i]:div(meanstd.std[i])
 31 |       end
 32 |       return img
 33 |    end
 34 | end
 35 | 
 36 | -- Scales the smaller edge to size
 37 | function M.Scale(size, interpolation)
 38 |    interpolation = interpolation or 'bicubic'
 39 |    return function(input)
 40 |       local w, h = input:size(3), input:size(2)
 41 |       if (w <= h and w == size) or (h <= w and h == size) then
 42 |          return input
 43 |       end
 44 |       if w < h then
 45 |          return image.scale(input, size, h/w * size, interpolation)
 46 |       else
 47 |          return image.scale(input, w/h * size, size, interpolation)
 48 |       end
 49 |    end
 50 | end
 51 | 
 52 | -- Crop to centered rectangle
 53 | function M.CenterCrop(size)
 54 |    return function(input)
 55 |       local w1 = math.ceil((input:size(3) - size)/2)
 56 |       local h1 = math.ceil((input:size(2) - size)/2)
 57 |       return image.crop(input, w1, h1, w1 + size, h1 + size) -- center patch
 58 |    end
 59 | end
 60 | 
 61 | -- Random crop form larger image with optional zero padding
 62 | function M.RandomCrop(size, padding)
 63 |    padding = padding or 0
 64 | 
 65 |    return function(input)
 66 |       if padding > 0 then
 67 |          local temp = input.new(3, input:size(2) + 2*padding, input:size(3) + 2*padding)
 68 |          temp:zero()
 69 |             :narrow(2, padding+1, input:size(2))
 70 |             :narrow(3, padding+1, input:size(3))
 71 |             :copy(input)
 72 |          input = temp
 73 |       end
 74 | 
 75 |       local w, h = input:size(3), input:size(2)
 76 |       if w == size and h == size then
 77 |          return input
 78 |       end
 79 | 
 80 |       local x1, y1 = torch.random(0, w - size), torch.random(0, h - size)
 81 |       local out = image.crop(input, x1, y1, x1 + size, y1 + size)
 82 |       assert(out:size(2) == size and out:size(3) == size, 'wrong crop size')
 83 |       return out
 84 |    end
 85 | end
 86 | 
 87 | -- Four corner patches and center crop from image and its horizontal reflection
 88 | function M.TenCrop(size)
 89 |    local centerCrop = M.CenterCrop(size)
 90 | 
 91 |    return function(input)
 92 |       local w, h = input:size(3), input:size(2)
 93 | 
 94 |       local output = {}
 95 |       for _, img in ipairs{input, image.hflip(input)} do
 96 |          table.insert(output, centerCrop(img))
 97 |          table.insert(output, image.crop(img, 0, 0, size, size))
 98 |          table.insert(output, image.crop(img, w-size, 0, w, size))
 99 |          table.insert(output, image.crop(img, 0, h-size, size, h))
100 |          table.insert(output, image.crop(img, w-size, h-size, w, h))
101 |       end
102 | 
103 |       -- View as mini-batch
104 |       for i, img in ipairs(output) do
105 |          output[i] = img:view(1, img:size(1), img:size(2), img:size(3))
106 |       end
107 | 
108 |       return input.cat(output, 1)
109 |    end
110 | end
111 | 
112 | -- Resized with shorter side randomly sampled from [minSize, maxSize] (ResNet-style)
113 | function M.RandomScale(minSize, maxSize)
114 |    return function(input)
115 |       local w, h = input:size(3), input:size(2)
116 | 
117 |       local targetSz = torch.random(minSize, maxSize)
118 |       local targetW, targetH = targetSz, targetSz
119 |       if w < h then
120 |          targetH = torch.round(h / w * targetW)
121 |       else
122 |          targetW = torch.round(w / h * targetH)
123 |       end
124 | 
125 |       return image.scale(input, targetW, targetH, 'bicubic')
126 |    end
127 | end
128 | 
129 | -- Random crop with size 8%-100% and aspect ratio 3/4 - 4/3 (Inception-style)
130 | function M.RandomSizedCrop(size)
131 |    local scale = M.Scale(size)
132 |    local crop = M.CenterCrop(size)
133 | 
134 |    return function(input)
135 |       local attempt = 0
136 |       repeat
137 |          local area = input:size(2) * input:size(3)
138 |          local targetArea = torch.uniform(0.08, 1.0) * area
139 | 
140 |          local aspectRatio = torch.uniform(3/4, 4/3)
141 |          local w = torch.round(math.sqrt(targetArea * aspectRatio))
142 |          local h = torch.round(math.sqrt(targetArea / aspectRatio))
143 | 
144 |          if torch.uniform() < 0.5 then
145 |             w, h = h, w
146 |          end
147 | 
148 |          if h <= input:size(2) and w <= input:size(3) then
149 |             local y1 = torch.random(0, input:size(2) - h)
150 |             local x1 = torch.random(0, input:size(3) - w)
151 | 
152 |             local out = image.crop(input, x1, y1, x1 + w, y1 + h)
153 |             assert(out:size(2) == h and out:size(3) == w, 'wrong crop size')
154 | 
155 |             return image.scale(out, size, size, 'bicubic')
156 |          end
157 |          attempt = attempt + 1
158 |       until attempt >= 10
159 | 
160 |       -- fallback
161 |       return crop(scale(input))
162 |    end
163 | end
164 | 
165 | function M.HorizontalFlip(prob)
166 |    return function(input)
167 |       if torch.uniform() < prob then
168 |          input = image.hflip(input)
169 |       end
170 |       return input
171 |    end
172 | end
173 | 
174 | function M.Rotation(deg)
175 |    return function(input)
176 |       if deg ~= 0 then
177 |          input = image.rotate(input, (torch.uniform() - 0.5) * deg * math.pi / 180, 'bilinear')
178 |       end
179 |       return input
180 |    end
181 | end
182 | 
183 | -- Lighting noise (AlexNet-style PCA-based noise)
184 | function M.Lighting(alphastd, eigval, eigvec)
185 |    return function(input)
186 |       if alphastd == 0 then
187 |          return input
188 |       end
189 | 
190 |       local alpha = torch.Tensor(3):normal(0, alphastd)
191 |       local rgb = eigvec:clone()
192 |          :cmul(alpha:view(1, 3):expand(3, 3))
193 |          :cmul(eigval:view(1, 3):expand(3, 3))
194 |          :sum(2)
195 |          :squeeze()
196 | 
197 |       input = input:clone()
198 |       for i=1,3 do
199 |          input[i]:add(rgb[i])
200 |       end
201 |       return input
202 |    end
203 | end
204 | 
205 | local function blend(img1, img2, alpha)
206 |    return img1:mul(alpha):add(1 - alpha, img2)
207 | end
208 | 
209 | local function grayscale(dst, img)
210 |    dst:resizeAs(img)
211 |    dst[1]:zero()
212 |    dst[1]:add(0.299, img[1]):add(0.587, img[2]):add(0.114, img[3])
213 |    dst[2]:copy(dst[1])
214 |    dst[3]:copy(dst[1])
215 |    return dst
216 | end
217 | 
218 | function M.Saturation(var)
219 |    local gs
220 | 
221 |    return function(input)
222 |       gs = gs or input.new()
223 |       grayscale(gs, input)
224 | 
225 |       local alpha = 1.0 + torch.uniform(-var, var)
226 |       blend(input, gs, alpha)
227 |       return input
228 |    end
229 | end
230 | 
231 | function M.Brightness(var)
232 |    local gs
233 | 
234 |    return function(input)
235 |       gs = gs or input.new()
236 |       gs:resizeAs(input):zero()
237 | 
238 |       local alpha = 1.0 + torch.uniform(-var, var)
239 |       blend(input, gs, alpha)
240 |       return input
241 |    end
242 | end
243 | 
244 | function M.Contrast(var)
245 |    local gs
246 | 
247 |    return function(input)
248 |       gs = gs or input.new()
249 |       grayscale(gs, input)
250 |       gs:fill(gs[1]:mean())
251 | 
252 |       local alpha = 1.0 + torch.uniform(-var, var)
253 |       blend(input, gs, alpha)
254 |       return input
255 |    end
256 | end
257 | 
258 | function M.RandomOrder(ts)
259 |    return function(input)
260 |       local img = input.img or input
261 |       local order = torch.randperm(#ts)
262 |       for i=1,#ts do
263 |          img = ts[order[i]](img)
264 |       end
265 |       return input
266 |    end
267 | end
268 | 
269 | function M.ColorJitter(opt)
270 |    local brightness = opt.brightness or 0
271 |    local contrast = opt.contrast or 0
272 |    local saturation = opt.saturation or 0
273 | 
274 |    local ts = {}
275 |    if brightness ~= 0 then
276 |       table.insert(ts, M.Brightness(brightness))
277 |    end
278 |    if contrast ~= 0 then
279 |       table.insert(ts, M.Contrast(contrast))
280 |    end
281 |    if saturation ~= 0 then
282 |       table.insert(ts, M.Saturation(saturation))
283 |    end
284 | 
285 |    if #ts == 0 then
286 |       return function(input) return input end
287 |    end
288 | 
289 |    return M.RandomOrder(ts)
290 | end
291 | 
292 | return M
293 | 
294 | 


--------------------------------------------------------------------------------
/visual.lua:
--------------------------------------------------------------------------------
 1 | 
 2 | require 'nn'
 3 | require 'cunn'
 4 | require 'cudnn'
 5 | require 'image'
 6 | local t = require 'transforms'
 7 | local imagenetLabel = require './imagenet'
 8 | 
 9 | if #arg < 2 then
10 |    io.stderr:write('Usage: th visual.lua [MODEL] [FILE]...\n')
11 |    os.exit(1)
12 | end
13 | for _, f in ipairs(arg) do
14 |    if not paths.filep(f) then
15 |       io.stderr:write('file not found: ' .. f .. '\n')
16 |       os.exit(1)
17 |    end
18 | end
19 | 
20 | 
21 | -- Load the model
22 | local model = torch.load(arg[1])
23 | local net = model:clone()
24 | local softMaxLayer = cudnn.SoftMax():cuda()
25 | 
26 | -- add Softmax layer
27 | model:add(softMaxLayer)
28 | 
29 | -- Evaluate mode
30 | model:evaluate()
31 | 
32 | -- For visualization
33 | weight_lr = net:get(#net.modules).weight:clone()
34 | bias_lr = net:get(#net.modules).bias:clone()
35 | net:remove(#net.modules)
36 | net:remove(#net.modules)
37 | net:remove(#net.modules)
38 | net:add(cudnn.SpatialConvolution(2048,1000,1,1,1,1,0,0))
39 | net:get(#net.modules).weight = torch.reshape(weight_lr, net:get(#net.modules).weight:size())
40 | net:get(#net.modules).bias   = torch.reshape(bias_lr, net:get(#net.modules).bias:size())
41 | net:cuda()
42 | net:evaluate()
43 | 
44 | -- The model was trained with this input normalization
45 | local meanstd = {
46 |    mean = { 0.485, 0.456, 0.406 },
47 |    std = { 0.229, 0.224, 0.225 },
48 | }
49 | 
50 | local transform = t.Compose{
51 |    t.Scale(256),
52 |    t.ColorNormalize(meanstd),
53 |    t.CenterCrop(224),
54 | }
55 | 
56 | local N = 3
57 | 
58 | for i=2,#arg do
59 |   -- load the image as a RGB float tensor with values 0..1
60 |   local img = image.load(arg[i], 3, 'float')
61 |   local im = img:clone()
62 |   local name = paths.basename(arg[i], 'JPEG')
63 | 
64 |   -- Scale, normalize, and crop the image
65 |   img = transform(img)
66 | 
67 |    -- View as mini-batch of size 1
68 |   local batch = img:view(1, table.unpack(img:size():totable()))
69 | 
70 |    -- Get the output of the softmax
71 |   local output = model:forward(batch:cuda()):squeeze()
72 | 
73 |    -- Get the featuremap
74 |   local featuremap = net:forward(batch:cuda()):squeeze()
75 | 
76 |    -- Get the top N class indexes and probabilities
77 |   local probs, indexes = output:topk(N, true, true)
78 |   print('Classes for', arg[i])
79 | 
80 |   for n=1,1 do
81 |     print(probs[n], imagenetLabel[indexes[n]])
82 |     map = image.scale(featuremap[indexes[n]]:float(), img:size(2),img:size(3))
83 |     im = image.scale(im, img:size(2),img:size(3))
84 |     map = (torch.div((map - torch.min(map)), (torch.max(map)-torch.min(map))):mul(255)+1):round()
85 |     map = image.y2jet(map)
86 |     map = img:mul(0.2)+map:mul(0.8)
87 |     frame = torch.cat(im, map, 3)
88 |     frame=image.drawText(frame, imagenetLabel[indexes[n]], 10, 10)
89 |     image.save(string.format('images/%s.png', name), frame)
90 |   end
91 | end
92 | 
93 | 


--------------------------------------------------------------------------------