Cifar_100 Targets Labels doubt

1

I am writing a way to assign coarse label (super class) for each label in the CIFAR 100 dataset.
Is it like this below how they are assigned in pytorch while reading the data. like class 0 =beaver ,class 1 is dolphin …etc or anyother way?

(when I print classes it’s printing in alphabetical order in pytorch like 'apple', 'aquarium_fish', 'baby', 'bear', 'beaver which mean apple -0 , aquarium_fish 1 ,,etc ? AM I correct or now? or the below Image is correct?).

Just want to want to which one is correct. So I can create superclass for all labels.(went through images but as they are of small sizes it’s a bit confusion) . Any help appreciated.

Screen Shot 2020-05-15 at 1.28.11 AM
Screen Shot 2020-05-15 at 1.28.11 AM2222×1080 321 KB

2

I did this not sure right or wrong. I read the test file downloaded by the pytorch dataloader

class Dictlist(dict):
    def __setitem__(self, key, value):
        try:
            self[key]
        except KeyError:
            super(Dictlist, self).__setitem__(key, [])
        self[key].append(value)

def unpickle(file):
    import pickle
    with open(file, 'rb') as fo:
        dict = pickle.load(fo, encoding='bytes')
    return dict
x=unpickle('cifar-100-python/test')
fine_to_coarse=Dictlist()

for i in range(0,len(x[b'coarse_labels'])):
    fine_to_coarse[x[b'coarse_labels'][i]]=x[ b'fine_labels'][i]
d=dict(fine_to_coarse)
for i in d.keys():
    d[i]=list(dict.fromkeys(d[i]))

Finally got the below

0: [72, 4, 95, 30, 55]
1: [73, 32, 67, 91, 1]
2: [92, 70, 82, 54, 62]
3: [16, 61, 9, 10, 28]
4: [51, 0, 53, 57, 83]
5: [40, 39, 22, 87, 86]
6: [20, 25, 94, 84, 5]
7: [14, 24, 6, 7, 18]
8: [43, 97, 42, 3, 88]
9: [37, 17, 76, 12, 68]
10: [49, 33, 71, 23, 60]
11: [15, 21, 19, 31, 38]
12: [75, 63, 66, 64, 34]
13: [77, 26, 45, 99, 79]
14: [11, 2, 35, 46, 98]
15: [29, 93, 27, 78, 44]
16: [65, 50, 74, 36, 80]
17: [56, 52, 47, 59, 96]
18: [8, 58, 90, 13, 48]
19: [81, 69, 41, 89, 85]

for i in range(0,20):
    print("Sub classes for Super class ",i,end=': ')
    for j in d[i]:
        print(dataset.cifar100.classes[j],end=',')
    print("")

Final Output is

Sub classes for Super class  0: seal,beaver,whale,dolphin,otter,
Sub classes for Super class  1: shark,flatfish,ray,trout,aquarium_fish,
Sub classes for Super class  2: tulip,rose,sunflower,orchid,poppy,
Sub classes for Super class  3: can,plate,bottle,bowl,cup,
Sub classes for Super class  4: mushroom,apple,orange,pear,sweet_pepper,
Sub classes for Super class  5: lamp,keyboard,clock,television,telephone,
Sub classes for Super class  6: chair,couch,wardrobe,table,bed,
Sub classes for Super class  7: butterfly,cockroach,bee,beetle,caterpillar,
Sub classes for Super class  8: lion,wolf,leopard,bear,tiger,
Sub classes for Super class  9: house,castle,skyscraper,bridge,road,
Sub classes for Super class  10: mountain,forest,sea,cloud,plain,
Sub classes for Super class  11: camel,chimpanzee,cattle,elephant,kangaroo,
Sub classes for Super class  12: skunk,porcupine,raccoon,possum,fox,
Sub classes for Super class  13: snail,crab,lobster,worm,spider,
Sub classes for Super class  14: boy,baby,girl,man,woman,
Sub classes for Super class  15: dinosaur,turtle,crocodile,snake,lizard,
Sub classes for Super class  16: rabbit,mouse,shrew,hamster,squirrel,
Sub classes for Super class  17: palm_tree,oak_tree,maple_tree,pine_tree,willow_tree,
Sub classes for Super class  18: bicycle,pickup_truck,train,bus,motorcycle,
Sub classes for Super class  19: streetcar,rocket,lawn_mower,tractor,tank,

I hope I am right as it matched the actual image from CIFAR website