Overlapping between meta-training classes and meta-testing classes

Hi. Thank you for publishing the code. I am trying to do an experiment with Few-shot learning benchmark datasets (i.e., CIFAR-FS and FC100) with the data loader. While as I understand, there is no overlapping between meta-training classes and meta-testing classes (i.e., as the description of the FC100 dataset from the link here https://paperswithcode.com/dataset/fc100).

However, when I check the label sets of train, validation, and test set, there is still overlapping (very seriously) as below :

print('** Labels set: TRAIN **')
print(np.unique(dataset_train.labels))
print(np.shape(dataset_train.data))
print('** Labels set: VAL **')
print(np.unique(dataset_val.labels))
print(np.shape(dataset_val.data))
print('** Labels set: TEST **')
print(np.unique(dataset_test.labels))
print(np.shape(dataset_test.data))

Output:

** Labels set: TRAIN **                                                                                                                                             
[ 0 1 5 8 9 10 12 13 16 17 20 22 23 25 27 28 29 32 33 37 39 40 41 44 47 48 49 51 52 53 54 56 57 58 59 60 61 62 67 68 69 70 71 73 76 78 81 82 83 84 85 86 87 89 90 91 92 93 94 96] 
(36000, 32, 32, 3)
** Labels set: VAL **                                                                                                                                               
[ 0 1 3 5 8 9 10 12 13 15 16 17 19 20 21 22 23 25 26 27 28 29 31 32 33 36 37 38 39 40 41 42 43 44 45 47 48 49 50 51 52 53 54 56 57 58 59 60  61 62 65 67 68 69 70 71 73 74 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 96 97 99]                                                                                                                                           
(48000, 32, 32, 3)                                                                                                                                                  
** Labels set: TEST **                                                                                                                                              
[ 0 1 2 4 5 6 7 8 9 10 11 12 13 14 16 17 18 20 22 23 24 25 27 28 29 30 32 33 34 35 37 39 40 41 44 46 47 48 49 51 52 53 54 55 56 57 58 59 60 61 62 63 64 66 67 68 69 70 71 72 73 75 76 78 81 82 83 84 85 86 87 89 90 91 92 93 94 95 96 98]                                                                                                                                           
(48000, 32, 32, 3)

Because I am a newbie with this few-shot learning setting therefore it should be my miss understanding in some parts of the few-shot learning setting but I don’t know what it is? Could you please help? Thanks a lot.

kjunelee / MetaOptNet

Overlapping between meta-training classes and meta-testing classes #55