suvojit-0x55aa / a2s2k-resnet Goto Github PK

Colab link is showing error of, file doesnot exist

attention_vector = torch.cat( [ self.conv_ex(Z).unsqueeze(dim=1), self.conv_ex(Z).unsqueeze(dim=1) ], dim=1)
attention_vector = self.softmax(attention_vector)
and self.softmax = nn.Softmax(dim=1)
it seems that the elements of the attention_vector are the same, so if you apply softmax on dim=1，the result of the softmax will all be the same, 0.5 for sure

so why are we doing this，i don't know if i have missed something

Thank you !The model you proposed was very useful to me !!

record.record_output( OA, AA, KAPPA, ELEMENT_ACC, TRAINING_TIME, TESTING_TIME, './report/' + 'SSRNpatch:' + str(img_rows) + '_' + Dataset + 'split' + str(VALIDATION_SPLIT) + 'lr' + str(lr) + PARAM_OPTIM + '.txt')
Utils.generate_png( all_iter, net, gt_hsi, Dataset, device, total_indices, './classification_maps/' + 'SSRNpatch:' + str(img_rows) + '_' + Dataset + 'split' + str(VALIDATION_SPLIT) + 'lr' + str(lr) + PARAM_OPTIM)
the report and classification folder need to be created additionally,Right?
And there is no output '.txt' and '.png'file generated as program descripition.

Good afternoon,

I have question, how do you calculate what would be your TOTAL_SIZE value in data loader:

def load_dataset(Dataset, split=0.9):
data_path = '../dataset/'
if Dataset == 'IN':
mat_data = sio.loadmat(data_path + 'Indian_pines_corrected.mat')
mat_gt = sio.loadmat(data_path + 'Indian_pines_gt.mat')
data_hsi = mat_data['indian_pines_corrected']
gt_hsi = mat_gt['indian_pines_gt']
K = 200
TOTAL_SIZE = 10249 # THIS VALUE
VALIDATION_SPLIT = split
TRAIN_SIZE = math.ceil(TOTAL_SIZE * VALIDATION_SPLIT)

I read in your paper that you call it Total sample pixels, but not sure how this value can be calculated?

Thank you and appreciate your effort for this paper, very interesting approach.