imagerettrieval -pytorch

[shiva1407]

i am using Kaggle to run my code
for i, row in df_train.iterrows():
id = int(row.id)
for j in range(aug_size):
im = cv2.imread(path.join(dir,f'{id:.0f}.{j}.png'))
x[c * aug_size + j, 0, :, :] = t.from_numpy(im).squeeze()
y[c * aug_size + j][0] = row.malignancy_th
c += 1

errors met:
RuntimeError: expand(torch.ByteTensor{[32, 32, 3]}, size=[32, 32]): the number of sizes provided (2) must be greater or equal to the number of dimensions in the tensor (3)

[shiva1407]

please help me to fix this bug

[filipradenovic]

Please take a look at our cirtorch.examples.test script to see how we load and preprocess images.