将矩阵从 (100, ) 重塑为 (100, 405416)

Question

I am trying to reshape a one-hot encoded array to (samples, 405416) . The output shape after one-hot encoded is (405416, 405416) and I want it to reshape to (100, 405416) where 100 is the samples.

y = []
for i in range(100):
    random_num = randint(0, 405416, 1)
    y.append(data_enc[random_number])

The output y shape is (100, )

The output y[0] shape is (1, 405416)

data_enc is the matrix of shape (405416, 405416) after one-hot encoding

Answer 1

Reshape of a matrix should contain same number of elements from input to reshaped matrix. so reshaping (405416, 405416) to (100, 405416) is not possible.

When you have 100 samples then the onehot encoding should result (100, features) not (features, features). Check your implementation of Onehot clearly.