they mean the number of the training samples? `X_final = np.zeros((3516, img_size, img_size, 3), dtype=np.uint8)` `X_final[i+3268]=image`