In a sense, we converted one image into billions of training data points, massively reducing the amount of data needed for training.

从某种意义上说， 我们将一张图像转换为数十亿个训练数据点，从而大大减少了训练所需的数据量。