@符号增加正态扰动的含义？ about eat_tensorflow2_in_30_days HOT 5 CLOSED

This means noise with normal distribution, which refers to the second term tf.random.normal.

from eat_tensorflow2_in_30_days.

那增加正态扰动的意思，是为了让随机数“更随机”？，但即使@符号与tf.random.normal不在一起也能影响到tf.random.normal吗还是说@符号会影响整条语句？ 谢谢！

from eat_tensorflow2_in_30_days.

X@w0 means matrix multiplication between X and w0. The @ mark has a higher priority than adding. b0 is a bias. This part (X@w0+b0) represents the fundamental linear operator in the neural network, which is irrelevant to the random number.

On the other hand, tf.random.normal is for adding a (pseudo-) randomized (or stochastic) noise.

from eat_tensorflow2_in_30_days.

好的，虽然只听懂了前半部分，感觉后半部分的知识我还没接触过，日后再来请教！感谢

from eat_tensorflow2_in_30_days.

我感觉作者的意思应该是：
对于【X@w0 + b0】这部分，“@”表示矩阵乘法
对于【torch.normal( 0.0,2.0,size = [n,1])】这部分，则是为了增加正态扰动

from eat_tensorflow2_in_30_days.