Missed TensorFlow World? Check out the recap. Learn more

tf.random.normal

View source on GitHub

Outputs random values from a normal distribution.

Aliases:

  • tf.compat.v1.random.normal
  • tf.compat.v1.random_normal
  • tf.compat.v2.random.normal
tf.random.normal(
    shape,
    mean=0.0,
    stddev=1.0,
    dtype=tf.dtypes.float32,
    seed=None,
    name=None
)

Used in the guide:

Used in the tutorials:

Args:

  • shape: A 1-D integer Tensor or Python array. The shape of the output tensor.
  • mean: A 0-D Tensor or Python value of type dtype. The mean of the normal distribution.
  • stddev: A 0-D Tensor or Python value of type dtype. The standard deviation of the normal distribution.
  • dtype: The type of the output.
  • seed: A Python integer. Used to create a random seed for the distribution. See tf.compat.v1.set_random_seed for behavior.
  • name: A name for the operation (optional).

Returns:

A tensor of the specified shape filled with random normal values.