[PYTHON] Tensorflow API: tf.truncated_normal

https://www.tensorflow.org/api_docs/python/tf/truncated_normal

Overview

Generate a tensor randomly extracted from the truncated normal distribution of ± 2σ.

truncated_normal(
    shape,
    mean=0.0,
    stddev=1.0,
    dtype=tf.float32,
    seed=None,
    name=None
)

• shape: Return tensor dimension (specify a one-dimensional integer list or tensor)
• mean: Mean of the generated truncated normal distribution
• stddev: Standard deviation of the generated cut normal distribution
• dtype: Return tensor type
• If seed: integer is specified, the same value will be extracted each time the session changes.
• name: The name of the operation

Example of use

import tensorflow as tf

x = tf.truncated_normal(shape=[2, 2], mean=0.0, stddev=1.0, dtype=tf.float32)
with tf.Session() as sess:
    x.eval()