当前位置：网站首页>random.normal() and random.truncated_normal() in TF

random.normal() and random.truncated_normal() in TF

2022-08-09 23:11:00 【phac123】

Brief

The tf.random.normal() and tf.random.truncated_normal() methods are commonly used in TF to create Tensors that conform to the normal distribution.
The difference between them is that the latter guarantees the value of randomly generated dataBetween (μ-2σ, μ+2σ), if the generated data exceeds this range, re-generate.
μ: mean, σ: standard deviation

Code

import tensorflow as tfd = tf.random.normal([2,span> 2], mean = 0.5, stddev = 1)print("d: ", d)e = tf.random.truncated_normal([2,span> 2], mean = 0.5, stddev = 1)print("e: ", e)