Truncated normal initialization (approximation)

trunc_normal_(x, mean = 0, std = 1)

Arguments

x

tensor

mean

mean

std

standard deviation

Value

tensor