gaussian_cdf_pytorch.py

import math
import torch
from torch.distributions import Normal

# standard univariate Gaussian (Normal)
mean = torch.zeros(1)
std = torch.ones(1)

# evaluate from -0.5 to 0.5
x_min = -0.5 * torch.ones(1)
x_max = 0.5 * torch.ones(1)

# using built-in functions
normal = Normal(mean, std)
prob_mass = normal.cdf(x_max) - normal.cdf(x_min)

# implemented from scratch
def cdf(mu, sigma, val):
    return 0.5 * (1 + torch.erf((value - mu) / (math.sqrt(2) * sigma)))

prob_mass = cdf(mean, std, x_max) - cdf(mean, std, x_min)