def _test(p, n):
rv = Bernoulli(p=p)
rv_sample = rv.sample(n)
x = rv_sample.eval()
x_tf = tf.constant(x, dtype=tf.float32)
p = p.eval()
assert np.allclose(rv.log_prob(x_tf).eval(), stats.bernoulli.logpmf(x, p))
class ZeroInflatedRV(RandomVariable, Distribution):
"""
A zero-inflated random variable. The prob_zero parameter defines the
probability of inflation.
"""
def __init__(self, prob_zero, underlying, *args, **kwargs):
self.prob_zero = prob_zero
self.underlying = underlying
self.bernoulli = Bernoulli(probs=self.prob_zero) # for sampling
super(ZeroInflatedRV, self).__init__(
*args,
**kwargs,
dtype=underlying.dtype,
validate_args=underlying.validate_args,
allow_nan_stats=underlying.allow_nan_stats,
reparameterization_type=underlying.reparameterization_type)
def _log_prob(self, value):
not_zero_lp = self.underlying.log_prob(value)
return tf.where(
tf.equal(value, tf.zeros_like(value)),
tf.log(self.prob_zero +
(1. - self.prob_zero * tf.exp(not_zero_lp))),
tf.log(1. - self.prob_zero) + not_zero_lp)
def _sample_n(self, n, seed=None):
zero = self.bernoulli.sample(n, seed=seed)
return tf.where(tf.equal(tf.constant(1), zero),
tf.zeros_like(zero, dtype=self.dtype),
self.underlying.sample(n))
def _test(shape, n):
rv = Bernoulli(shape, p=tf.zeros(shape) + 0.5)
rv_sample = rv.sample(n)
x = rv_sample.eval()
x_tf = tf.constant(x, dtype=tf.float32)
p = rv.p.eval()
for idx in range(shape[0]):
assert np.allclose(rv.log_prob_idx((idx,), x_tf).eval(), stats.bernoulli.logpmf(x[:, idx], p[idx]))
def _test(p, n):
rv = Bernoulli(p=p)
rv_sample = rv.sample(n)
x = rv_sample.eval()
x_tf = tf.constant(x, dtype=tf.float32)
p = p.eval()
assert np.allclose(rv.log_prob(x_tf).eval(),
stats.bernoulli.logpmf(x, p))
def _test(shape, n):
rv = Bernoulli(shape, p=tf.zeros(shape) + 0.5)
rv_sample = rv.sample(n)
x = rv_sample.eval()
x_tf = tf.constant(x, dtype=tf.float32)
p = rv.p.eval()
for idx in range(shape[0]):
assert np.allclose(
rv.log_prob_idx((idx, ), x_tf).eval(),
stats.bernoulli.logpmf(x[:, idx], p[idx]))
def _test(shape, n):
# using Bernoulli's internally implemented log_prob_idx() to check
# Distribution's log_prob()
rv = Bernoulli(shape, p=tf.zeros(shape)+0.5)
rv_sample = rv.sample(n)
x = rv_sample.eval()
x_tf = tf.constant(x, dtype=tf.float32)
p = rv.p.eval()
val_ed = rv.log_prob(x_tf).eval()
val_true = 0.0
for idx in range(shape[0]):
val_true += stats.bernoulli.logpmf(x[:, idx], p[idx])
assert np.allclose(val_ed, val_true)
def _test(shape, n):
# using Bernoulli's internally implemented log_prob_idx() to check
# Distribution's log_prob()
rv = Bernoulli(shape, p=tf.zeros(shape)+0.5)
rv_sample = rv.sample(n)
with sess.as_default():
x = rv_sample.eval()
x_tf = tf.constant(x, dtype=tf.float32)
p = rv.p.eval()
val_ed = rv.log_prob(x_tf).eval()
val_true = 0.0
for idx in range(shape[0]):
val_true += stats.bernoulli.logpmf(x[:, idx], p[idx])
assert np.allclose(val_ed, val_true)
def _test(p, n):
x = Bernoulli(p=p)
val_est = get_dims(x.sample(n))
val_true = n + get_dims(p)
assert val_est == val_true
def _test(shape, p, n):
x = Bernoulli(shape, p)
val_est = tuple(get_dims(x.sample(n)))
val_true = (n, ) + shape
assert val_est == val_true
from __future__ import print_function, division, absolute_import
from data import get_value
import tensorflow as tf
import edward as ed
from edward.models import Beta, Bernoulli
theta = Beta(a=1.0, b=1.0)
# 100-dimensional Bernoulli
x = Bernoulli(p=tf.ones(12) * theta)
# ====== sampling from each marginal variables
theta_sample = theta.sample()
x_sample = x.sample()
print("Marginal theta samples:", get_value(theta_sample))
print("Marginal X samples:", get_value(x_sample))
# ====== sampling from the joint distribution
samples = get_value([x.value(), theta.value()])
print("From joint distribution:")
print("- X:", samples[0])
print("- theta:", samples[1])
def _test(shape, p, n):
x = Bernoulli(shape, p)
val_est = tuple(get_dims(x.sample(n)))
val_true = (n, ) + shape
assert val_est == val_true
def _test(p, n): x = Bernoulli(p=p) val_est = get_dims(x.sample(n)) val_true = n + get_dims(p) assert val_est == val_true
def _test(self, probs, n):
rv = Bernoulli(probs)
dist = ds.Bernoulli(probs)
x = rv.sample(n).eval()
self.assertAllEqual(rv.log_prob(x).eval(), dist.log_prob(x).eval())
def _test(self, probs, n):
rv = Bernoulli(probs)
dist = ds.Bernoulli(probs)
self.assertEqual(rv.sample(n).shape, dist.sample(n).shape)
def _test(self, probs, n): rv = Bernoulli(probs) dist = ds.Bernoulli(probs) x = rv.sample(n).eval() self.assertAllEqual(rv.log_prob(x).eval(), dist.log_prob(x).eval())
def _test(shape, p, size):
x = Bernoulli(shape, p)
val_est = tuple(get_dims(x.sample(size=size)))
val_true = (size, ) + shape
assert val_est == val_true
