def test_logpdf_1d():
x = tf.constant([0.5])
val_true = stats.beta.logpdf([0.5], 0.5, 0.5)
_assert_eq(beta.logpdf(x, tf.constant(0.5), tf.constant(0.5)), val_true)
_assert_eq(beta.logpdf(x, tf.constant([0.5]), tf.constant(0.5)), val_true)
_assert_eq(beta.logpdf(x, tf.constant(0.5), tf.constant([0.5])), val_true)
_assert_eq(beta.logpdf(x, tf.constant([0.5]), tf.constant([0.5])), val_true)
def _test_logpdf_scalar(x, a=0.5, b=0.5):
xtf = tf.constant(x)
val_true = stats.beta.logpdf(x, a, b)
_assert_eq(beta.logpdf(xtf, tf.constant(a), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([a]), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant(a), tf.constant([b])), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([a]), tf.constant([b])), val_true)
def _test(self, x, a, b):
xtf = tf.constant(x)
val_true = stats.beta.logpdf(x, a, b)
with self.test_session():
self.assertAllClose(beta.logpdf(xtf, a, b).eval(), val_true)
self.assertAllClose(beta.logpdf(xtf, tf.constant(a),
tf.constant(b)).eval(), val_true)
def _test_logpdf_scalar(x, a=0.5, b=0.5):
xtf = tf.constant(x)
val_true = stats.beta.logpdf(x, a, b)
_assert_eq(beta.logpdf(xtf, tf.constant(a), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([a]), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant(a), tf.constant([b])), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([a]), tf.constant([b])), val_true)
def _test_logpdf_scalar(scalar, a=.5, b=.5):
x = tf.constant(scalar)
val_true = stats.beta.logpdf(scalar, a, b)
_assert_eq(beta.logpdf(x, tf.constant(a), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(x, tf.constant([a]), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(x, tf.constant(a), tf.constant([b])), val_true)
_assert_eq(beta.logpdf(x, tf.constant([a]), tf.constant([b])), val_true)
def _test(x, a, b):
xtf = tf.constant(x)
val_true = stats.beta.logpdf(x, a, b)
_assert_eq(beta.logpdf(xtf, a, b), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant(a), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([a]), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant(a), tf.constant([b])), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([a]), tf.constant([b])), val_true)
def _test(x, a, b):
xtf = tf.constant(x)
val_true = stats.beta.logpdf(x, a, b)
_assert_eq(beta.logpdf(xtf, a, b), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant(a), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([a]), tf.constant(b)), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant(a), tf.constant([b])), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([a]), tf.constant([b])), val_true)
def _test(self, x, a, b):
xtf = tf.constant(x)
val_true = stats.beta.logpdf(x, a, b)
with self.test_session():
self.assertAllClose(beta.logpdf(xtf, a, b).eval(), val_true)
self.assertAllClose(beta.logpdf(xtf, tf.constant(a), tf.constant(b)).eval(), val_true)
self.assertAllClose(beta.logpdf(xtf, tf.constant([a]), tf.constant(b)).eval(), val_true)
self.assertAllClose(beta.logpdf(xtf, tf.constant(a), tf.constant([b])).eval(), val_true)
self.assertAllClose(beta.logpdf(xtf, tf.constant([a]), tf.constant([b])).eval(), val_true)
def test_logpdf_1d():
x = [0.5]
xtf = tf.constant([0.5])
val_true = stats.beta.logpdf(x, 0.5, 0.5)
_assert_eq(beta.logpdf(xtf, tf.constant(0.5), tf.constant(0.5)), val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([0.5]), tf.constant(0.5)),
val_true)
_assert_eq(beta.logpdf(xtf, tf.constant(0.5), tf.constant([0.5])),
val_true)
_assert_eq(beta.logpdf(xtf, tf.constant([0.5]), tf.constant([0.5])),
val_true)
def log_prob(self, xs, zs):
log_prior = beta.logpdf(zs, a=1.0, b=1.0)
log_lik = tf.pack([
tf.reduce_sum(bernoulli.logpmf(xs['x'], z)) for z in tf.unpack(zs)
])
return log_lik + log_prior
def log_prob_idx(self, idx, xs):
full_idx = (slice(0, None), ) + idx # slice over batch size
return beta.logpdf(xs[full_idx], self.alpha[idx], self.beta[idx])
def log_prob_zi(self, i, zs):
"""log q(z_i | lambda)"""
if i >= self.num_factors:
raise IndexError()
return beta.logpdf(zs[:, i], self.a[i], self.b[i])
def log_prob_i(self, i, xs):
"""log p(x_i | params)"""
if i >= self.num_factors:
raise IndexError()
return beta.logpdf(xs[:, i], self.alpha[i], self.beta[i])
def log_prob_idx(self, idx, xs):
full_idx = (slice(0, None), ) + idx # slice over sample size
return beta.logpdf(xs[full_idx], self.alpha[idx], self.beta[idx])
def log_prob_zi(self, i, z):
"""log q(z_i | lambda_i)"""
if i >= self.num_vars:
raise
return beta.logpdf(z[:, i], self.a[i], self.b[i])
def _test(self, x, a, b):
val_true = stats.beta.logpdf(x, a, b)
with self.test_session():
self.assertAllClose(beta.logpdf(x, a=a, b=b).eval(), val_true)
def log_prob(self, xs, zs): log_prior = beta.logpdf(zs['p'], a=1.0, b=1.0) log_lik = tf.reduce_sum(bernoulli.logpmf(xs['x'], p=zs['p'])) return log_lik + log_prior
def log_prob(self, xs, zs):
log_prior = beta.logpdf(zs, a=1.0, b=1.0)
log_lik = tf.pack([tf.reduce_sum(bernoulli.logpmf(xs['x'], z))
for z in tf.unpack(zs)])
return log_lik + log_prior
def _test(self, x, a, b):
val_true = stats.beta.logpdf(x, a, b)
with self.test_session():
self.assertAllClose(beta.logpdf(x, a=a, b=b).eval(), val_true)
def log_prob_zi(self, i, zs):
"""log q(z_i | lambda)"""
if i >= self.num_factors:
raise IndexError()
return beta.logpdf(zs[:, i], self.a[i], self.b[i])
def log_prob_zi(self, i, z):
"""log q(z_i | lambda_i)"""
if i >= self.num_vars:
raise
return beta.logpdf(z[:, i], self.a[i], self.b[i])
def log_prob(self, xs, zs):
log_prior = beta.logpdf(zs['p'], a=1.0, b=1.0)
log_lik = tf.reduce_sum(bernoulli.logpmf(xs['x'], p=zs['p']))
return log_lik + log_prior
def log_prob(self, xs, zs):
log_prior = beta.logpdf(zs[:, 0], a=1.0, b=1.0)
log_lik = tf.concat(0, [
tf.reduce_sum(bernoulli.logpmf(xs, z)) \
for z in tf.unpack(zs)])
return log_lik + log_prior