def test_sum_squared_error():
"""Tests probflow.utils.metrics.sum_squared_error"""
# Predictive dist
preds = tf.constant([0, 1, 2, 0, 0, 0], dtype=tf.float32)
pred_dist = Normal(preds, 1)
y_true = np.array([0, 0, 0, 0, 1, 2]).astype("float32")
# Compare metric
assert is_close(metrics.sum_squared_error(y_true, pred_dist.mean()), 10.0)
def test_r_squared():
"""Tests probflow.utils.metrics.r_squared"""
# Predictive dist
preds = tf.constant([0, 1, 2, 2, 2], dtype=tf.float32)
pred_dist = Normal(preds, 1)
y_true = np.array([0, 1, 2, 3, 4]).astype("float32")
# Compare metric
assert is_close(metrics.r_squared(y_true, pred_dist.mean()), 0.5)
def test_Normal():
"""Tests Normal distribution"""
# Create the distribution
dist = Normal()
# Check default params
assert dist.loc == 0
assert dist.scale == 1
# Call should return backend obj
assert isinstance(dist(), tod.normal.Normal)
# Test methods
npdf = lambda x, m, s: (1.0 / np.sqrt(2 * np.pi * s * s) * np.exp(
-np.power(x - m, 2) / (2 * s * s)))
assert is_close(dist.prob(0).numpy(), npdf(0, 0, 1))
assert is_close(dist.prob(1).numpy(), npdf(1, 0, 1))
assert is_close(dist.log_prob(0).numpy(), np.log(npdf(0, 0, 1)))
assert is_close(dist.log_prob(1).numpy(), np.log(npdf(1, 0, 1)))
assert dist.mean().numpy() == 0.0
# Test sampling
samples = dist.sample()
assert isinstance(samples, torch.Tensor)
assert samples.ndim == 0
samples = dist.sample(10)
assert isinstance(samples, torch.Tensor)
assert samples.ndim == 1
assert samples.shape[0] == 10
# Should be able to set params
dist = Normal(loc=3, scale=2)
assert dist.loc == 3
assert dist.scale == 2
# But only with Tensor-like objs
with pytest.raises(TypeError):
dist = Normal(loc="lalala", scale="lalala")
with pytest.raises(TypeError):
dist = Normal(loc=0, scale="lalala")
with pytest.raises(TypeError):
dist = Normal(loc="lalala", scale=1)
