def test_poisson_likelihood(rate: float, hybridize: bool) -> None:
"""
Test to check that maximizing the likelihood recovers the parameters
"""
# generate samples
rates = mx.nd.zeros(NUM_SAMPLES) + rate
distr = Poisson(rates)
samples = distr.sample()
init_biases = [inv_softplus(rate - START_TOL_MULTIPLE * TOL * rate)]
rate_hat = maximum_likelihood_estimate_sgd(
PoissonOutput(),
samples,
init_biases=init_biases,
hybridize=hybridize,
learning_rate=PositiveFloat(0.05),
num_epochs=PositiveInt(20),
)
print("rate:", rate_hat)
assert (np.abs(rate_hat[0] - rate) < TOL *
rate), f"mu did not match: rate = {rate}, rate_hat = {rate_hat}"
(3, 4, 5),
(),
),
(
MixtureDistributionOutput([
MultivariateGaussianOutput(dim=5),
MultivariateGaussianOutput(dim=5),
]),
mx.nd.random.normal(shape=(3, 4, 10)),
[None, mx.nd.ones(shape=(3, 4, 5))],
[None, mx.nd.ones(shape=(3, 4, 5))],
(3, 4),
(5, ),
),
(
PoissonOutput(),
mx.nd.random.normal(shape=(3, 4, 5, 6)),
[None],
[None, mx.nd.ones(shape=(3, 4, 5))],
(3, 4, 5),
(),
),
(
DeterministicOutput(42.0),
mx.nd.random.normal(shape=(3, 4, 5, 6)),
[None, mx.nd.ones(shape=(3, 4, 5))],
[None, mx.nd.ones(shape=(3, 4, 5))],
(3, 4, 5),
(),
),
],
