pdf = MultivariateNormal(torch.zeros(2, dtype=pdf_dtype), covariance_matrix=torch.eye(2, dtype=pdf_dtype))
def log_pdf(theta, x, y):
return pdf.log_prob(theta)
model = DistributionModel(log_pdf, 2, dtype=pdf.loc.dtype)
# %% Setup PowerPosteriorSampler
num_chains = 5
per_chain_samplers = [['MALA', {'step': 0.25}] for _ in range(num_chains)]
sampler = PowerPosteriorSampler(
model,
DataLoader(EmptyXYDataset()),
per_chain_samplers,
theta0=torch.tensor([-1, 1], dtype=model.dtype),
between_step=1,
check_input=True
)
# %% Run PowerPosteriorSampler
start_time = timer()
sampler.run(num_epochs=11000, num_burnin_epochs=1000, verbose=True, verbose_step=1000)
end_time = timer()
print("Time taken: {}".format(timedelta(seconds=end_time-start_time)))
pdf = MultivariateNormal(torch.zeros(2, dtype=pdf_dtype),
covariance_matrix=torch.eye(2, dtype=pdf_dtype))
def log_pdf(theta, x, y):
return pdf.log_prob(theta)
model = DistributionModel(log_pdf, 2, dtype=pdf.loc.dtype)
# %% Setup HMC sampler
sampler = HMC(model,
theta0=torch.tensor([-1, 1], dtype=model.dtype),
dataloader=DataLoader(EmptyXYDataset()),
tuner=HMCDATuner(1., e0=2.))
# %% Run HMC sampler
sampler.run(num_epochs=11000,
num_burnin_epochs=1000,
verbose=True,
verbose_step=1000)
# %% Compute acceptance rate
print('Acceptance rate: {}'.format(sampler.get_chain().acceptance_rate()))
# %% Compute Monte Carlo mean
# [
# MultivariateNormal(means[0], covariance_matrix=covs[0]),
# MultivariateNormal(means[1], covariance_matrix=covs[1])
# ]))
def log_pdf(theta, x, y):
return torch.log(
torch.exp(-0.5 * (torch.dot(theta-means[0], theta-means[0])))
+ torch.exp(-0.5 * torch.dot(theta-means[1], theta-means[1]))
)
model = DistributionModel(log_pdf, 2, dtype=pdf_dtype)
# %% Setup RAM sampler
sampler = RAM(model, theta0=torch.tensor([0., 0.], dtype=model.dtype), dataloader=DataLoader(EmptyXYDataset()))
# %% Run RAM sampler
sampler.run(num_epochs=11000, num_burnin_epochs=1000)
# %% Compute acceptance rate
print('Acceptance rate: {}'.format(sampler.get_chain().acceptance_rate()))
# %% Compute Monte Carlo mean
print('Monte Carlo mean: {}'.format(sampler.get_chain().mean()))
# %% Compute posterior covariance matrix