variational_samplers = [
ProbabilisticModel([
NormalVariable(loc=loc1,
scale=0.1,
name="weights1",
learnable=True),
NormalVariable(loc=loc2,
scale=0.1,
name="weights2",
learnable=True)
]) for loc1, loc2 in zip(initial_locations1, initial_locations2)
]
# Inference
inference_method = WVGD(variational_samplers=variational_samplers,
particles=particles,
biased=False)
inference.perform_inference(model,
inference_method=inference_method,
number_iterations=1000,
number_samples=100,
optimizer="Adam",
lr=0.0025,
posterior_model=particles,
pretraining_iterations=0)
loss_list = model.diagnostics["loss curve"]
# Test accuracy
test_size = len(ind[dataset_size:])
num_images = test_size * 3
test_indices = RandomIndices(dataset_size=test_size,
particle_2 = DeterministicVariable(initial_location_2, name="weights", learnable=True)
particle_locations = [particle_1, particle_2]
particles = [ProbabilisticModel([l]) for l in particle_locations]
# Importance sampling distributions
voranoi_set = VoronoiSet(particle_locations) #TODO: Bug if you use variables instead of probabilistic models
variational_samplers = [ProbabilisticModel([TruncatedNormalVariable(mu=initial_location_1, sigma=0.1,
truncation_rule=lambda a: voranoi_set(a, 0),
name="weights", learnable=True)]),
ProbabilisticModel([TruncatedNormalVariable(mu=initial_location_2, sigma=0.1,
truncation_rule=lambda a: voranoi_set(a, 1),
name="weights", learnable=True)])]
# Inference
inference.perform_inference(model,
inference_method=WVGD(biased=True),
number_iterations=1000,
number_samples=50,
optimizer=chainer.optimizers.Adam(0.005),
posterior_model=particles,
sampler_model=variational_samplers,
pretraining_iterations=0)
loss_list = model.diagnostics["loss curve"]
# Local variational models
plt.plot(loss_list)
plt.show()
# Test accuracy
num_images = 2000
test_size = len(test)
#initial_locations = [0, 0.1]
particles = [
ProbabilisticModel([RootVariable(p, name="theta", learnable=True)])
for p in initial_locations
]
# Importance sampling distributions
variational_samplers = [
ProbabilisticModel([
NormalVariable(loc=location, scale=0.2, name="theta", learnable=True)
]) for location in initial_locations
]
# Inference
inference_method = WVGD(variational_samplers=variational_samplers,
particles=particles,
biased=False,
number_post_samples=80000)
inference.perform_inference(model,
inference_method=inference_method,
number_iterations=1500,
number_samples=50,
optimizer="Adam",
lr=0.005,
posterior_model=particles,
pretraining_iterations=0)
loss_list = model.diagnostics["loss curve"]
# Local variational models
plt.plot(loss_list)
plt.show()
RootVariable(b, name="b", learnable=True)
]) for wk, wl, b in zip(wk_locations, wl_locations, b_locations)
]
# Importance sampling distributions
variational_samplers = [
ProbabilisticModel([
NormalVariable(wk, 0.1 + 0 * wk, name="Wk", learnable=True),
NormalVariable(wl, 0.1 + 0 * wl, name="Wl", learnable=True),
NormalVariable(b, 0.1 + 0 * b, name="b", learnable=True)
]) for wk, wl, b in zip(wk_locations, wl_locations, b_locations)
]
# Inference
inference_method = WVGD(variational_samplers=variational_samplers,
particles=particles,
number_post_samples=500,
biased=True)
inference.perform_inference(
model,
inference_method=inference_method,
number_iterations=1500, #4000
number_samples=20, #10
optimizer="Adam",
lr=0.05,
posterior_model=particles,
pretraining_iterations=0)
loss_list = model.diagnostics["loss curve"]
plt.plot(loss_list)
# ELBO
ELBO = model.posterior_model.estimate_log_model_evidence(