def model():
p2 = torch.tensor(torch.ones(2) / 2)
p3 = torch.tensor(torch.ones(3) / 3)
x2 = pyro.sample("x2", dist.OneHotCategorical(p2))
x3 = pyro.sample("x3", dist.OneHotCategorical(p3))
assert x2.shape == torch.Size([2]) + iarange_shape + p2.shape
assert x3.shape == torch.Size([3, 1]) + iarange_shape + p3.shape
def model(self, xs, ys=None, aux_scale=46):
pyro.module("ss_vae", self)
batch_size = xs.size(0)
options = dict(out=None,
dtype=xs.dtype,
layout=torch.strided,
device=xs.device,
requires_grad=False)
with pyro.plate("data"):
prior_loc = torch.zeros(batch_size, 50, **options)
prior_scale = torch.ones(batch_size, 50, **options)
zs2 = pyro.sample("z2",
dist.Normal(prior_loc, prior_scale).to_event(1))
alpha_prior = torch.ones(batch_size, 4, **options) / 4.0
ys_ = pyro.sample("y", dist.OneHotCategorical(alpha_prior), obs=ys)
z1_mean, z1_std = self.decoder_z1(zs2, ys_)
zs1 = pyro.sample("z1", dist.Normal(z1_mean, z1_std).to_event(1))
x_mean, x_std = self.decoder_x(zs1)
pyro.sample('x', dist.Normal(x_mean, x_std).to_event(3), obs=xs)
if ys is not None:
alpha = self.encoder_y(zs1)
# with pyro.poutine.scale(scale=self.aux_scale):
with pyro.poutine.scale(scale=aux_scale):
pyro.sample("y_aux", dist.OneHotCategorical(alpha), obs=ys)
def test_batch_log_dims(dim, probs):
batch_pdf_shape = (3,) + (1,) * (dim-1)
expected_log_prob_sum = np.array(wrap_nested(list(np.log(probs)), dim-1)).reshape(*batch_pdf_shape)
probs = modify_params_using_dims(probs, dim)
support = dist.OneHotCategorical(probs).enumerate_support()
log_prob = dist.OneHotCategorical(probs).log_prob(support)
assert_equal(log_prob.detach().cpu().numpy(), expected_log_prob_sum)
def generate_name(self,
lstm: Decoder,
address: str,
batch_size: int,
hidd_cell_states: tuple = None,
sample: bool = True):
"""
lstm: Decoder associated with name being generated
address: The address to correlate pyro distribution with latent variables
hidd_cell_states: Previous LSTM hidden state or empty hidden state
max_name_length: The max name length allowed
"""
# If no hidden state is provided, initialize it with all 0s
if hidd_cell_states == None:
hidd_cell_states = lstm.init_hidden(batch_size=batch_size)
input_tensor = strings_to_tensor([SOS] * batch_size, 1,
letter_to_index)
names = [''] * batch_size
for index in range(MAX_NAME_LENGTH):
char_dist, hidd_cell_states = lstm.forward(input_tensor,
hidd_cell_states)
if sample:
# Next LSTM input is the sampled character
input_tensor = pyro.sample(f"unsup_{address}_{index}",
dist.OneHotCategorical(char_dist))
chars_at_indexes = list(
map(lambda index: MODEL_CHARS[int(index.item())],
torch.argmax(input_tensor, dim=2).squeeze(0)))
else:
# Next LSTM input is the character with the highest probability of occurring
pyro.sample(f"unsup_{address}_{index}",
dist.OneHotCategorical(char_dist))
chars_at_indexes = list(
map(lambda index: MODEL_CHARS[int(index.item())],
torch.argmax(char_dist, dim=2).squeeze(0)))
input_tensor = strings_to_tensor(chars_at_indexes, 1,
letter_to_index)
# Add sampled characters to names
for i, char in enumerate(chars_at_indexes):
names[i] += char
# Discard everything after EOS character
# names = list(map(lambda name: name[:name.find(EOS)] if name.find(EOS) > -1 else name, names))
return hidd_cell_states, names
def model_classify(self, xs, lengths, ys=None):
"""
this model is used to add an auxiliary (supervised) loss as described in the
Kingma et al., "Semi-Supervised Learning with Deep Generative Models".
"""
xs = pad_sequence(xs, batch_first=True, padding_value=self.padding_idx)
if ys is not None:
ys = torch.stack(ys)
if self.use_cuda:
xs = xs.cuda()
lengths = lengths.cuda()
if ys is not None:
ys = ys.cuda()
#Get embeddings
xs = self.embeddings(xs)
# register all pytorch (sub)modules with pyro
pyro.module("text_ss_vae", self)
# inform Pyro that the variables in the batch of xs, ys are conditionally independent
with pyro.plate("data"):
# this here is the extra term to yield an auxiliary loss that we do gradient descent on
if ys is not None:
alpha = self.encoder_y.forward(xs, lengths)
with pyro.poutine.scale(scale=self.aux_loss_multiplier):
pyro.sample("y_aux", dist.OneHotCategorical(alpha), obs=ys)
def one_hot_model(pseudocounts, classes=None):
probs_prior = dist.Dirichlet(pseudocounts)
probs = pyro.sample("probs", probs_prior)
with pyro.plate("classes",
classes.size(0) if classes is not None else 1,
dim=-1):
return pyro.sample("obs", dist.OneHotCategorical(probs), obs=classes)
def model(self, data, label=None):
pyro.module("decoder", self)
batch_size = data.size(0)
with pyro.plate("data"):
deformation_loc = torch.zeros([batch_size, 1])
deformation_scale = torch.ones([batch_size, 1])
label_prior = data.new_ones([batch_size, self.num_classes
]) / (1.0 * self.num_classes)
deformation = pyro.sample(
"deformation",
dist.Normal(deformation_loc, deformation_scale).to_event(1))
label = pyro.sample("label",
dist.OneHotCategorical(label_prior),
obs=label)
final_image = self.decoder(deformation, label, data)
pyro.sample("image",
dist.Bernoulli(final_image).to_event(1),
obs=data)
def test_one_hot_categorical_shape():
ps = ng_ones(3, 2) / 2
d = dist.OneHotCategorical(ps)
assert d.batch_shape() == (3, )
assert d.event_shape() == (2, )
assert d.shape() == (3, 2)
assert d.sample().size() == d.shape()
def model(self, x, y=None):
# Register various nn.Modules with Pyro
pyro.module("scanvi", self)
# This gene-level parameter modulates the variance of the observation distribution
theta = pyro.param("inverse_dispersion", 10.0 * x.new_ones(self.num_genes),
constraint=constraints.positive)
# We scale all sample statements by scale_factor so that the ELBO is normalized
# wrt the number of datapoints and genes
with pyro.plate("batch", len(x)), poutine.scale(scale=self.scale_factor):
z1 = pyro.sample("z1", dist.Normal(0, x.new_ones(self.latent_dim)).to_event(1))
# Note that if y is None (i.e. y is unobserved) then y will be sampled;
# otherwise y will be treated as observed.
y = pyro.sample("y", dist.OneHotCategorical(logits=x.new_zeros(self.num_labels)),
obs=y)
z2_loc, z2_scale = self.z2_decoder(z1, y)
z2 = pyro.sample("z2", dist.Normal(z2_loc, z2_scale).to_event(1))
l_scale = self.l_scale * x.new_ones(1)
l = pyro.sample("l", dist.LogNormal(self.l_loc, l_scale).to_event(1))
# Note that by construction mu is normalized (i.e. mu.sum(-1) == 1) and the
# total scale of counts for each cell is determined by `l`
gate_logits, mu = self.x_decoder(z2)
# TODO revisit this parameterization if torch.distributions.NegativeBinomial changes
# from failure to success parametrization;
# see https://github.com/pytorch/pytorch/issues/42449
nb_logits = (l * mu + self.epsilon).log() - (theta + self.epsilon).log()
x_dist = dist.ZeroInflatedNegativeBinomial(gate_logits=gate_logits, total_count=theta,
logits=nb_logits)
# Observe the datapoint x using the observation distribution x_dist
pyro.sample("x", x_dist.to_event(1), obs=x)
def guide(self, x, y=None):
pyro.module("scanvi", self)
with pyro.plate("batch",
len(x)), poutine.scale(scale=self.scale_factor):
z2_loc, z2_scale, l_loc, l_scale = self.z2l_encoder(x)
pyro.sample("l", dist.LogNormal(l_loc, l_scale).to_event(1))
z2 = pyro.sample("z2", dist.Normal(z2_loc, z2_scale).to_event(1))
y_logits = self.classifier(z2)
y_dist = dist.OneHotCategorical(logits=y_logits)
if y is None:
# x is unlabeled so sample y using q(y|z2)
y = pyro.sample("y", y_dist)
else:
# x is labeled so add a classification loss term
# (this way q(y|z2) learns from both labeled and unlabeled data)
classification_loss = y_dist.log_prob(y)
# Note that the negative sign appears because we're adding this term in the guide
# and the guide log_prob appears in the ELBO as -log q
pyro.factor(
"classification_loss",
-self.alpha * classification_loss,
has_rsample=False,
)
z1_loc, z1_scale = self.z1_encoder(z2, y)
pyro.sample("z1", dist.Normal(z1_loc, z1_scale).to_event(1))
def model(self, xs, ys):
# register this pytorch module and all of its sub-modules with pyro
pyro.module("ss_vae", self)
batch_size = xs.size(0)
# inform Pyro that the variables in the batch of xs, ys are conditionally independent
with pyro.plate("data"):
# sample the handwriting style from the constant prior distribution
prior_loc = xs.new_zeros([batch_size, self.z_dim])
prior_scale = xs.new_ones([batch_size, self.z_dim])
zs = pyro.sample("z",
dist.Normal(prior_loc, prior_scale).to_event(1))
# if the label y (which digit to write) is supervised, sample from the
# constant prior, otherwise, observe the value (i.e. score it against the constant prior)
alpha_prior = xs.new_ones([batch_size, self.output_size
]) / (1.0 * self.output_size)
ys = pyro.sample("y", dist.OneHotCategorical(alpha_prior), obs=ys)
# finally, score the image (x) using the handwriting style (z) and
# the class label y (which digit to write) against the
# parametrized distribution p(x|y,z) = bernoulli(decoder(y,z))
# where `decoder` is a neural network
loc = self.decoder.forward([zs, ys])
pyro.sample("x", dist.Bernoulli(loc).to_event(1), obs=xs)
def guide(self, xs, ys=None):
"""
The guide corresponds to the following:
q(y|x) = categorical(alpha(x)) # infer digit from an image
q(z|x,y) = normal(loc(x,y),scale(x,y)) # infer handwriting style from an image and the digit
loc, scale are given by a neural network `encoder_z`
alpha is given by a neural network `encoder_y`
:param xs: a batch of scaled vectors of pixels from an image
:param ys: (optional) a batch of the class labels i.e.
the digit corresponding to the image(s)
:return: None
"""
# inform Pyro that the variables in the batch of xs, ys are conditionally independent
with pyro.plate("data"):
# if the class label (the digit) is not supervised, sample
# (and score) the digit with the variational distribution
# q(y|x) = categorical(alpha(x))
if ys is None:
alpha = self.encoder_y.forward(xs)
ys = pyro.sample("y", dist.OneHotCategorical(alpha))
# sample (and score) the latent handwriting-style with the variational
# distribution q(z|x,y) = normal(loc(x,y),scale(x,y))
loc, scale = self.encoder_z.forward([xs, ys])
pyro.sample("z", dist.Normal(loc, scale).to_event(1))
def model_classify(self, xs, ys=None):
pyro.module("ss_vae", self)
with pyro.plate("data"):
if ys is not None:
z1, _ = self.encoder_z1(xs)
alpha = self.encoder_y(z1)
pyro.sample("y_aux", dist.OneHotCategorical(alpha), obs=ys)
def gmm_batch_guide(data):
with pyro.iarange("data", len(data)) as batch:
n = len(batch)
probs = pyro.param("probs", torch.tensor(torch.ones(n, 1) * 0.6, requires_grad=True))
probs = torch.cat([probs, 1 - probs], dim=1)
z = pyro.sample("z", dist.OneHotCategorical(probs))
assert z.shape[-2:] == (n, 2)
def model(self, xs, ys=None):
# can pass in ys as labels or not pass in anything for unlabelled
# register PyTorch module `decoder` with Pyro
pyro.module("decoder", self.decoder)
# with pyro.plate("data"):
with pyro.iarange("data", xs.shape[0]):
# setup hyperparameters for prior p(z)
z_loc = xs.new_zeros(torch.Size((xs.shape[0], self.z_dim)))
z_scale = xs.new_ones(torch.Size((xs.shape[0], self.z_dim)))
# sample from prior (value will be sampled by guide when computing the ELBO)
zs = pyro.sample("z", dist.Normal(z_loc, z_scale).independent(1))
# if there is a label y, sample from the constant prior
# otherwise, observe the value (i.e. score against constant prior)
# alpha_prior = xs.new_ones(torch.Size((xs.shape[0], self.output_size))) / (1.0*self.output_size)
# prior is the actual train data label distribution (not uniform)
alpha_prior = self.prior_probs.repeat(xs.shape[0], 1)
ys = pyro.sample("y", dist.OneHotCategorical(alpha_prior), obs=ys)
# decoder outputs mean and sqroot cov, sample from normal
recon_loc, recon_scale = self.decoder.forward([zs, ys])
pyro.sample("x",
dist.Normal(recon_loc, recon_scale).independent(1),
obs=xs.reshape(-1, 1335))
def model(self, X_u: list, X_s: list, Z_s: dict, observations=None):
"""
Model for generating names representing p(x,z)
x: Training data (name string)
z: Optionally supervised latent values (dictionary of name/format values)
"""
pyro.module("model_fn_lstm", self.model_fn_lstm)
formatted_X_u = strings_to_tensor(X_u, MAX_NAME_LENGTH,
printable_to_index)
formatted_X_s = strings_to_tensor(X_s, MAX_NAME_LENGTH,
printable_to_index)
with pyro.plate("sup_batch", len(X_s)):
_, first_names = self.generate_name_supervised(
self.model_fn_lstm,
FIRST_NAME_ADD,
len(X_s),
observed=Z_s[FIRST_NAME_ADD])
full_names = list(
map(lambda name: pad_string(name, MAX_NAME_LENGTH),
first_names))
probs = strings_to_probs(full_names,
MAX_NAME_LENGTH,
printable_to_index,
true_index_prob=self.peak_prob)
pyro.sample("sup_output",
dist.OneHotCategorical(probs.transpose(0,
1)).to_event(1),
obs=formatted_X_s.transpose(0, 1))
with pyro.plate("unsup_batch", len(X_u)):
_, first_names = self.generate_name(self.model_fn_lstm,
FIRST_NAME_ADD, len(X_u))
full_names = list(
map(lambda name: pad_string(name, MAX_NAME_LENGTH),
first_names))
probs = strings_to_probs(full_names,
MAX_NAME_LENGTH,
printable_to_index,
true_index_prob=self.peak_prob)
pyro.sample("unsup_output",
dist.OneHotCategorical(probs.transpose(0,
1)).to_event(1),
obs=formatted_X_u.transpose(0, 1))
return full_names
def remap_y(self, ys):
new_ys = []
options = dict(dtype=ys.dtype, device=ys.device)
for i, label_length in enumerate(self.label_shape):
prior = torch.ones(ys.size(0), label_length, **options) / (1.0 * label_length)
new_ys.append(pyro.sample("y_%s" % self.label_names[i], dist.OneHotCategorical(prior),
obs=torch.nn.functional.one_hot(ys[:,i].to(torch.int64), int(label_length))))
new_ys = torch.cat(new_ys, -1)
return new_ys.to(torch.float32)
def model(self, images, labels=None, kl_factor=1.0):
images = images.view(-1, 784)
n_images = images.size(0)
# Set-up parameters for the distribution of weights for each layer `a<n>`
a1_mean = torch.zeros(784, self.n_hidden)
a1_scale = torch.ones(784, self.n_hidden)
a1_dropout = torch.tensor(0.25)
a2_mean = torch.zeros(self.n_hidden + 1, self.n_classes)
a2_scale = torch.ones(self.n_hidden + 1, self.n_hidden)
a2_dropout = torch.tensor(1.0)
a3_mean = torch.zeros(self.n_hidden + 1, self.n_classes)
a3_scale = torch.ones(self.n_hidden + 1, self.n_hidden)
a3_dropout = torch.tensor(1.0)
a4_mean = torch.zeros(self.n_hidden + 1, self.n_classes)
a4_scale = torch.ones(self.n_hidden + 1, self.n_classes)
# Mark batched calculations to be conditionally independent given parameters using `plate`
with pyro.plate('data', size=n_images):
# Sample first hidden layer
h1 = pyro.sample(
'h1',
bnn.HiddenLayer(images,
a1_mean,
a1_dropout * a1_scale,
non_linearity=nnf.leaky_relu,
KL_factor=kl_factor))
# Sample second hidden layer
h2 = pyro.sample(
'h2',
bnn.HiddenLayer(h1,
a2_mean,
a2_dropout * a2_scale,
non_linearity=nnf.leaky_relu,
KL_factor=kl_factor))
# Sample third hidden layer
h3 = pyro.sample(
'h3',
bnn.HiddenLayer(h2,
a3_mean,
a3_dropout * a3_scale,
non_linearity=nnf.leaky_relu,
KL_factor=kl_factor))
# Sample output logits
logits = pyro.sample(
'logits',
bnn.HiddenLayer(
h3,
a4_mean,
a4_scale,
non_linearity=lambda x: nnf.log_softmax(x, dim=-1),
KL_factor=kl_factor,
include_hidden_bias=False))
# One-hot encode labels
labels = nnf.one_hot(labels) if labels is not None else None
# Condition on observed labels, so it calculates the log-likehood loss when training using VI
return pyro.sample('label',
dist.OneHotCategorical(logits=logits),
obs=labels)
def model_classify(self, xs, ys=None):
# register PyTorch module `encoder` with Pyro
pyro.module("encoder_y", self.encoder_y)
# with pyro.plate("data")
with pyro.iarange("data", xs.shape[0]):
# this here is the extra term to yield an auxiliary loss that we do gradient descent on
if ys is not None:
alpha = self.encoder_y.forward(xs)
with pyro.poutine.scale(scale=self.aux_loss_multiplier):
pyro.sample("y_aux", dist.OneHotCategorical(alpha), obs=ys)
def gmm_batch_model(data):
p = pyro.param("p", torch.tensor([0.3], requires_grad=True))
p = torch.cat([p, 1 - p])
scale = pyro.param("scale", torch.tensor([1.0], requires_grad=True))
mus = torch.tensor([-1.0, 1.0])
with pyro.iarange("data", len(data)) as batch:
n = len(batch)
z = pyro.sample("z", dist.OneHotCategorical(p).expand_by([n]))
assert z.shape[-2:] == (n, 2)
loc = (z * mus).sum(-1)
pyro.sample("x", dist.Normal(loc, scale.expand(n)), obs=data[batch])
def test_basevae_encode_xy():
data_dim = (2, 64)
x = torch.randn(*data_dim)
alpha = torch.ones(data_dim[0], 3) / 3
y = dist.OneHotCategorical(alpha).sample()
vae = models.base.baseVAE(data_dim[1:], None)
encoder_net = nets.fcEncoderNet(data_dim[1:], 2, 3)
vae.set_encoder(encoder_net)
encoded = vae._encode(x, y)
assert_equal(encoded[:, :2].shape, (data_dim[0], 2))
assert_equal(encoded[:, 2:].shape, (data_dim[0], 2))
def test_posterior_predictive_svi_one_hot():
pseudocounts = torch.ones(3) * 0.1
true_probs = torch.tensor([0.15, 0.6, 0.25])
classes = dist.OneHotCategorical(true_probs).sample((10000,))
guide = AutoDelta(one_hot_model)
svi = SVI(one_hot_model, guide, optim.Adam(dict(lr=0.1)), Trace_ELBO())
for i in range(1000):
svi.step(pseudocounts, classes=classes)
posterior_samples = Predictive(guide, num_samples=10000).get_samples(pseudocounts)
posterior_predictive = Predictive(one_hot_model, posterior_samples)
marginal_return_vals = posterior_predictive.get_samples(pseudocounts)["obs"]
assert_close(marginal_return_vals.mean(dim=0), true_probs.unsqueeze(0), rtol=0.1)
def guide_ncls(self, xs, ys=None):
with pyro.plate("data"):
z1_mean, z1_std = self.encoder_z1(xs)
zs1 = pyro.sample("z1", dist.Normal(z1_mean, z1_std).to_event(1))
if ys is None:
alpha = self.encoder_y(zs1)
ys = pyro.sample("y", dist.OneHotCategorical(alpha))
z2_mean, z2_std = self.encoder_z2(zs1, ys)
zs2 = pyro.sample("z2", dist.Normal(z2_mean, z2_std).to_event(1))
pass
def guide(self, data, label=None):
pyro.module("encoder", self)
with pyro.plate("data"):
label_prior, deformation_loc, deformation_scale = self.encoder(
data)
pyro.sample("label", dist.OneHotCategorical(label_prior))
pyro.sample("deformation",
dist.Normal(deformation_loc, deformation_scale))
def test_support_dims(dim, probs):
probs = modify_params_using_dims(probs, dim)
d = dist.OneHotCategorical(probs)
support = d.enumerate_support()
for s in support:
assert_correct_dimensions(s, probs)
n = len(support)
assert support.shape == (n,) + d.batch_shape + d.event_shape
support_expanded = d.enumerate_support(expand=True)
assert support_expanded.shape == (n,) + d.batch_shape + d.event_shape
support_unexpanded = d.enumerate_support(expand=False)
assert support_unexpanded.shape == (n,) + (1,) * len(d.batch_shape) + d.event_shape
def model(self, xs, ys=None):
"""
The model corresponds to the following generative process:
p(z) = normal(0,I) # handwriting style (latent)
p(y|x) = categorical(I/10.) # which digit (semi-supervised)
p(x|y,z) = bernoulli(mu(y,z)) # an image
mu is given by a neural network `decoder`
:param xs: a batch of scaled vectors of pixels from an image
:param ys: (optional) a batch of the class labels i.e.
the digit corresponding to the image(s)
:return: None
"""
# register this pytorch module and all of its sub-modules with pyro
pyro.module("ss_vae", self)
# inform Pyro that the variables in the batch of xs, ys are conditionally independent
batch_size = xs.size(0)
with pyro.iarange("independent"):
# sample the handwriting style from the constant prior distribution
prior_mu = Variable(torch.zeros([batch_size, self.z_dim]))
prior_sigma = Variable(torch.ones([batch_size, self.z_dim]))
zs = pyro.sample("z", dist.Normal(prior_mu, prior_sigma).reshape(extra_event_dims=1))
# if the label y (which digit to write) is supervised, sample from the
# constant prior, otherwise, observe the value (i.e. score it against the constant prior)
alpha_prior = Variable(torch.ones([batch_size, self.y_dim]) / (1.0 * self.y_dim))
if ys is None:
ys = pyro.sample("y", dist.OneHotCategorical(alpha_prior))
else:
pyro.sample("y", dist.OneHotCategorical(alpha_prior), obs=ys)
# finally, score the image (x) using the handwriting style (z) and
# the class label y (which digit to write) against the
# parametrized distribution p(x|y,z) = bernoulli(decoder(y,z))
# where `decoder` is a neural network
mu = self.decoder.forward(zs, ys)
pyro.sample("x", dist.Bernoulli(mu).reshape(extra_event_dims=1), obs=xs)
def model_classify(self, xs, ys=None):
"""
this model is used to add an auxiliary (supervised) loss as described in the
Kingma et al., "Semi-Supervised Learning with Deep Generative Models".
"""
# register all pytorch (sub)modules with pyro
pyro.module("ss_vae", self)
# inform Pyro that the variables in the batch of xs, ys are conditionally independent
with pyro.plate("data"):
# this here is the extra term to yield an auxiliary loss that we do gradient descent on
if ys is not None:
alpha = self.encoder_y.forward(xs)
with pyro.poutine.scale(scale=self.aux_loss_multiplier):
pyro.sample("y_aux", dist.OneHotCategorical(alpha), obs=ys)
def model_aux(self,
xs: torch.Tensor,
ys: Optional[torch.Tensor] = None,
**kwargs: float) -> None:
"""
Models an auxiliary (supervised) loss
"""
pyro.module("ss_vae", self)
with pyro.plate("data"):
# the extra term to yield an auxiliary loss
aux_loss_multiplier = kwargs.get("aux_loss_multiplier", 20)
if ys is not None:
alpha = self.encoder_y.forward(xs)
with pyro.poutine.scale(scale=aux_loss_multiplier):
pyro.sample("y_aux", dist.OneHotCategorical(alpha), obs=ys)
def test_auxsvi_trainer_cls(invariances):
data_dim = (5, 8, 8)
train_unsup = torch.randn(data_dim[0], torch.prod(tt(data_dim[1:])).item())
train_sup = train_unsup + .1 * torch.randn_like(train_unsup)
labels = dist.OneHotCategorical(torch.ones(data_dim[0], 3)).sample()
loader_unsup, loader_sup, loader_val = utils.init_ssvae_dataloaders(
train_unsup, (train_sup, labels), (train_sup, labels), batch_size=2)
vae = models.ssiVAE(data_dim[1:], 2, 3, invariances)
trainer = trainers.auxSVItrainer(vae)
weights_before = dc(vae.state_dict())
for _ in range(2):
trainer.step(loader_unsup, loader_sup, loader_val)
weights_after = vae.state_dict()
assert_(not torch.isnan(tt(trainer.history["training_loss"])).any())
assert_(not assert_weights_equal(weights_before, weights_after))
def model(self,
batch,
reversed_batch,
batch_mask,
batch_seqlens,
kl_anneal=1.0):
"""
the model defines p(x_{1:T}|z_{1:T}) and p(z_{1:T})
"""
# maximum duration of batch
Tmax = batch.size(1)
# register torch submodules w/ pyro
pyro.module("dmm", self)
# setup recursive conditioning for p(z_t|z_{t-1})
z_prev = self.z_0.expand(batch.size(0), self.z_0.size(0))
# sample conditionally indepdent text across the batch
with pyro.plate("z_batch", len(batch)):
# sample latent vars z and observed x w/ multiple samples from the guide for each z
for t in pyro.markov(range(1, Tmax + 1)):
# compute params of diagonal gaussian p(z_t|z_{t-1})
z_loc, z_scale = self.transition(z_prev)
# sample latent variable
with poutine.scale(scale=kl_anneal):
z_t = pyro.sample(
"z_%d" % t,
dist.Normal(z_loc, z_scale).mask(
batch_mask[:, t - 1:t]).to_event(1),
)
# compute emission probability from latent variable
emission_prob = self.emitter(z_t)
# observe x_t according to the Categorical distribution defined by the emitter probability
pyro.sample(
"obs_x_%d" % t,
dist.OneHotCategorical(emission_prob).mask(
batch_mask[:, t - 1:t]).to_event(1),
obs=batch[:, t - 1, :],
)
# set conditional var for next time step
z_prev = z_t
pass
