def sample_and_log_prob(self, num_samples, context=None):
"""Generates samples from the distribution together with their log probability.
Args:
num_samples: int, number of samples to generate.
context: Tensor or None, conditioning variables. If None, the context is ignored.
Returns:
A tuple of:
* A Tensor containing the samples, with shape [num_samples, ...] if context is None,
or [context_size, num_samples, ...] if context is given.
* A Tensor containing the log probabilities of the samples, with shape
[num_samples, ...] if context is None, or [context_size, num_samples, ...] if
context is given.
"""
samples = self.sample(num_samples, context=context)
if context is not None:
# Merge the context dimension with sample dimension in order to call log_prob.
samples = various.merge_leading_dims(samples, num_dims=2)
context = various.repeat_rows(context, num_reps=num_samples)
assert samples.shape[0] == context.shape[0]
log_prob = self.log_prob(samples, context=context)
if context is not None:
# Split the context dimension from sample dimension.
samples = various.split_leading_dim(samples,
shape=[-1, num_samples])
log_prob = various.split_leading_dim(log_prob,
shape=[-1, num_samples])
return samples, log_prob
def _sample(self, num_samples, context):
if context is None:
return torch.randn(num_samples, *self._shape)
else:
# The value of the context is ignored, only its size is taken into account.
context_size = context.shape[0]
samples = torch.randn(context_size * num_samples, *self._shape)
return various.split_leading_dim(samples,
[context_size, num_samples])
def _sample(self, num_samples, context):
# Compute parameters.
logits = self._compute_params(context)
probs = torch.sigmoid(logits)
probs = various.repeat_rows(probs, num_samples)
# Generate samples.
context_size = context.shape[0]
noise = torch.rand(context_size * num_samples, *self._shape)
samples = (noise < probs).float()
return various.split_leading_dim(samples, [context_size, num_samples])
def _sample(self, num_samples, context):
# Compute parameters.
means, log_stds = self._compute_params(context)
stds = torch.exp(log_stds)
means = various.repeat_rows(means, num_samples)
stds = various.repeat_rows(stds, num_samples)
# Generate samples.
context_size = context.shape[0]
noise = torch.randn(context_size * num_samples, *self._shape)
samples = means + stds * noise
return various.split_leading_dim(samples, [context_size, num_samples])
def _sample(self, num_samples, context):
if context is None:
return self.std * torch.randn(num_samples, *self._shape)
else:
stds = torch.exp(self.log_stds).unsqueeze(0)
# The value of the context is ignored, only its size is taken into account.
context_size = context.shape[0]
samples = torch.randn(context_size * num_samples, *self._shape)
samples = various.split_leading_dim(samples,
[context_size, num_samples])
samples = stds * samples
return samples