def inverse(self, inputs, context=None):
if torch.min(inputs) <= -1 or torch.max(inputs) >= 1:
raise transforms.InputOutsideDomain()
outputs = 0.5 * torch.log((1 + inputs) / (1 - inputs))
logabsdet = - torch.log(1 - inputs ** 2)
logabsdet = utils.sum_except_batch(logabsdet, num_batch_dims=1)
return outputs, logabsdet
def _spline(self, inputs, inverse=False):
batch_size = inputs.shape[0]
unnormalized_widths=_share_across_batch(self.unnormalized_widths, batch_size)
unnormalized_heights=_share_across_batch(self.unnormalized_heights, batch_size)
unnormalized_derivatives=_share_across_batch(self.unnormalized_derivatives, batch_size)
if self.tails is None:
spline_fn = splines.rational_quadratic_spline
spline_kwargs = {}
else:
spline_fn = splines.unconstrained_rational_quadratic_spline
spline_kwargs = {
'tails': self.tails,
'tail_bound': self.tail_bound
}
outputs, logabsdet = spline_fn(
inputs=inputs,
unnormalized_widths=unnormalized_widths,
unnormalized_heights=unnormalized_heights,
unnormalized_derivatives=unnormalized_derivatives,
inverse=inverse,
min_bin_width=self.min_bin_width,
min_bin_height=self.min_bin_height,
min_derivative=self.min_derivative,
**spline_kwargs
)
return outputs, utils.sum_except_batch(logabsdet)
def _elementwise_inverse(self, inputs, autoregressive_params):
unconstrained_scale, shift = self._unconstrained_scale_and_shift(autoregressive_params)
scale = torch.sigmoid(unconstrained_scale + 2.) + 1e-3
log_scale = torch.log(scale)
outputs = (inputs - shift) / scale
logabsdet = -utils.sum_except_batch(log_scale, num_batch_dims=1)
return outputs, logabsdet
def _log_prob(self, inputs, context):
# Note: the context is ignored.
if inputs.shape[1:] != self._shape:
raise ValueError('Expected input of shape {}, got {}'.format(
self._shape, inputs.shape[1:]))
neg_energy = -0.5 * utils.sum_except_batch(inputs**2, num_batch_dims=1)
return neg_energy - self._log_z
def _elementwise(self, inputs, autoregressive_params, inverse=False):
batch_size = inputs.shape[0]
transform_params = autoregressive_params.view(batch_size,
self.features,
self.num_bins * 2 + 2)
unnormalized_widths = transform_params[...,:self.num_bins]
unnormalized_heights = transform_params[...,self.num_bins:2*self.num_bins]
derivatives = transform_params[...,2*self.num_bins:]
unnorm_derivatives_left = derivatives[..., 0][..., None]
unnorm_derivatives_right = derivatives[..., 1][..., None]
if hasattr(self.autoregressive_net, 'hidden_features'):
unnormalized_widths /= np.sqrt(self.autoregressive_net.hidden_features)
unnormalized_heights /= np.sqrt(self.autoregressive_net.hidden_features)
outputs, logabsdet = splines.cubic_spline(
inputs=inputs,
unnormalized_widths=unnormalized_widths,
unnormalized_heights=unnormalized_heights,
unnorm_derivatives_left=unnorm_derivatives_left,
unnorm_derivatives_right=unnorm_derivatives_right,
inverse=inverse
)
return outputs, utils.sum_except_batch(logabsdet)
def inverse(self, inputs, context=None):
outputs = F.leaky_relu(inputs,
negative_slope=(1 / self.negative_slope))
mask = (inputs < 0).type(torch.Tensor)
logabsdet = -self.log_negative_slope * mask
logabsdet = utils.sum_except_batch(logabsdet, num_batch_dims=1)
return outputs, logabsdet
def _log_prob(self, inputs, context):
if inputs.shape[1:] != self._shape:
raise ValueError('Expected input of shape {}, got {}'.format(
self._shape, inputs.shape[1:]))
# Compute parameters.
means, log_stds = self._compute_params(context)
assert means.shape == inputs.shape and log_stds.shape == inputs.shape
# Compute log prob.
norm_inputs = (inputs - means) * torch.exp(-log_stds)
log_prob = -0.5 * utils.sum_except_batch(norm_inputs**2,
num_batch_dims=1)
log_prob -= utils.sum_except_batch(log_stds, num_batch_dims=1)
log_prob -= self._log_z
return log_prob
def forward(self, inputs, context=None):
inputs = self.temperature * inputs
outputs = torch.sigmoid(inputs)
logabsdet = utils.sum_except_batch(
torch.log(self.temperature) - F.softplus(-inputs) - F.softplus(inputs)
)
return outputs, logabsdet
def inverse(self, inputs, context=None):
if torch.min(inputs) < 0 or torch.max(inputs) > 1:
raise transforms.InputOutsideDomain()
outputs = torch.tan(np.pi * (inputs - 0.5))
logabsdet = -utils.sum_except_batch(-np.log(np.pi) -
torch.log(1 + outputs**2))
return outputs, logabsdet
def _elementwise_forward(self, inputs, autoregressive_params):
unconstrained_scale, shift = self._unconstrained_scale_and_shift(
autoregressive_params)
# scale = torch.sigmoid(unconstrained_scale + 2.0) + self._epsilon
scale = F.softplus(unconstrained_scale) + self._epsilon
log_scale = torch.log(scale)
outputs = scale * inputs + shift
logabsdet = utils.sum_except_batch(log_scale, num_batch_dims=1)
return outputs, logabsdet
def _elementwise(self, inputs, autoregressive_params, inverse=False):
batch_size = inputs.shape[0]
unnormalized_pdf = autoregressive_params.view(
batch_size, self.features, self._output_dim_multiplier())
outputs, logabsdet = splines.linear_spline(
inputs=inputs, unnormalized_pdf=unnormalized_pdf, inverse=inverse)
return outputs, utils.sum_except_batch(logabsdet)
def inverse(self, inputs, context=None):
if torch.min(inputs) < 0 or torch.max(inputs) > 1:
raise transforms.InputOutsideDomain()
inputs = torch.clamp(inputs, self.eps, 1 - self.eps)
outputs = (1 / self.temperature) * (torch.log(inputs) - torch.log1p(-inputs))
logabsdet = - utils.sum_except_batch(
torch.log(self.temperature) - F.softplus(
-self.temperature * outputs) - F.softplus(self.temperature * outputs)
)
return outputs, logabsdet
def _lu_forward_inverse(self, inputs, inverse=False):
b, c, h, w = inputs.shape
inputs = inputs.permute(0, 2, 3, 1).reshape(b * h * w, c)
if inverse:
outputs, logabsdet = super().inverse(inputs)
else:
outputs, logabsdet = super().forward(inputs)
outputs = outputs.reshape(b, h, w, c).permute(0, 3, 1, 2)
logabsdet = logabsdet.reshape(b, h, w)
return outputs, utils.sum_except_batch(logabsdet)
def _coupling_transform(self, inputs, transform_params, inverse=False):
if inputs.dim() == 4:
b, c, h, w = inputs.shape
# For images, reshape transform_params from Bx(C*?)xHxW to BxCxHxWx?
transform_params = transform_params.reshape(b, c, -1, h, w).permute(0, 1, 3, 4, 2)
elif inputs.dim() == 2:
b, d = inputs.shape
# For 2D data, reshape transform_params from Bx(D*?) to BxDx?
transform_params = transform_params.reshape(b, d, -1)
outputs, logabsdet = self._piecewise_cdf(inputs, transform_params, inverse)
return outputs, utils.sum_except_batch(logabsdet)
def _log_prob(self, inputs, context):
if inputs.shape[1:] != self._shape:
raise ValueError('Expected input of shape {}, got {}'.format(
self._shape, inputs.shape[1:]))
# Compute parameters.
logits = self._compute_params(context)
assert logits.shape == inputs.shape
# Compute log prob.
log_prob = -inputs * F.softplus(-logits) - (
1.0 - inputs) * F.softplus(logits)
log_prob = utils.sum_except_batch(log_prob, num_batch_dims=1)
return log_prob
def forward(self, inputs, context=None):
mask_right = (inputs > self.cut_point)
mask_left = (inputs < -self.cut_point)
mask_middle = ~(mask_right | mask_left)
outputs = torch.zeros_like(inputs)
outputs[mask_middle] = torch.tanh(inputs[mask_middle])
outputs[mask_right] = self.alpha * torch.log(self.beta * inputs[mask_right])
outputs[mask_left] = self.alpha * -torch.log(-self.beta * inputs[mask_left])
logabsdet = torch.zeros_like(inputs)
logabsdet[mask_middle] = torch.log(1 - outputs[mask_middle] ** 2)
logabsdet[mask_right] = torch.log(self.alpha / inputs[mask_right])
logabsdet[mask_left] = torch.log(-self.alpha / inputs[mask_left])
logabsdet = utils.sum_except_batch(logabsdet, num_batch_dims=1)
return outputs, logabsdet
def inverse(self, inputs, context=None):
mask_right = (inputs > self.inv_cut_point)
mask_left = (inputs < -self.inv_cut_point)
mask_middle = ~(mask_right | mask_left)
outputs = torch.zeros_like(inputs)
outputs[mask_middle] = 0.5 * torch.log((1 + inputs[mask_middle])
/ (1 - inputs[mask_middle]))
outputs[mask_right] = torch.exp(inputs[mask_right] / self.alpha) / self.beta
outputs[mask_left] = -torch.exp(-inputs[mask_left] / self.alpha) / self.beta
logabsdet = torch.zeros_like(inputs)
logabsdet[mask_middle] = -torch.log(1 - inputs[mask_middle] ** 2)
logabsdet[mask_right] = -np.log(self.alpha * self.beta) + inputs[mask_right] / self.alpha
logabsdet[mask_left] = -np.log(self.alpha * self.beta) - inputs[mask_left] / self.alpha
logabsdet = utils.sum_except_batch(logabsdet, num_batch_dims=1)
return outputs, logabsdet
def _elementwise(self, inputs, autoregressive_params, inverse=False):
batch_size, features = inputs.shape[0], inputs.shape[1]
transform_params = autoregressive_params.view(
batch_size,
features,
self._output_dim_multiplier()
)
unnormalized_widths = transform_params[...,:self.num_bins]
unnormalized_heights = transform_params[...,self.num_bins:2*self.num_bins]
unnormalized_derivatives = transform_params[...,2*self.num_bins:]
if hasattr(self.autoregressive_net, 'hidden_features'):
unnormalized_widths /= np.sqrt(self.autoregressive_net.hidden_features)
unnormalized_heights /= np.sqrt(self.autoregressive_net.hidden_features)
if self.tails is None:
spline_fn = splines.rational_quadratic_spline
spline_kwargs = {}
elif self.tails == 'linear':
spline_fn = splines.unconstrained_rational_quadratic_spline
spline_kwargs = {
'tails': self.tails,
'tail_bound': self.tail_bound
}
else:
raise ValueError
outputs, logabsdet = spline_fn(
inputs=inputs,
unnormalized_widths=unnormalized_widths,
unnormalized_heights=unnormalized_heights,
unnormalized_derivatives=unnormalized_derivatives,
inverse=inverse,
min_bin_width=self.min_bin_width,
min_bin_height=self.min_bin_height,
min_derivative=self.min_derivative,
**spline_kwargs
)
return outputs, utils.sum_except_batch(logabsdet)
def _spline(self, inputs, inverse=False):
batch_size = inputs.shape[0]
unnormalized_pdf = _share_across_batch(self.unnormalized_pdf,
batch_size)
if self.tails is None:
outputs, logabsdet = splines.linear_spline(
inputs=inputs,
unnormalized_pdf=unnormalized_pdf,
inverse=inverse)
else:
outputs, logabsdet = splines.unconstrained_linear_spline(
inputs=inputs,
unnormalized_pdf=unnormalized_pdf,
inverse=inverse,
tails=self.tails,
tail_bound=self.tail_bound)
return outputs, utils.sum_except_batch(logabsdet)
def forward(self, x, context=None):
x2_size = self.input_dim - self.split_dim
x1, x2 = x.split([self.split_dim, x2_size], dim=self.event_dim)
nn_input = torch.cat(
(x1, context), dim=self.event_dim) if self.context_dim != 0 else x1
nn_out = torch.utils.checkpoint.checkpoint(
self.nn, nn_input, preserve_rng_state=False)
unnormalized_widths, unnormalized_heights, unnormalized_derivatives = nn_out.reshape(
nn_input.shape[:2]+(-1, self._output_dim_multiplier())).split([self.num_bins, self.num_bins, self.num_bins+1], dim=self.event_dim)
# Inverse not specified as default is false
y2, ldj = torch.utils.checkpoint.checkpoint(unconstrained_rational_quadratic_spline,
x2,
unnormalized_widths,
unnormalized_heights,
unnormalized_derivatives,
preserve_rng_state=False)
ldj = sum_except_batch(ldj, num_dims=2)
y1 = x1
return torch.cat([y1, y2], dim=self.event_dim), ldj
def log_prob(self, x, context):
dist = self.cond_dist(context)
return sum_except_batch(dist.log_prob(x), num_dims=2)
def _coupling_transform_inverse(self, inputs, transform_params):
scale, shift = self._scale_and_shift(transform_params)
log_scale = torch.log(scale)
outputs = (inputs - shift) / scale
logabsdet = -utils.sum_except_batch(log_scale, num_batch_dims=1)
return outputs, logabsdet
def _coupling_transform_forward(self, inputs, transform_params):
scale, shift = self._scale_and_shift(transform_params)
log_scale = torch.log(scale)
outputs = inputs * scale + shift
logabsdet = utils.sum_except_batch(log_scale, num_batch_dims=1)
return outputs, logabsdet
def evaluate_test_samples(args, simulator, filename, model=None, ood=False, n_save_reco=100):
""" Likelihood evaluation """
logger.info(
"Evaluating %s samples according to %s, %s likelihood evaluation, saving in %s",
"the ground truth" if model is None else "a trained model",
"ood" if ood else "test",
"with" if not args.skiplikelihood else "without",
filename,
)
# Prepare
x, _ = simulator.load_dataset(
train=False, numpy=True, ood=ood, dataset_dir=create_filename("dataset", None, args), true_param_id=args.trueparam, joint_score=False, limit_samplesize=args.evaluate,
)
parameter_grid = [None] if simulator.parameter_dim() is None else simulator.eval_parameter_grid(resolution=args.gridresolution)
log_probs = []
x_recos = None
reco_error = None
# Evaluate
for i, params in enumerate(parameter_grid):
logger.debug("Evaluating grid point %s / %s", i + 1, len(parameter_grid))
if model is None:
params_ = None if params is None else np.asarray([params for _ in x])
log_prob = simulator.log_density(x, parameters=params_)
else:
log_prob = []
reco_error_ = []
x_recos_ = []
n_batches = (args.evaluate - 1) // args.evalbatchsize + 1
for j in range(n_batches):
x_ = torch.tensor(x[j * args.evalbatchsize : (j + 1) * args.evalbatchsize], dtype=torch.float)
if params is None:
params_ = None
else:
params_ = np.asarray([params for _ in x_])
params_ = torch.tensor(params_, dtype=torch.float)
if args.algorithm == "flow":
x_reco, log_prob_, _ = model(x_, context=params_)
elif args.algorithm in ["pie", "slice"]:
x_reco, log_prob_, _ = model(x_, context=params_, mode=args.algorithm if not args.skiplikelihood else "projection")
else:
x_reco, log_prob_, _ = model(x_, context=params_, mode="mf" if not args.skiplikelihood else "projection")
if not args.skiplikelihood:
log_prob.append(log_prob_.detach().numpy())
reco_error_.append((sum_except_batch((x_ - x_reco) ** 2) ** 0.5).detach().numpy())
x_recos_.append(x_reco.detach().numpy())
if not args.skiplikelihood:
log_prob = np.concatenate(log_prob, axis=0)
if reco_error is None:
reco_error = np.concatenate(reco_error_, axis=0)
if x_recos is None:
x_recos = np.concatenate(x_recos_, axis=0)
if not args.skiplikelihood:
log_probs.append(log_prob)
# Save results
if len(log_probs) > 0:
if simulator.parameter_dim() is None:
log_probs = log_probs[0]
np.save(create_filename("results", filename.format("log_likelihood"), args), log_probs)
if len(x_recos) > 0:
np.save(create_filename("results", filename.format("x_reco"), args), x_recos[:n_save_reco])
if reco_error is not None:
np.save(create_filename("results", filename.format("reco_error"), args), reco_error)
if parameter_grid is not None:
np.save(create_filename("results", "parameter_grid_test", args), parameter_grid)
def log_prob(self, value, context):
return utils.sum_except_batch(super().log_prob(value))
def log_prob(self, x, context=None):
log_base = -0.5 * math.log(2 * math.pi)
log_inner = -0.5 * x**2
return sum_except_batch(log_base + log_inner, num_dims=2)
def sample_with_log_prob(self, context):
dist = self.cond_dist(context)
z = dist.rsample()
log_prob = dist.log_prob(z)
log_prob = sum_except_batch(log_prob, num_dims=2)
return z, log_prob
def forward(self, inputs, context=None):
outputs = (1 / np.pi) * torch.atan(inputs) + 0.5
logabsdet = utils.sum_except_batch(
- np.log(np.pi) - torch.log(1 + inputs ** 2)
)
return outputs, logabsdet
def forward(self, inputs, context=None):
outputs = torch.tanh(inputs)
logabsdet = torch.log(1 - outputs ** 2)
logabsdet = utils.sum_except_batch(logabsdet, num_batch_dims=1)
return outputs, logabsdet