skwdro package

Subpackages

Submodules

skwdro.torch module

skwdro.wrap_problem module

skwdro.wrap_problem.dualize_primal_loss(loss_: Module, transform_: Module | None, rho: Tensor, xi_batchinit: Tensor, xi_labels_batchinit: Tensor | None, post_sample: bool = True, cost_spec: str | None = None, n_samples: int = 10, seed: int = 42, *, epsilon: float | None = None, sigma: float | None = None, l2reg: float | None = None, adapt: str | None = 'prodigy', imp_samp: bool = True) → _DualLoss[source]

Provide the wrapped version of the primal loss.

Parameters:

loss_: nn.Module: the primal loss
transform_: nn.Module: the transformation to apply to the data before feeding it to the loss
rho: Tensor, shape (n_samples,): Wasserstein radius
xi_batchinit: Tensor, shape (n_samples, n_features): Data points to initialize the samplers and $\lambda_0$
xi_labels_batchinit: Optional[Tensor], shape (n_samples, n_features): Labels to initialize the samplers and $\lambda_0$
post_sample: bool: whether to use a post-sampled dual loss
cost_spec: str|None: the cost specification in the format (k, p) for a sample k-norm and p-power. None to use the default (2, 2).
n_samples: int: number of $\zeta$ samples to draw before the gradient descent begins (can be changed if needed between inferences)
seed: int: the seed for the samplers
epsilon: float|None: Epsilon if hard coded, None to let the algo find it.
sigma: float|None: Sigma if hard coded, None to let the algo find it.
l2reg: float|None: L2 regularization if needed
adapt: str|None: the adaptative step to use between “prodigy” and “mechanic”.
imp_samp: bool: whether to use importance sampling (will work only for (2, 2) costs).

skwdro.wrap_problem.expert_hyperparams(rho: Tensor, p: float, epsilon: float | None, epsilon_sigma_factor: float, sigma: float | None, sigma_factor: float) → Tuple[Tensor, Tensor][source]

Tuning of the hyperparameters for the dual loss.