treeflow.vi.fixed_topology_advi module

treeflow.vi.fixed_topology_advi.fit_fixed_topology_variational_approximation(model: ~tensorflow_probability.python.distributions.distribution.Distribution, topologies: ~typing.Dict[str, ~treeflow.tree.topology.tensorflow_tree_topology.TensorflowTreeTopology], optimizer: ~keras.src.optimizers.optimizer.Optimizer, num_steps: int, trace_fn: ~typing.Callable[[~tensorflow_probability.python.math.minimize.MinimizeTraceableQuantities], object] | None = None, convergence_criterion: ~tensorflow_probability.python.optimizer.convergence_criteria.convergence_criterion.ConvergenceCriterion | None = None, init_loc: object | None = None, return_full_length_trace: bool = True, progress_bar: bool | ~treeflow.vi.progress_bar.ProgressBarFunc = False, progress_bar_step: int = 10, approx_fn: ~treeflow.vi.fixed_topology_advi.ApproximationBuilder = <function get_fixed_topology_mean_field_approximation>, approx_kwargs: ~typing.Dict[str, object] | None = None, **vi_kwargs) → Tuple[Distribution, object]