treeflow.debug.minimize_eager module

treeflow.debug.minimize_eager.minimize_eager(loss_fn, num_steps, optimizer, trainable_variables=None, iter=<class 'range'>, convergence_criterion=None, trace_fn=<function <lambda>>, batch_convergence_reduce_fn=<function all>)

treeflow.debug.minimize_eager.fit_surrogate_posterior(target_log_prob_fn, surrogate_posterior, optimizer, num_steps, variational_loss_fn=None, discrepancy_fn=<function kl_reverse>, sample_size=1, importance_sample_size=1, seed=None, **opt_kwargs)