VI node sampling with multiple parameters

ferrine · December 26, 2019, 3:05pm

Hi, you should pass same argument in pm.fit it will be bypassed to this:

pymc-devs/pymc3/blob/1b49e5eb533e6c05568fbd4592af658a98e15cc6/pymc3/variational/opvi.py#L220


    else:
        if tf_n_mc is not None:
            _warn_not_used('tf_n_mc', self.op)
        if more_tf_params:
            _warn_not_used('more_tf_params', self.op)
    self.add_obj_updates(
        resulting_updates,
        obj_n_mc=obj_n_mc,
        obj_optimizer=obj_optimizer,
        more_obj_params=more_obj_params,
        more_replacements=more_replacements,
        total_grad_norm_constraint=total_grad_norm_constraint
    )
    resulting_updates.update(more_updates)
    return resulting_updates


def add_test_updates(self, updates, tf_n_mc=None, test_optimizer=adagrad_window,
                     more_tf_params=None, more_replacements=None,
                     total_grad_norm_constraint=None):
    if more_tf_params is None:
        more_tf_params = []

Topic		Replies	Views
Combining parameters into single tensor Questions	19	3538	October 30, 2017
Weird thing about approximation.sample_node Questions	6	597	August 14, 2018
Sampling after minibatch training Questions	1	1049	November 28, 2017
`sample_node` no longer working? Questions	10	763	September 11, 2017
Input arguments in theano likelihood function that are not random variables Questions	0	385	May 27, 2020

VI node sampling with multiple parameters

Related topics