def __init__(self, source=None, **kwargs):
kwargs = super(SimTB, self).__init__(**kwargs)
if source is None:
raise ValueError('No source provided')
# Fetch simTB data from "source" source can be file, directory, etc.
self.X = self.get_data(source)
self.n = self.X.shape[0]
# Reference for the dimension of the dataset. A dict is used for
# multimodal data (e.g., mri and labels)
self.dims = dict()
self.dims[self.name] = self.X.shape[1]
# This is reference for models to decide how the data should be modelled
# E.g. with a binomial or gaussian variable
self.distributions = dict()
self.distributions[self.name] = 'gaussian'
# We will probably center the data in the main script using this
# global mean image.
self.mean_image = self.X.mean(axis=0)
warn_kwargs(self, kwargs)
def __init__(self, source=None, **kwargs):
kwargs = super(SimTB, self).__init__(**kwargs)
if source is None:
raise ValueError('No source provided')
# Fetch simTB data from "source" source can be file, directory, etc.
self.X = self.get_data(source)
self.n = self.X.shape[0]
# Reference for the dimension of the dataset. A dict is used for
# multimodal data (e.g., mri and labels)
self.dims = dict()
self.dims[self.name] = self.X.shape[1]
# This is reference for models to decide how the data should be modelled
# E.g. with a binomial or gaussian variable
self.distributions = dict()
self.distributions[self.name] = 'gaussian'
# We will probably center the data in the main script using this
# global mean image.
self.mean_image = self.X.mean(axis=0)
warn_kwargs(self, kwargs)
def __init__(self, name='', excludes=[], learn=True, **kwargs):
self.name = name
self.params = None
self.excludes = excludes
self.learn = learn
self.set_params()
self.n_params = len(self.params)
warn_kwargs(kwargs)
def __init__(self,
model,
name='IRVI',
inference_rate=0.1,
n_inference_samples=20,
n_inference_steps=20,
pass_gradients=True,
init_inference='recognition_network',
**kwargs):
self.name = name
self.model = model
self.init_inference = init_inference
self.inference_rate = inference_rate
self.n_inference_steps = n_inference_steps
self.n_inference_samples = n_inference_samples
self.pass_gradients = pass_gradients
warn_kwargs(self, **kwargs)
def __init__(self, model, name='RWS', **kwargs):
self.name = name
self.model = model
warn_kwargs(self, **kwargs)