def __init__(self):
self.net = pf.Sequential([
pf.Dense(5, 128),
tf.nn.relu,
pf.Dense(128, 64),
tf.nn.relu,
pf.Dense(64, 1),
])
self.s = pf.ScaleParameter()
def __init__(self, d, k):
self.m = pf.Parameter([d, k])
self.s = pf.ScaleParameter([d, k])
self.w = pf.DirichletParameter(k)
def __init__(self):
self.mu = pf.Parameter(name='mu')
self.sig = pf.ScaleParameter(name='sig')
def __init__(self, dims):
self.net = DenseNetwork(dims)
self.s = pf.ScaleParameter([1, 1])
def __init__(self):
self.mu = pf.Parameter(name="mu")
self.sig = pf.ScaleParameter(name="sig")
def __init__(self):
self.weight = pf.Parameter(name="Weight")
self.bias = pf.Parameter(name="Bias")
self.std = pf.ScaleParameter(name="Noise Std Dev",
prior=pf.Gamma(1.0, 1.0))
def __init__(self):
self.weight = pf.Parameter(name="Weight")
self.bias = pf.Parameter(name="Bias")
self.std = pf.ScaleParameter(name="Noise Std Dev")
def __init__(self):
self.weight = pf.Parameter(name='Weight')
self.bias = pf.Parameter(name='Bias')
self.std = pf.ScaleParameter(name='Noise Std Dev')
def __init__(self, d, q):
self.W = pf.Parameter(shape=[d, q])
self.sigma = pf.ScaleParameter()
def __init__(self, dims):
self.w = pf.Parameter([dims, 1])
self.b = pf.Parameter()
self.s = pf.ScaleParameter()
def __init__(self):
self.w = pf.Parameter()
self.b = pf.Parameter()
self.s = pf.ScaleParameter()
def __init__(self):
self.weight = pf.Parameter(name='Weight')
self.bias = pf.Parameter(name='Bias')
self.std = pf.ScaleParameter(name='Noise Std Dev',
prior=pf.Gamma(1., 1.))
