def test_06(self):
lmbda = 1e-1
W = np.array([1.0])
opt = parcnsdl.ConvBPDNMaskDcplDictLearn_Consensus.Options(
{'MaxMainIter': 10})
try:
b = parcnsdl.ConvBPDNMaskDcplDictLearn_Consensus(self.D0,
self.S[...,0], lmbda, W, opt=opt, nproc=2, dimK=0)
b.solve()
except Exception as e:
print(e)
assert(0)
def test_12(self):
lmbda = 1e-1
W = np.ones(self.S.shape[0:2] + (1, 1, 1))
opt = parcnsdl.ConvBPDNMaskDcplDictLearn_Consensus.Options(
{'MaxMainIter': 10})
try:
b = parcnsdl.ConvBPDNMaskDcplDictLearn_Consensus(self.D0,
self.S, lmbda, W, opt=opt, nproc=2)
b.solve()
except Exception as e:
print(e)
assert(0)
def test_18(self):
lmbda = 1e-1
W = np.ones(self.S.shape[0:2] + (1, self.S.shape[2], 1))
Nit = 10
opts = cbpdndl.ConvBPDNMaskDcplDictLearn.Options({
'MaxMainIter': Nit,
'AccurateDFid': True,
'CBPDN': {
'RelaxParam': 1.8,
'AutoRho': {
'Enabled': False
}
},
'CCMOD': {
'RelaxParam': 1.8,
'AutoRho': {
'Enabled': False
}
}
})
bs = cbpdndl.ConvBPDNMaskDcplDictLearn(self.D0,
self.S,
lmbda,
W,
opt=opts,
method='cns')
Ds = bs.solve()
optp = parcnsdl.ConvBPDNDictLearn_Consensus.Options({
'MaxMainIter': Nit,
'CBPDN': {
'RelaxParam': 1.8
},
'CCMOD': {
'RelaxParam': 1.8
}
})
bp = parcnsdl.ConvBPDNMaskDcplDictLearn_Consensus(self.D0,
self.S,
lmbda,
W,
opt=optp,
nproc=2)
Dp = bp.solve()
assert (np.linalg.norm(Ds - Dp) < 1e-7)
assert (np.abs(bs.getitstat().ObjFun[-1] - bp.getitstat().ObjFun[-1]) <
1e-7)
def test_14(self):
N = 16
Nc = 3
Nd = 5
M = 4
K = 2
D0 = np.random.randn(Nd, Nd, Nc, M)
S = np.random.randn(N, N, Nc, K)
lmbda = 1e-1
W = np.ones(S.shape + (1,))
opt = parcnsdl.ConvBPDNMaskDcplDictLearn_Consensus.Options(
{'MaxMainIter': 10})
try:
b = parcnsdl.ConvBPDNMaskDcplDictLearn_Consensus(D0,
S, lmbda, W, opt=opt, nproc=2)
b.solve()
except Exception as e:
print(e)
assert(0)