def build_factor_graph_model(labels,
feats,
factor_types,
edge_table,
infer_alg=GRAPH_CUT):
""" Build factor graph model
Args:
labels: matrix of labels [num_train_samples*len_label]
feats: maxtrix of feats [num_train_samples*len_feat]
factory_types: vectors of all factor types
edge_table: matrix of pairwised edges, each row is a pair of node indeces
infer_alg: inference algorithm (GRAPH_CUT)
Returns:
labels_fg: matrix of labels in factor graph format
feats_fg: matrix of features in factor graph format
"""
labels = labels.astype(np.int32)
num_train_samples = labels.shape[0]
num_vars = labels.shape[1]
num_edges = edge_table.shape[0]
n_stats = 2
feats_fg = FactorGraphFeatures(num_train_samples)
labels_fg = FactorGraphLabels(num_train_samples)
for i in range(num_train_samples):
cardinaities = np.array([n_stats] * num_vars, np.int32)
fg = FactorGraph(cardinaities)
# add unary factors
for u in range(num_vars):
data_u = np.array(feats[i, :], np.float64)
inds_u = np.array([u], np.int32)
factor_u = Factor(factor_types[u], inds_u, data_u)
fg.add_factor(factor_u)
# add pairwise factors
for v in range(num_edges):
data_p = np.array([1.0])
inds_p = np.array(edge_table[v, :], np.int32)
factor_p = Factor(factor_types[v + num_vars], inds_p, data_p)
fg.add_factor(factor_p)
# add factor graph
feats_fg.add_sample(fg)
# add corresponding label
loss_weights = np.array([1.0 / num_vars] * num_vars)
fg_obs = FactorGraphObservation(labels[i, :], loss_weights)
labels_fg.add_label(fg_obs)
return (labels_fg, feats_fg)
def build_factor_graph_model(labels, feats, factor_types, edge_table, infer_alg = GRAPH_CUT):
""" Build factor graph model
Args:
labels: matrix of labels [num_train_samples*len_label]
feats: maxtrix of feats [num_train_samples*len_feat]
factory_types: vectors of all factor types
edge_table: matrix of pairwised edges, each row is a pair of node indeces
infer_alg: inference algorithm (GRAPH_CUT)
Returns:
labels_fg: matrix of labels in factor graph format
feats_fg: matrix of features in factor graph format
"""
labels = labels.astype(np.int32)
num_train_samples = labels.shape[0]
num_vars = labels.shape[1]
num_edges = edge_table.shape[0]
n_stats = 2
feats_fg = FactorGraphFeatures(num_train_samples)
labels_fg = FactorGraphLabels(num_train_samples)
for i in range(num_train_samples):
cardinaities = np.array([n_stats]*num_vars, np.int32)
fg = FactorGraph(cardinaities)
# add unary factors
for u in range(num_vars):
data_u = np.array(feats[i,:], np.float64)
inds_u = np.array([u], np.int32)
factor_u = Factor(factor_types[u], inds_u, data_u)
fg.add_factor(factor_u)
# add pairwise factors
for v in range(num_edges):
data_p = np.array([1.0])
inds_p = np.array(edge_table[v, :], np.int32)
factor_p = Factor(factor_types[v + num_vars], inds_p, data_p)
fg.add_factor(factor_p)
# add factor graph
feats_fg.add_sample(fg)
# add corresponding label
loss_weights = np.array([1.0/num_vars]*num_vars)
fg_obs = FactorGraphObservation(labels[i,:], loss_weights)
labels_fg.add_label(fg_obs)
return (labels_fg, feats_fg)
def gen_data(ftype, num_samples, show_data = False): from modshogun import Math from modshogun import FactorType, Factor, TableFactorType, FactorGraph from modshogun import FactorGraphObservation, FactorGraphLabels, FactorGraphFeatures from modshogun import MAPInference, TREE_MAX_PROD Math.init_random(17) samples = FactorGraphFeatures(num_samples) labels = FactorGraphLabels(num_samples) for i in xrange(num_samples): vc = np.array([2,2,2], np.int32) fg = FactorGraph(vc) data1 = np.array([2.0*Math.random(0.0,1.0)-1.0 for i in xrange(2)]) vind1 = np.array([0,1], np.int32) fac1 = Factor(ftype[0], vind1, data1) fg.add_factor(fac1) data2 = np.array([2.0*Math.random(0.0,1.0)-1.0 for i in xrange(2)]) vind2 = np.array([1,2], np.int32) fac2 = Factor(ftype[0], vind2, data2) fg.add_factor(fac2) data3 = np.array([2.0*Math.random(0.0,1.0)-1.0 for i in xrange(2)]) vind3 = np.array([0], np.int32) fac3 = Factor(ftype[1], vind3, data3) fg.add_factor(fac3) data4 = np.array([2.0*Math.random(0.0,1.0)-1.0 for i in xrange(2)]) vind4 = np.array([1], np.int32) fac4 = Factor(ftype[1], vind4, data4) fg.add_factor(fac4) data5 = np.array([2.0*Math.random(0.0,1.0)-1.0 for i in xrange(2)]) vind5 = np.array([2], np.int32) fac5 = Factor(ftype[1], vind5, data5) fg.add_factor(fac5) data6 = np.array([1.0]) vind6 = np.array([0], np.int32) fac6 = Factor(ftype[2], vind6, data6) fg.add_factor(fac6) data7 = np.array([1.0]) vind7 = np.array([2], np.int32) fac7 = Factor(ftype[2], vind7, data7) fg.add_factor(fac7) samples.add_sample(fg) fg.connect_components() fg.compute_energies() infer_met = MAPInference(fg, TREE_MAX_PROD) infer_met.inference() fg_obs = infer_met.get_structured_outputs() labels.add_label(fg_obs) if show_data: state = fg_obs.get_data() print state return samples, labels
def gen_data(ftype, num_samples, show_data=False):
from modshogun import Math
from modshogun import FactorType, Factor, TableFactorType, FactorGraph
from modshogun import FactorGraphObservation, FactorGraphLabels, FactorGraphFeatures
from modshogun import MAPInference, TREE_MAX_PROD
Math.init_random(17)
samples = FactorGraphFeatures(num_samples)
labels = FactorGraphLabels(num_samples)
for i in range(num_samples):
vc = np.array([2, 2, 2], np.int32)
fg = FactorGraph(vc)
data1 = np.array([2.0 * Math.random(0.0, 1.0) - 1.0 for i in range(2)])
vind1 = np.array([0, 1], np.int32)
fac1 = Factor(ftype[0], vind1, data1)
fg.add_factor(fac1)
data2 = np.array([2.0 * Math.random(0.0, 1.0) - 1.0 for i in range(2)])
vind2 = np.array([1, 2], np.int32)
fac2 = Factor(ftype[0], vind2, data2)
fg.add_factor(fac2)
data3 = np.array([2.0 * Math.random(0.0, 1.0) - 1.0 for i in range(2)])
vind3 = np.array([0], np.int32)
fac3 = Factor(ftype[1], vind3, data3)
fg.add_factor(fac3)
data4 = np.array([2.0 * Math.random(0.0, 1.0) - 1.0 for i in range(2)])
vind4 = np.array([1], np.int32)
fac4 = Factor(ftype[1], vind4, data4)
fg.add_factor(fac4)
data5 = np.array([2.0 * Math.random(0.0, 1.0) - 1.0 for i in range(2)])
vind5 = np.array([2], np.int32)
fac5 = Factor(ftype[1], vind5, data5)
fg.add_factor(fac5)
data6 = np.array([1.0])
vind6 = np.array([0], np.int32)
fac6 = Factor(ftype[2], vind6, data6)
fg.add_factor(fac6)
data7 = np.array([1.0])
vind7 = np.array([2], np.int32)
fac7 = Factor(ftype[2], vind7, data7)
fg.add_factor(fac7)
samples.add_sample(fg)
fg.connect_components()
fg.compute_energies()
infer_met = MAPInference(fg, TREE_MAX_PROD)
infer_met.inference()
fg_obs = infer_met.get_structured_outputs()
labels.add_label(fg_obs)
if show_data:
state = fg_obs.get_data()
print(state)
return samples, labels
