[opengm] 131/386: bla

[Ghislain Vaillant]

This is an automated email from the git hooks/post-receive script.

ghisvail-guest pushed a commit to branch debian/master
in repository opengm.

commit 30448b3a6564150534feae62e258b7c26d5cdd56
Author: DerThorsten <thorsten.beier at iwr.uni-heidelberg.de>
Date:   Thu Dec 18 16:25:52 2014 +0100

    bla
---
 fubar/example1.py     |   4 +-
 fubar/real_example.py | 105 ++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 106 insertions(+), 3 deletions(-)

diff --git a/fubar/example1.py b/fubar/example1.py
index e5e49e7..9f5f845 100644
--- a/fubar/example1.py
+++ b/fubar/example1.py
@@ -16,8 +16,6 @@ dataset =learning.createDataset(loss='h')
 print "type of dataset", dataset
 
 
-
-
 # for grid search learner
 lowerBounds = numpy.zeros(nWeights)
 upperBounds = numpy.ones(nWeights)
@@ -34,4 +32,4 @@ smm_learnerParam = learning.StructMaxMargin_Bundle_HammingLossParameter(1.0, 0.0
 smm_learner = learning.StructMaxMargin_Bundle_HammingLoss(dataset, smm_learnerParam)
 smm_learner.learn(infCls=opengm.inference.Icm)
 smm_learner2 = learning.structMaxMarginLearner(dataset, 1.0, 0.001, 0)
-smm_learner2.learn(infCls=opengm.inference.BeliefPropagation, parameter=opengm.InfParam(damping=0.5))
\ No newline at end of file
+smm_learner2.learn(infCls=opengm.inference.BeliefPropagation, parameter=opengm.InfParam(damping=0.5))
diff --git a/fubar/real_example.py b/fubar/real_example.py
new file mode 100644
index 0000000..bf3d9c3
--- /dev/null
+++ b/fubar/real_example.py
@@ -0,0 +1,105 @@
+import opengm
+import opengm.learning as learning
+from opengm import numpy
+import vigra
+
+nModels = 4
+nLables = 2 
+shape = [10, 10]
+numVar = shape[0]*shape[1]
+nWeights = 12
+
+def makeGt(shape):
+    gt=numpy.ones(shape,dtype='uint8')
+    gt[0:shape[0]/2,:] = 0
+    return gt
+
+
+
+weightVals = numpy.ones(nWeights)
+weights = opengm.learning.Weights(weightVals)
+
+uWeightIds = numpy.arange(8,dtype='uint64').reshape(2,4)
+
+
+print uWeightIds
+
+bWeightIds = numpy.array([8,9,10,11],dtype='uint64')
+
+
+dataset = learning.createDataset(loss='h')
+
+
+
+def makeFeatures(gt):
+    random  = numpy.random.rand(*gt.shape)-0.5
+    randGt = random + gt
+    feat = []
+    for sigma in [1.0, 1.5, 2.0]:
+        feat.append(vigra.filters.gaussianSmoothing(randGt.astype('float32'),sigma) )
+
+    featB = []
+    for sigma in [1.0, 1.5, 2.0]:
+        featB.append(vigra.filters.gaussianGradientMagnitude(randGt.astype('float32'),sigma) )
+
+
+
+    a =  numpy.rollaxis(numpy.array(feat), axis=0, start=3)
+    b =  numpy.rollaxis(numpy.array(featB), axis=0, start=3)
+    return a,b
+
+for mi in range(nModels):
+
+
+    gm = opengm.gm(numpy.ones(numVar)*nLables)
+    gt = makeGt(shape) 
+    gtFlat = gt.reshape([-1])
+
+    unaries,binaries = makeFeatures(gt)
+
+    print unaries, binaries
+
+
+    for x in range(shape[0]):
+        for y in range(shape[1]):
+            uFeat = numpy.append(unaries[x,y,:], [1]).astype(opengm.value_type)
+            uFeat = numpy.repeat(uFeat[:,numpy.newaxis],2,axis=1).T
+
+            lu = opengm.LUnaryFunction(weights=weights,numberOfLabels=nLables, features=uFeat, weightIds=uWeightIds)
+
+
+            fid= gm.addFunction(lu)
+            gm.addFactor(fid, y+x*shape[1])
+
+
+
+    for x in range(shape[0]):
+        for y in range(shape[1]):
+
+            if x+1 < shape[0]:
+                bFeat = numpy.append(binaries[x,y,:], [1]).astype(opengm.value_type) +  numpy.append(binaries[x+1,y,:], [1]).astype(opengm.value_type)
+                pf = opengm.LPottsFunction(weights=weights,numberOfLabels=nLables, features=bFeat, weightIds=bWeightIds)
+                fid= gm.addFunction(pf)
+                gm.addFactor(fid, [y+x*shape[1], y+(x+1)*shape[1]])
+            if y+1 < shape[1]:
+                bFeat = numpy.append(binaries[x,y,:], [1]).astype(opengm.value_type) + numpy.append(binaries[x,y+1,:], [1]).astype(opengm.value_type)
+                pf = opengm.LPottsFunction(weights=weights,numberOfLabels=nLables, features=bFeat, weightIds=bWeightIds)
+                fid= gm.addFunction(pf)
+                gm.addFactor(fid, [y+x*shape[1], y+1+x*shape[1]])
+
+
+
+    dataset.pushBackInstance(gm,gtFlat.astype(opengm.label_type))
+
+
+
+# for grid search learner
+lowerBounds = numpy.ones(nWeights)*-2.0
+upperBounds = numpy.ones(nWeights)*2.0
+nTestPoints  =numpy.ones(nWeights).astype('uint64')*10
+
+
+learner = learning.gridSearchLearner(dataset=dataset,lowerBounds=lowerBounds, upperBounds=upperBounds,nTestPoints=nTestPoints)
+
+learner.learn(infCls=opengm.inference.BeliefPropagation, 
+              parameter=opengm.InfParam(damping=0.5))

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-science/packages/opengm.git