def __init__(self, rs):
regression.__init__(self,rs)
self.activation = []
self.W=[]
self.b=[]
self.size=0
#no hidden Layer
if(len(rs.hiddenLayers)==0):
#connection between input and output Layer
self.W.append(np.empty((rs.inputDim,rs.outputDim)))
self.b.append(np.empty(rs.outputDim))
self.activation.append(layersDict[rs.outputLayer])
self.size=rs.inputDim*rs.outputDim+rs.outputDim
else :
#hidden Layers
precDim=rs.inputDim
for layer in rs.hiddenLayers:
self.W.append(np.empty((precDim,layer[1])))
self.b.append(np.empty(layer[1]))
self.activation.append(layersDict[layer[0]])
precDim=layer[1]
self.size+=precDim*layer[1]+layer[1]
#outputLayer
self.W.append(np.empty((precDim,rs.outputDim)))
self.b.append(np.empty(rs.outputDim))
self.activation.append(layersDict[rs.outputLayer])
self.size+=precDim*rs.outputDim+rs.outputDim
self.theta=np.empty(self.size)
def __init__(self, rs):
'''
Initializes class parameters
Input: -nbFeature: int, number of feature in order to perform the regression
'''
regression.__init__(self, rs)
self.nbFeat = rs.numfeats
self.theta = np.zeros((self.nbFeat, self.outputDimension))
def __init__(self, rs):
regression.__init__(self,rs)
self.learningRate=rs.learningRate
self.momentum=rs.momentum
self.nbW=0
self.nbB=0
self.x = tf.placeholder(tf.float32, shape=[None, rs.inputDim])
self.y_ = tf.placeholder(tf.float32, shape=[None, rs.outputDim])
#no hidden Layer
if(len(rs.hiddenLayers)==0):
#connection between input and output Layer
in_to_out = [self.weight_variable([rs.inputDim,rs.outputDim]),self.bias_variable([rs.outputDim])]
self.y = layersDict[rs.outputLayer](tf.matmul(self.x,in_to_out[0]) + in_to_out[1])
self.listTheta=[in_to_out]
self.listBias=[in_to_out[1]]
self.listWeight=[in_to_out[0]]
self.theta = np.empty(rs.inputDim*rs.outputDim+rs.outputDim)
else :
#hidden Layers
self.listTheta=[]
self.listBias=[]
self.listWeight=[]
precDim=rs.inputDim
y=self.x
size=0
for layer in rs.hiddenLayers:
tmp=[self.weight_variable([precDim,layer[1]]),self.bias_variable([layer[1]])]
self.listTheta.append(tmp)
self.listBias.append(tmp[1])
self.listWeight.append(tmp[0])
y=layersDict[layer[0]](tf.matmul(y,tmp[0]) + tmp[1])
size+=precDim*layer[1]+layer[1]
precDim=layer[1]
tmp=[self.weight_variable([precDim,rs.outputDim]),self.bias_variable([rs.outputDim])]
self.listTheta.append(tmp)
self.listBias.append(tmp[1])
self.listWeight.append(tmp[0])
self.y=layersDict[rs.outputLayer](tf.matmul(y,tmp[0]) + tmp[1])
size+=precDim*rs.outputDim+rs.outputDim
self.theta=np.empty(size)
self.meanSquareError = tf.reduce_mean(tf.square(tf.sub(self.y_,self.y)))
self.train_step = tf.train.AdamOptimizer(rs.learningRate).minimize(self.meanSquareError)
self.init_op = tf.initialize_all_variables()
self.saver = tf.train.Saver(self.listBias+self.listWeight)
self.sess = tf.Session()
self.sess.run(self.init_op)
self.sess.as_default()
def __init__(self, rs):
regression.__init__(self,rs)
self.learningRate=rs.learningRate
self.momentum=rs.momentum
self.net = FeedForwardNetwork()
#input Layer
inLayer = layersDict[rs.inputLayer](rs.inputDim)
self.net.addInputModule(inLayer)
#outputLayer
outLayer = layersDict[rs.outputLayer](rs.outputDim)
self.net.addOutputModule(outLayer)
#no hidden Layer
if(len(rs.hiddenLayers)==0):
#connection between input and output Layer
in_to_out = FullConnection(inLayer, outLayer)
self.net.addConnection(in_to_out)
if(rs.bias==True):
bias= BiasUnit('bias')
self.net.addModule(bias)
bias_to_out = FullConnection(bias, outLayer)
self.net.addConnection(bias_to_out)
else :
#hidden Layers
hiddenLayers=[]
for layer in rs.hiddenLayers:
tmp=layersDict[layer[0]](layer[1])
self.net.addModule(tmp)
hiddenLayers.append(tmp)
#connection between input and first hidden Layer
in_to_hidden=FullConnection(inLayer,hiddenLayers[0])
self.net.addConnection(in_to_hidden)
#connection between hidden Layers
i=0
for i in range(1,len(hiddenLayers)):
hidden_to_hidden=FullConnection(hiddenLayers[i-1],hiddenLayers[i])
self.net.addConnection(hidden_to_hidden)
#connection between last hidden Layer and output Layer
hidden_to_out= FullConnection(hiddenLayers[i],outLayer)
self.net.addConnection(hidden_to_out)
if(rs.bias==True):
bias=BiasUnit('bias')
self.net.addModule(bias)
for layer in hiddenLayers :
bias_to_hidden = FullConnection(bias, layer)
self.net.addConnection(bias_to_hidden)
bias_to_out = FullConnection(bias, outLayer)
self.net.addConnection(bias_to_out)
#initilisation of weight
self.net.sortModules()
self.shape=self.net.params.shape
self.net._setParameters(np.random.normal(0.0,0.1,self.shape))
self.ds = SupervisedDataSet(self.inputDimension, self.outputDimension)
