lr=0.001
hidden0=500
hidden1=400
ut.log_p('drop_mlp4da.py|ad:'+advertiser+'|drop:'+str(dropout)+'|b_size:'+str(batch_size)+' | X:'+str(x_dim) + ' | Hidden 0:'+str(hidden0)+ ' | Hidden 1:'+str(hidden1)+ ' | Hidden 2:'+str(hidden2)+
' | L_r:'+str(lr)+ ' | activation1:'+ str(acti_type)+
' | lambda:'+str(lambda1)
)
# initialise parameters
arr=[]
arr.append(x_dim)
arr.append(hidden0)
arr.append(hidden1)
arr.append(hidden2)
ww0,bb0=ut.init_weight(x_dim,hidden0,'sigmoid')
ww1,bb1=ut.init_weight(hidden0,hidden1,'sigmoid')
ww2,bb2=ut.init_weight(hidden1,hidden2,'sigmoid')
# ww0,bb0,ww1,bb1,ww2,bb2=da.get_da_weights(train_file,arr,ncases=train_size,batch_size=100000)
# pickle.dump( (ww0,bb0,ww1,bb1,ww2,bb2), open( "2997_da_4l_10.p", "wb" ))
# (ww0,bb0,ww1,bb1,ww2,bb2)=pickle.load( open( "2997_da_4l_10.p", "rb" ) )
# ww2,bb2=ut.init_weight(hidden1,hidden2,'sigmoid')
ww3=rng.uniform(-0.05,0.05,hidden2)
ww3=numpy.zeros(hidden2)
#
bb3=0.
ut.log_p('bat size:'+str(batch_size)+'|X:'+str(x_dim) + ' | Hidden 0:'+str(hidden0)+ ' | Hidden 1:'+str(hidden1)+ ' | Hidden 2:'+str(hidden2)+
' | L rate:'+str(lr)+ ' | activation1:'+ str(acti_type)+
' | lambda:'+str(lambda1)
)
# initialise parameters
arr=[]
arr.append(x_dim)
arr.append(hidden0)
arr.append(hidden1)
arr.append(hidden2)
# ww0,bb0=ut.init_weight(x_dim,hidden0,'sigmoid')
ww0=numpy.asarray(clicks)
bb0=numpy.zeros(hidden0)
ww1,bb1=ut.init_weight(hidden0,hidden1,'sigmoid')
ww2,bb2=ut.init_weight(hidden1,hidden2,'sigmoid')
# ww0,bb0,ww1,bb1,ww2,bb2=da.get_da_weights(train_file,arr,ncases=train_size,batch_size=100000)
# pickle.dump( (ww0,bb0,ww1,bb1,ww2,bb2), open( "2997_da_4l_10.p", "wb" ))
# (ww0,bb0,ww1,bb1,ww2,bb2)=pickle.load( open( "2997_da_4l_10.p", "rb" ) )
# ww2,bb2=ut.init_weight(hidden1,hidden2,'sigmoid')
ww3=rng.uniform(-0.05,0.05,hidden2)
ww3=numpy.zeros(hidden2)
#
bb3=0.
arr=[]
log_p('drop_mlp4da.py|ad:' + advertiser + '|drop:' + str(dropout) +
'|b_size:' + str(batch_size) + ' | X:' + str(x_dim) + ' | Hidden 0:' +
str(hidden0) + ' | Hidden 1:' + str(hidden1) + ' | Hidden 2:' +
str(hidden2) + ' | L_r:' + str(lr) + ' | activation1:' + str(acti_type) +
' | lambda:' + str(lambda1))
# initialise parameters
arr = []
arr.append(x_dim)
arr.append(hidden0)
arr.append(hidden1)
arr.append(hidden2)
ww0, bb0 = ut.init_weight(x_dim, hidden0, 'sigmoid')
ww1, bb1 = ut.init_weight(hidden0, hidden1, 'sigmoid')
ww2, bb2 = ut.init_weight(hidden1, hidden2, 'sigmoid')
wfile = "rbm_" + str(advertiser) + "_.p"
if os.path.isfile(wfile):
(ww0, bb0, ww1, bb1, ww2, bb2) = pickle.load(open(wfile, "rb"))
else:
ww0, bb0, ww1, bb1, ww2, bb2 = gbrbm.get_rbm_weights(
train_file,
arr,
ncases=train_size,
batch_size=100000,
fm_model_file=fm_model_file)
pickle.dump((ww0, bb0, ww1, bb1, ww2, bb2), open(wfile, "wb"))
return 1 + feat_field[feat] * k + l
def feats_to_layer_one_array(feats):
x = numpy.zeros(xdim)
x[0] = w_0
for feat in feats:
x[feat_layer_one_index(feat, 0):feat_layer_one_index(feat, k)] = feat_weights[feat]
return x
ut.log_p('X:'+str(xdim) + ' | Hidden 1:'+str(hidden1)+ ' | Hidden 2:'+str(hidden2)+
' | L rate:'+str(lr)+ ' | activation1:'+ str(acti_type)+
' | lambda:'+str(lambda1)
)
# initialise parameters
ww1,bb1=ut.init_weight(xdim,hidden1,acti_type)
ww2,bb2=ut.init_weight(hidden1,hidden2,acti_type)
ww3,bb3=ut.init_weight(hidden2,hidden3,acti_type)
ww4,bb4=ut.init_weight(hidden3,hidden4,acti_type)
ww5=numpy.zeros(hidden2)
bb5=0.
# Declare Theano symbolic variables
x = T.matrix("x")
y = T.vector("y")
w1 = theano.shared(ww1, name="w1")
w2 = theano.shared(ww2, name="w2")
w3 = theano.shared(ww3, name="w3")
w4 = theano.shared(ww4, name="w4")
w5 = theano.shared(ww5, name="w5")
b1 = theano.shared(bb1, name="b1")
+ " | L_r:"
+ str(lr)
+ " | activation1:"
+ str(acti_type)
+ " | lambda:"
+ str(lambda1)
)
# initialise parameters
arr = []
arr.append(x_dim)
arr.append(hidden0)
arr.append(hidden1)
arr.append(hidden2)
ww0, bb0 = ut.init_weight(x_dim, hidden0, "sigmoid")
ww1, bb1 = ut.init_weight(hidden0, hidden1, "sigmoid")
ww2, bb2 = ut.init_weight(hidden1, hidden2, "sigmoid")
wfile = "rbm_" + str(advertiser) + "_.p"
if os.path.isfile(wfile):
(ww0, bb0, ww1, bb1, ww2, bb2) = pickle.load(open(wfile, "rb"))
else:
ww0, bb0, ww1, bb1, ww2, bb2 = gbrbm.get_rbm_weights(
train_file, arr, ncases=train_size, batch_size=100000, fm_model_file=fm_model_file
)
pickle.dump((ww0, bb0, ww1, bb1, ww2, bb2), open(wfile, "wb"))
# ww2,bb2=ut.init_weight(hidden1,hidden2,'sigmoid')
ww3 = rng.uniform(-0.05, 0.05, hidden2)
