def __init__(self, inputdim, outputdim, hidden_unit_lim, rng):
self.inputdim = inputdim
self.outputdim = outputdim
#self.inputarr=inputarr #self explanatory
self.hidden_unit_lim = hidden_unit_lim
self.rng = rng
rest_set, test_set = pimadataf.give_data()
self.restx = rest_set[0]
resty = rest_set[1]
self.testx = test_set[0]
testy = test_set[1]
self.resty = np.ravel(resty)
self.testy = np.ravel(testy)
self.rest_setx = tf.Variable(initial_value=self.restx,
name='rest_setx',
dtype=tf.float32)
self.rest_sety = tf.Variable(initial_value=self.resty,
name='rest_sety',
dtype=tf.int32)
self.test_setx = tf.Variable(initial_value=self.testx,
name='rest_sety',
dtype=tf.float32)
self.test_sety = tf.Variable(initial_value=self.testy,
name='test_sety',
dtype=tf.int32)
#self.inputarr = inputarr
self.inputarr = self.restx
def __init__(self,
max_no_of_hidden_units,
dimtup,
size=5,
limittup=(-1, 1)):
self.size = size
self.max_no_of_hidden_units = max_no_of_hidden_units
self.dimtup = dimtup
self.list_chromo = self.aux_pop(size, limittup) #a numpy array
self.fits_pops = []
rest_set, test_set = pimadataf.give_data(
) #one time thing #RTC required here
self.trainx = rest_set[0]
self.trainy = rest_set[1]
#print("hmm",self.trainy)
self.testx = test_set[0]
self.testy = test_set[1]
self.net_err = network.Neterr(inputdim=self.dimtup[0],
outputdim=self.dimtup[1],
arr_of_net=self.list_chromo,
trainx=self.trainx,
trainy=self.trainy,
testx=self.testx,
testy=self.testy)
def __init__(self,rng, max_hidden_units, size=50, limittup=(-1,1)):
self.dimtup = pimadataf.get_dimension()
rest_set, test_set = pimadataf.give_data()
tup = pimadataf.give_datainshared()
self.rng=rng
self.size = size
self.max_hidden_units = max_hidden_units
self.list_chromo = self.aux_pop(size, limittup) #a numpy array
self.fits_pops = []
self.trainx = rest_set[0]
self.trainy = rest_set[1]
self.testx = test_set[0]
self.testy = test_set[1]
self.strainx, self.strainy = tup[0]
self.stestx, self.stesty = tup[1]
self.net_err = Network.Neterr(inputdim=self.dimtup[0], outputdim=self.dimtup[1], arr_of_net=self.list_chromo, trainx=self.trainx, trainy=self.trainy, testx=self.testx, testy=self.testy,strainx=self.strainx, strainy=self.strainy, stestx=self.stestx, stesty=self.stesty)
self.net_dict={} #dictionary of networks for back-propagation, one for each n_hid
def __init__(self,rng, max_hidden_units, size=5, limittup=(-1,1)):
self.dimtup = pimadataf.get_dimension()
rest_set, test_set = pimadataf.give_data()
restx=rest_set[0]
resty=rest_set[1]
testx=test_set[0]
testy=test_set[1]
resty=np.ravel(resty)
testy=np.ravel(testy)
self.rng=rng
self.size = size
self.max_hidden_units = max_hidden_units
self.list_chromo = self.aux_pop(size, limittup) #a numpy array
self.fits_pops = []
restn=538 #a flaw here ,one has to know no. of datapoints in both set before opening it(inside program)
testn=230
print("here you",rest_set[1].shape)
self.rest_setx=tf.Variable(initial_value=np.zeros((restn,self.dimtup[0])),name='rest_setx',dtype=tf.float64)
self.rest_sety=tf.Variable(initial_value=np.zeros((restn,)),name='rest_sety',dtype=tf.int32)
self.test_setx=tf.Variable(initial_value=np.zeros((testn,self.dimtup[0])),name='rest_sety',dtype=tf.float64)
self.test_sety=tf.Variable(initial_value=np.zeros((testn,)),name='test_sety',dtype=tf.int32)
if not os.path.isfile('/home/robita/forgit/neuro-evolution/05/state/tf/indep_pima/input/model.ckpt.meta'):
rxn=self.rest_setx.assign(restx)
ryn=self.rest_sety.assign(resty)
txn=self.test_setx.assign(testx)
tyn=self.test_sety.assign(testy)
var_lis=[self.rest_setx,self.rest_sety,self.test_setx,self.test_sety]
nodelis=[rxn,ryn,txn,tyn]
savo=tf.train.Saver(var_list=var_lis)
with tf.Session() as sess:
sess.run([i for i in nodelis])
print("saving checkpoint")
save_path = savo.save(sess, "/home/robita/forgit/neuro-evolution/05/state/tf/indep_pima/input/model.ckpt")
self.net_err = Network.Neterr(inputdim=self.dimtup[0], outputdim=self.dimtup[1], arr_of_net=self.list_chromo,rest_setx=self.rest_setx,rest_sety=self.rest_sety,test_setx=self.test_setx,test_sety=self.test_sety,rng=self.rng)
self.net_dict={} #dictionary of networks for back-propagation, one for each n_hid