def buildNetwork(self):
"""
Builds the neural network with a fixed structure,
and a variable number of outputs.
"""
self.inputLayer = InputLayer()
convLayer = ConvolutionalLayer(5,10)
poolLayer = PoolingLayer(4)
reluLayer = ReluLayer()
convLayer2 = ConvolutionalLayer(4,20)
pool2Layer = PoolingLayer(2)
flattenLayer = FlattenLayer()
reluLayer2 = ReluLayer()
fullLayer = FullyConnectedLayer(20)
self.outputLayer = OutputLayer(self.numOutputs)
fullLayer.connect(self.outputLayer)
flattenLayer.connect(fullLayer)
reluLayer2.connect(flattenLayer)
pool2Layer.connect(reluLayer2)
convLayer2.connect(pool2Layer)
reluLayer.connect(convLayer2)
poolLayer.connect(reluLayer)
convLayer.connect(poolLayer)
self.inputLayer.connect(convLayer)
n_sample = 500 # Using for Error, Accuracy samples alpha = 1 # how many times do you want to learn in 1 epoch. if 1, whole data learns. input_train, input_test, correct_train, correct_test = cifar10_call(N) n_train = input_train.shape[0] n_test = input_test.shape[0] img_h = 32 img_w = 32 img_ch = 3 # -- 각 층의 초기화 -- cl1 = ConvLayer(img_ch, img_h, img_w, 30, 3, 3, stride=1, pad=1) # 앞3개:인풋 중간3개:필터 cl2 = ConvLayer(cl1.y_ch, cl1.y_h, cl1.y_w, 30, 3, 3, stride=1, pad=1) pl1 = PoolingLayer(cl2.y_ch, cl2.y_h, cl2.y_w, pool=2, pad=0) # pool:풀링크기(2*2), pad:패딩 너비 c_dr1 = fn.dropout(0.25) cl3 = ConvLayer(pl1.y_ch, pl1.y_h, pl1.y_w, 60, 3, 3, stride=1, pad=1) pl2 = PoolingLayer(cl3.y_ch, cl3.y_h, cl3.y_w, pool=2, pad=0) c_dr2 = fn.dropout(0.25) cl4 = ConvLayer(pl2.y_ch, pl2.y_h, pl2.y_w, 120, 3, 3, stride=1, pad=1) pl3 = PoolingLayer(cl4.y_ch, cl4.y_h, cl4.y_w, pool=2, pad=0) n_fc_in = pl3.y_ch * pl3.y_h * pl3.y_w ml1 = MiddleLayer(n_fc_in, 500) dr1 = fn.dropout(0.5) ml2 = MiddleLayer(500, 500) dr2 = fn.dropout(0.5) ol1 = OutputLayer(500, 10)
def __init__(self, phase, start_from_scratch= False ):
encoding_t = 10
self.encoding_t = encoding_t
self.phase = phase
path = dirname(dirname(realpath(__file__)))
self.path_to_img_folder = path + '/datasets/'+ phase + 'Set/'
classifier_dataset_path = path + '/datasets/ClassifierSet/'
self.classifier_training_dataset_path = classifier_dataset_path +'TrainingData.csv'
self.classifier_testing_dataset_path = classifier_dataset_path +'TestingData.csv'
self.start_from_scratch = start_from_scratch
self.path_to_log_file = path+ '/logs/log.csv'
self.pathWeights = path + '/weights/'
strides_conv= [1,1,1,1]
padding= "SAME"
pooling_type= "MAX"
stdp_flag = phase == 'Learning'
''' after 20 img the first layer converges
the second has some weights around 1 and the others
are in the initial value range : .68 score
encoding_t,.02,-.0,-.02, stdp_flag ),
encoding_t,.006,-.0,-.002, stdp_flag),
'''
'''after some img both layers converges
with scary simmetry of weights around 0
and around 1: same number
score 0.67
encoding_t,.01,-.0,-.012, stdp_flag ),
encoding_t,.01,-.0,-.02, stdp_flag),
maybe try a slower convergence
'''
'''
self.layers = [
ConvolutionalLayer(padding, strides_conv,
[5,5,1,4],6.5, [1,160,250,1], [1,160,250,4],
encoding_t,5,.004,-.0,-.008, stdp_flag ),
PoolingLayer(padding, [6,6], [7,7], pooling_type, [1,27,42,4]),
ConvolutionalLayer(padding,strides_conv,
[17,17,4,20], 35., [1,27,42,4], [1,27,42,20],
encoding_t,25,.002,-.0,-.004, stdp_flag),
PoolingLayer(padding, [5,5], [5,5], pooling_type, [1,6,9,20]),
ConvolutionalLayer(padding, strides_conv,
[5,5,20,20], math.inf , [1,6,9,20], [1,6,9,20],
encoding_t,0,.0,-.0,-.0, stdp_flag)
]
'''
self.layers = [
ConvolutionalLayer(padding, strides_conv,
[5,5,1,4],3.7, [1,160,250,1], [1,160,250,4],
encoding_t,18,.004,-.0,-.008, stdp_flag ),
PoolingLayer(padding, [6,6], [7,7], pooling_type, [1,27,42,4]),
ConvolutionalLayer(padding,strides_conv,
[17,17,4,20], 25., [1,27,42,4], [1,27,42,20],
encoding_t,28,.002,-.0,-.004, stdp_flag),
PoolingLayer(padding, [5,5], [5,5], pooling_type, [1,6,9,20]),
ConvolutionalLayer(padding, strides_conv,
[5,5,20,20], math.inf , [1,6,9,20], [1,6,9,20],
encoding_t,0,.0,-.0,-.0, stdp_flag)
]
if start_from_scratch:
self.writeFeatureNamesinDataset( classifier_dataset_path +'TrainingData.csv')
self.writeFeatureNamesinDataset( classifier_dataset_path +'TestingData.csv')
