def _declare_variables(self): with tf.variable_scope('vfeedbacknet_model1'): with tf.variable_scope('convlstm1'): with tf.variable_scope('rnn'): with tf.variable_scope('conv_lstm_cell'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() n = 512 m = 4*n input_size = [7, 7, n] kernel2d_size = [3, 3] kernel_size = kernel2d_size + [2*n] + [m] with tf.variable_scope('convlstm'): kernel = tf.get_variable('kernel', kernel_size, initializer=initializer, regularizer=regularizer) W_ci = tf.get_variable('W_ci', input_size, initializer=initializer, regularizer=regularizer) W_cf = tf.get_variable('W_cf', input_size, initializer=initializer, regularizer=regularizer) W_co = tf.get_variable('W_co', input_size, initializer=initializer, regularizer=regularizer) bias = tf.get_variable('bias', [m], initializer=tf.zeros_initializer(), regularizer=regularizer) self.convLSTMCell1 = ConvLSTMCell([7, 7], 512, [3, 3]) with tf.variable_scope('convlstm2'): with tf.variable_scope('rnn'): with tf.variable_scope('conv_lstm_cell'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() n = 512 m = 4*n input_size = [7, 7, n] kernel2d_size = [3, 3] kernel_size = kernel2d_size + [2*n] + [m] with tf.variable_scope('convlstm'): kernel = tf.get_variable('kernel', kernel_size, initializer=initializer, regularizer=regularizer) W_ci = tf.get_variable('W_ci', input_size, initializer=initializer, regularizer=regularizer) W_cf = tf.get_variable('W_cf', input_size, initializer=initializer, regularizer=regularizer) W_co = tf.get_variable('W_co', input_size, initializer=initializer, regularizer=regularizer) bias = tf.get_variable('bias', [m], initializer=tf.zeros_initializer(), regularizer=regularizer) self.convLSTMCell2 = ConvLSTMCell([7, 7], 512, [3, 3])
def _declare_variables(self): with tf.variable_scope('vfeedbacknet_{}'.format(Model.model_name)): # with tf.variable_scope('process_featurizer_output'): # with tf.variable_scope('conv1'): # regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) # initializer = tf.contrib.layers.xavier_initializer() # kernel = tf.get_variable('kernel', shape=[3, 3, 512, 512], dtype=tf.float32, regularizer=regularizer, initializer=initializer) # biases = tf.get_variable('biases', shape=[512], dtype=tf.float32, regularizer=regularizer, initializer=initializer) with tf.variable_scope('convlstm1'): with tf.variable_scope('rnn'): with tf.variable_scope('conv_lstm_cell'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() n = 512 m = 4 * n input_size = [7, 7, n] kernel2d_size = [3, 3] kernel_size = kernel2d_size + [2 * n] + [m] with tf.variable_scope('convlstm'): kernel = tf.get_variable('kernel', kernel_size, initializer=initializer, regularizer=regularizer) W_ci = tf.get_variable('W_ci', input_size, initializer=initializer, regularizer=regularizer) W_cf = tf.get_variable('W_cf', input_size, initializer=initializer, regularizer=regularizer) W_co = tf.get_variable('W_co', input_size, initializer=initializer, regularizer=regularizer) bias = tf.get_variable( 'bias', [m], initializer=tf.zeros_initializer(), regularizer=regularizer) self.convLSTMCell1 = ConvLSTMCell(input_size[:2], n, [3, 3]) # with tf.variable_scope('convlstm2'): # with tf.variable_scope('rnn'): # with tf.variable_scope('conv_lstm_cell'): # regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) # initializer = tf.contrib.layers.xavier_initializer() # n = 512 # m = 4*n # input_size = [7, 7, n] # kernel2d_size = [3, 3] # kernel_size = kernel2d_size + [2*n] + [m] # with tf.variable_scope('convlstm'): # kernel = tf.get_variable('kernel', kernel_size, initializer=initializer, regularizer=regularizer) # W_ci = tf.get_variable('W_ci', input_size, initializer=initializer, regularizer=regularizer) # W_cf = tf.get_variable('W_cf', input_size, initializer=initializer, regularizer=regularizer) # W_co = tf.get_variable('W_co', input_size, initializer=initializer, regularizer=regularizer) # bias = tf.get_variable('bias', [m], initializer=tf.zeros_initializer(), regularizer=regularizer) # self.convLSTMCell2 = ConvLSTMCell(input_size[:2], n, [3, 3]) with tf.variable_scope('reshape_convs'): with tf.variable_scope('conv1'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() kernel = tf.get_variable('kernel', shape=[3, 3, 256, 512], dtype=tf.float32, regularizer=regularizer, initializer=initializer) biases = tf.get_variable('biases', shape=[512], dtype=tf.float32, regularizer=regularizer, initializer=initializer) # with tf.variable_scope('conv2'): # regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) # initializer = tf.contrib.layers.xavier_initializer() # kernel = tf.get_variable('kernel', shape=[3, 3, 512, 1024], dtype=tf.float32, regularizer=regularizer, initializer=initializer) # biases = tf.get_variable('biases', shape=[1024], dtype=tf.float32, regularizer=regularizer, initializer=initializer) # with tf.variable_scope('conv3'): # regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) # initializer = tf.contrib.layers.xavier_initializer() # kernel = tf.get_variable('kernel', shape=[3, 3, 512, 1024], dtype=tf.float32, regularizer=regularizer, initializer=initializer) # biases = tf.get_variable('biases', shape=[1024], dtype=tf.float32, regularizer=regularizer, initializer=initializer) with tf.variable_scope('feedback_block1'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() input_size = [7, 7] kernel_size = [3, 3, 512, 512] W_xf = tf.get_variable('W_xf', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_xi = tf.get_variable('W_xi', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_xc = tf.get_variable('W_xc', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_xo = tf.get_variable('W_xo', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_hf = tf.get_variable('W_hf', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_hi = tf.get_variable('W_hi', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_hc = tf.get_variable('W_hc', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_ho = tf.get_variable('W_ho', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_cf = tf.get_variable( 'W_cf', [input_size[0], input_size[1], kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_ci = tf.get_variable( 'W_ci', [input_size[0], input_size[1], kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_co = tf.get_variable( 'W_co', [input_size[0], input_size[1], kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) b_f = tf.get_variable('b_f', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) b_i = tf.get_variable('b_i', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) b_c = tf.get_variable('b_c', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) b_o = tf.get_variable('b_o', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) # with tf.variable_scope('feedback_block2'): # regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) # initializer = tf.contrib.layers.xavier_initializer() # input_size = [14, 14] # kernel_size = [3, 3, 256, 256] # W_xf = tf.get_variable('W_xf', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_xi = tf.get_variable('W_xi', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_xc = tf.get_variable('W_xc', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_xo = tf.get_variable('W_xo', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_hf = tf.get_variable('W_hf', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_hi = tf.get_variable('W_hi', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_hc = tf.get_variable('W_hc', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_ho = tf.get_variable('W_ho', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_cf = tf.get_variable('W_cf', [input_size[0],input_size[1],kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_ci = tf.get_variable('W_ci', [input_size[0],input_size[1],kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_co = tf.get_variable('W_co', [input_size[0],input_size[1],kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) # b_f = tf.get_variable('b_f', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) # b_i = tf.get_variable('b_i', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) # b_c = tf.get_variable('b_c', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) # b_o = tf.get_variable('b_o', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) # with tf.variable_scope('feedback_block3'): # regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) # initializer = tf.contrib.layers.xavier_initializer() # input_size = [7, 7] # kernel_size = [3, 3, 512, 512] # W_xf = tf.get_variable('W_xf', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_xi = tf.get_variable('W_xi', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_xc = tf.get_variable('W_xc', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_xo = tf.get_variable('W_xo', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_hf = tf.get_variable('W_hf', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_hi = tf.get_variable('W_hi', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_hc = tf.get_variable('W_hc', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_ho = tf.get_variable('W_ho', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_cf = tf.get_variable('W_cf', [input_size[0],input_size[1],kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_ci = tf.get_variable('W_ci', [input_size[0],input_size[1],kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) # W_co = tf.get_variable('W_co', [input_size[0],input_size[1],kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) # b_f = tf.get_variable('b_f', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) # b_i = tf.get_variable('b_i', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) # b_c = tf.get_variable('b_c', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) # b_o = tf.get_variable('b_o', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) with tf.variable_scope('fc'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() trainable = False if self.train_fc == 'NO' else True weight = tf.get_variable('weights', shape=[512, self.num_classes], dtype=tf.float32, initializer=initializer, regularizer=regularizer, trainable=trainable) biases = tf.get_variable('biases', shape=[self.num_classes], dtype=tf.float32, initializer=initializer, regularizer=regularizer, trainable=trainable)
def _declare_variables(self): with tf.variable_scope('vfeedbacknet_{}'.format(Model.model_name)): with tf.variable_scope('feedbackcell1'): self.feedbackLSTMCell1 = FeedbackLSTMCell_stack1( [14, 14, 128], Model.NFEEDBACK) with tf.variable_scope('convlstm1'): with tf.variable_scope('rnn'): with tf.variable_scope('conv_lstm_cell'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() n = 128 m = 4 * n input_size = [7, 7, n] kernel2d_size = [3, 3] kernel_size = kernel2d_size + [2 * n] + [m] with tf.variable_scope('convlstm'): kernel = tf.get_variable('kernel', kernel_size, initializer=initializer, regularizer=regularizer) W_ci = tf.get_variable('W_ci', input_size, initializer=initializer, regularizer=regularizer) W_cf = tf.get_variable('W_cf', input_size, initializer=initializer, regularizer=regularizer) W_co = tf.get_variable('W_co', input_size, initializer=initializer, regularizer=regularizer) bias = tf.get_variable( 'bias', [m], initializer=tf.zeros_initializer(), regularizer=regularizer) self.convLSTMCell1 = ConvLSTMCell(input_size[:2], n, [3, 3]) with tf.variable_scope('reshape_convs'): with tf.variable_scope('conv1'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() kernel = tf.get_variable('kernel', shape=[7, 7, 3, 32], dtype=tf.float32, regularizer=regularizer, initializer=initializer) biases = tf.get_variable('biases', shape=[32], dtype=tf.float32, regularizer=regularizer, initializer=initializer) with tf.variable_scope('conv2'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() kernel = tf.get_variable('kernel', shape=[3, 3, 32, 64], dtype=tf.float32, regularizer=regularizer, initializer=initializer) biases = tf.get_variable('biases', shape=[64], dtype=tf.float32, regularizer=regularizer, initializer=initializer) with tf.variable_scope('conv3'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() kernel = tf.get_variable('kernel', shape=[3, 3, 64, 128], dtype=tf.float32, regularizer=regularizer, initializer=initializer) biases = tf.get_variable('biases', shape=[128], dtype=tf.float32, regularizer=regularizer, initializer=initializer) # with tf.variable_scope('conv4'): # regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) # initializer = tf.contrib.layers.xavier_initializer() # kernel = tf.get_variable('kernel', shape=[3, 3, 512, 1024], dtype=tf.float32, regularizer=regularizer, initializer=initializer) # biases = tf.get_variable('biases', shape=[1024], dtype=tf.float32, regularizer=regularizer, initializer=initializer) # with tf.variable_scope('conv5'): # regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) # initializer = tf.contrib.layers.xavier_initializer() # kernel = tf.get_variable('kernel', shape=[3, 3, 128, 256], dtype=tf.float32, regularizer=regularizer, initializer=initializer) # biases = tf.get_variable('biases', shape=[256], dtype=tf.float32, regularizer=regularizer, initializer=initializer) with tf.variable_scope('fc1'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() trainable = False if self.train_fc == 'NO' else True weight = tf.get_variable('weights', shape=[128, 128], dtype=tf.float32, initializer=initializer, regularizer=regularizer, trainable=trainable) biases = tf.get_variable('biases', shape=[128], dtype=tf.float32, initializer=initializer, regularizer=regularizer, trainable=trainable) with tf.variable_scope('fc2'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() trainable = False if self.train_fc == 'NO' else True weight = tf.get_variable('weights', shape=[128, self.num_classes], dtype=tf.float32, initializer=initializer, regularizer=regularizer, trainable=trainable) biases = tf.get_variable('biases', shape=[self.num_classes], dtype=tf.float32, initializer=initializer, regularizer=regularizer, trainable=trainable)
def _declare_variables(self): with tf.variable_scope('vfeedbacknet_{}'.format(Model.model_name)): with tf.variable_scope('convlstm1'): with tf.variable_scope('rnn'): with tf.variable_scope('conv_lstm_cell'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() n = 256 m = 4 * n input_size = [14, 14, n] kernel2d_size = [3, 3] kernel_size = kernel2d_size + [2 * n] + [m] with tf.variable_scope('convlstm'): kernel = tf.get_variable('kernel', kernel_size, initializer=initializer, regularizer=regularizer) W_ci = tf.get_variable('W_ci', input_size, initializer=initializer, regularizer=regularizer) W_cf = tf.get_variable('W_cf', input_size, initializer=initializer, regularizer=regularizer) W_co = tf.get_variable('W_co', input_size, initializer=initializer, regularizer=regularizer) bias = tf.get_variable( 'bias', [m], initializer=tf.zeros_initializer(), regularizer=regularizer) self.convLSTMCell1 = ConvLSTMCell([14, 14], 256, [3, 3]) with tf.variable_scope('conv1'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() kernel = tf.get_variable('kernel', shape=[14, 14, 256, 512], dtype=tf.float32, regularizer=regularizer, initializer=initializer) biases = tf.get_variable('biases', shape=[512], dtype=tf.float32, regularizer=regularizer, initializer=initializer) with tf.variable_scope('dconv1'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() kernel = tf.get_variable('kernel', [3, 3, 128, 256], initializer=initializer, regularizer=regularizer) biases = tf.get_variable('biases', [128], initializer=initializer, regularizer=regularizer)
def _declare_variables(self): with tf.variable_scope('vfeedbacknet_{}'.format(Model.model_name)): with tf.variable_scope('process_featurizer_output'): with tf.variable_scope('conv1'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() kernel = tf.get_variable('kernel', shape=[3, 3, 128, 128], dtype=tf.float32, regularizer=regularizer, initializer=initializer) biases = tf.get_variable('biases', shape=[128], dtype=tf.float32, regularizer=regularizer, initializer=initializer) with tf.variable_scope('convlstm1'): with tf.variable_scope('rnn'): with tf.variable_scope('conv_lstm_cell'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() n = 128 m = 4 * n input_size = [14, 14, n] kernel2d_size = [3, 3] kernel_size = kernel2d_size + [2 * n] + [m] with tf.variable_scope('convlstm'): kernel = tf.get_variable('kernel', kernel_size, initializer=initializer, regularizer=regularizer) W_ci = tf.get_variable('W_ci', input_size, initializer=initializer, regularizer=regularizer) W_cf = tf.get_variable('W_cf', input_size, initializer=initializer, regularizer=regularizer) W_co = tf.get_variable('W_co', input_size, initializer=initializer, regularizer=regularizer) bias = tf.get_variable( 'bias', [m], initializer=tf.zeros_initializer(), regularizer=regularizer) self.convLSTMCell1 = ConvLSTMCell([14, 14], 128, [3, 3]) with tf.variable_scope('feedback_block1'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() kernel_size = [3, 3, 128, 128] W_xf = tf.get_variable('W_xf', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_xi = tf.get_variable('W_xi', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_xc = tf.get_variable('W_xc', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_xo = tf.get_variable('W_xo', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_hf = tf.get_variable('W_hf', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_hi = tf.get_variable('W_hi', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_hc = tf.get_variable('W_hc', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_ho = tf.get_variable('W_ho', kernel_size, dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_cf = tf.get_variable('W_cf', [14, 14, kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_ci = tf.get_variable('W_ci', [14, 14, kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) W_co = tf.get_variable('W_co', [14, 14, kernel_size[-1]], dtype=tf.float32, initializer=initializer, regularizer=regularizer) b_f = tf.get_variable('b_f', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) b_i = tf.get_variable('b_i', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) b_c = tf.get_variable('b_c', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) b_o = tf.get_variable('b_o', [kernel_size[-1]], dtype=tf.float32, initializer=tf.zeros_initializer(), regularizer=regularizer) with tf.variable_scope('fc'): regularizer = None # tf.contrib.layers.l2_regularizer(scale=0.25) initializer = tf.contrib.layers.xavier_initializer() trainable = False if self.train_fc == 'NO' else True weight = tf.get_variable('weights', shape=[128, self.num_classes], dtype=tf.float32, initializer=initializer, regularizer=regularizer, trainable=trainable) biases = tf.get_variable('biases', shape=[self.num_classes], dtype=tf.float32, initializer=initializer, regularizer=regularizer, trainable=trainable)