# Ops for initializing the two different iterators training_init_op = train_iterator.make_initializer(tr_data.data) validation_init_op = valid_iterator.make_initializer(val_data.data) # TF placeholder for graph input and output x = tf.placeholder(tf.float32, [batch_size, 227, 227, 3]) y = tf.placeholder(tf.float32, [batch_size, num_classes]) keep_prob = tf.placeholder(tf.float32) # Initialize model model = AlexNet(x, keep_prob, num_classes) # Load the pretrained weights into the non-trainable layer load_op = model.load_initial_weights_ops() # Link variable to model output logits = model.fc8 # Op for calculating the loss with tf.name_scope("cross_ent"): loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=y)) var_list = [v for v in tf.trainable_variables()] for var in var_list: if 'weights' in var.name.split('/')[1]: weights[var.name.split('/')[0]] = var # Train op with tf.name_scope("train"):