class train_tf_cc_output:
decoded_data_train_cc1 = np.array([])
encoded_data_train_cc1 = np.array([])
decoded_data_valid_cc1 = np.array([])
encoded_data_valid_cc1 = np.array([])
decoded_data_test_cc1 = np.array([])
encoded_data_test_cc1 = np.array([])
obj_classifier = classifier_data()
def function(self):
print("This is train_tensorflow_cc_output class")
class train_tf_cc_input:
cc1_input_train_perm = np.array([])
cc1_output_train_perm = np.array([])
cc1_input_valid_perm = np.array([])
cc1_output_valid_perm = np.array([])
obj_classifier_data = classifier_data()
dimension_hidden_layer1 = []
EPOCHS_CC = []
classI = []
classJ = []
dataset_name = []
dim_feature = []
def function(self):
print("This is train_tensorflow_cc_input class")
#......................cc.......................
number_of_cc = number_of_train_classes * number_of_train_classes - number_of_train_classes
print "Number of c coders %d " % number_of_cc
cross_coders_train_data_input = []
cross_coders_train_data_output = []
# In[ ]:
#Get mean feature vector for each class
mean_feature_mat = np.empty((0, dimension_visual_data), float)
number_of_samples_per_class_train = []
number_of_samples_per_class_test = []
number_of_samples_per_class_valid = []
#pdb.set_trace()
obj_classifier = classifier_data()
cnt = 0
for classI in train_class_labels:
indices = np.flatnonzero(dataset_labels == classI)
classI_features = visual_features_dataset[indices.astype(int), :]
mean_feature = classI_features.mean(0)
mean_feature_mat = np.append(mean_feature_mat,
mean_feature.reshape(1,
dimension_visual_data),
axis=0)
number_of_samples_per_class_train.append(
int(TR_TS_VA_SPLIT[0] * np.size(indices)))
number_of_samples_per_class_test.append(
int(TR_TS_VA_SPLIT[2] * np.size(indices)))
number_of_samples_per_class_valid.append(
int(TR_TS_VA_SPLIT[1] * np.size(indices)))