MODEL_DIR = '/content/imagenet-vgg-f.mat' DATA_DIR = '/content/mirflickr25k.mat' phase = 'train' checkpoint_dir = './checkpoint' Savecode = './Savecode' dataset_dir = 'Flickr' netStr = 'alex' SEMANTIC_EMBED = 512 MAX_ITER = 100 batch_size = 128 image_size = 224 images, tags, labels = loading_data(DATA_DIR) dimTxt = tags.shape[1] dimLab = labels.shape[1] DATABASE_SIZE = 18015 TRAINING_SIZE = 10000 QUERY_SIZE = 2000 VERIFICATION_SIZE = 1000 X, Y, L = split_data(images, tags, labels, QUERY_SIZE, TRAINING_SIZE, DATABASE_SIZE) train_L = L['train'] train_x = X['train'] train_y = Y['train'] query_L = L['query'] query_x = X['query']
# Adding parameters from other python files # parser = add_parms_supervised(parser) parser = add_params_data(parser) parser = add_params_report(parser) parser = add_params_noise(parser) # parser = add_params_mt(parser) args = parser.parse_args() # lr=0.0001 # epochs=30 # batch_size= 64 # #for mean teacher # ratio =0.5 # alpha=0.99 #(0.90-0.99) # maxlen=100 x_train, y_train, x_test, y_test, x_unlabel = loading_data(args) x_train, x_test, x_unlabel, vocab_size, tokenizer = tokenization( args, x_train, x_test, x_unlabel, args.maxlen) for i in range(0, 1): x_train, y_train, x_test, y_test = Kfold_crossvalidation( args, x_train, y_train, x_test, y_test) print("train Data_Size:", np.shape(x_train)) print("test Data_Size:", np.shape(x_test)) print('Train Label count: True, Fake', np.count_nonzero(y_train == 1), np.count_nonzero(y_train == 0)) print('Test Label count : True, Fake', np.count_nonzero(y_test == 1), np.count_nonzero(y_test == 0)) # train_supervised(epochs, batch_size, lr,x_train, y_train, x_test, y_test,maxlen,vocab_size)