help="Multiple directories from other predictors", required=True) parser.add_argument( "--test-data-input-sep", default="\s+", help="Separator to use for loading test data CSV/TSV files", type=str) parser.add_argument( "--test-data-output-dir", help="Save combined test datasets to this directory", required=True) if __name__ == "__main__": args = parser.parse_args() dataframes, predictor_names = load_test_data( args.test_data_input_dirs, sep=args.test_data_input_sep) if not exists(args.test_data_output_dir): makedirs(args.test_data_output_dir) print("Loaded test data:") for (allele, df) in dataframes.items(): df.index.name = "sequence" print("%s: %d results" % (allele, len(df))) filename = "%s.csv" % allele filepath = join(args.test_data_output_dir, filename) df.to_csv(filepath)
model1.load_weights('./model_snapshots/X0001/weights-improvement-44-0.8930.h5') # model1.load_weights('./model_snapshots/X0002/weights-improvement-49-0.8977.h5') # model1.load_weights('./model_snapshots/X0003/weights-improvement-49-0.8874.h5') # model1.load_weights('./model_snapshots/X0004/weights-improvement-48-0.9002.h5') # model1.load_weights('./model_snapshots/X0005/weights-improvement-42-0.8846.h5') # 1,3,4,5 intermediate_model1 = Model( inputs=model1.input, outputs=[model1.get_layer('global_average_pooling2d_1').output]) # # 2 # intermediate_model1 = Model(inputs=model1.input, outputs=[model1.get_layer('global_average_pooling2d_17').output]) from test_data import load_test_data img_paths, test_data = load_test_data(input_size, preprocess_input) from svm_data import load_test_data_train train_data, train_label = load_test_data_train(input_size, preprocess_input) print( '################### intermediate_model train_going ##############################' ) from keras.preprocessing.image import ImageDataGenerator test_datagen1 = ImageDataGenerator(horizontal_flip=True, ) predictions_1 = [] for i in range(tta_steps): preds = intermediate_model1.predict_generator(test_datagen1.flow( train_data, batch_size=bs, shuffle=False),
from sklearn.ensemble import RandomForestClassifier from sklearn.metrics import accuracy_score import load_data as ld import text_tokenizing as tt import embedding_layer as emb_layer import test_data as td #Some constant values maxlen = 100 # We will cut reviews after 50 words max_words = 50000 # We will only consider the top 50,000 words in the dataset #labeling the data set texts,labels = ld.data_label() #sampling the text data x_train,y_train,x_val,y_val,word_index = tt.tokenize(texts,labels,maxlen,max_words) clf = RandomForestClassifier(n_estimators=25) clf.fit(x_train, y_train) x_test,y_test = td.load_test_data() clf_probs = clf.predict(x_test) print(accuracy_score(clf_probs,y_test))
def setUp(self): self.elements = segmentation.get_elements(test_data.load_test_data(), detailed=False)
def post_test_data(): """Loads test data""" load_test_data()