def init(data_dir, model_file):
""" Initialize web app """
global memn2n, test_story, test_questions, test_qstory
# Try to load model
memn2n = MemN2N(data_dir, model_file)
memn2n.load_model()
# SV create word vectors for all the dictionary words
print("Embedding dictionary words")
for word in memn2n.general_config.dictionary.keys():
# SV assign vector to each word in the dictionary
memn2n.dict_vectors[word] = memn2n.word_model[word]
# Read test data
# test_data_path = glob.glob('%s/qa*_*_test.txt' % memn2n.data_dir)
test_data_path = glob.glob(memn2n.data_path.format('test'))
# load different dataset with samples, e.g. real table data
# test_data_path = glob.glob('./data/table_data_{}.txt'.format('test'))
print("Reading test data from %s ..." % test_data_path)
# test_story, test_questions, test_qstory = \
# parse_babi_task(test_data_path, memn2n.general_config.dictionary, False)
# update dictionary with new words
test_story, test_questions, test_qstory = \
memn2n.parse_babi_task(test_data_path, False)
# SV expand reversed_dict with test data
print len(memn2n.general_config.dictionary)
# Get reversed dictionary mapping index to word
memn2n.reversed_dict = dict((ix, w) for w, ix in memn2n.general_config.dictionary.items())
def init(data_dir, model_file):
""" Initialize web app """
global memn2n, test_story, test_questions, test_qstory
# Try to load model
memn2n = MemN2N(data_dir, model_file)
memn2n.load_model()
# Read test data
print("Reading test data from %s ..." % memn2n.data_dir)
test_data_path = glob.glob('%s/qa*_*_test.txt' % memn2n.data_dir)
test_story, test_questions, test_qstory = \
parse_babi_task(test_data_path, memn2n.general_config.dictionary, False)
def predict_d():
word_idx = pickle.load(open(model_dir + 'word_idx.pkl', 'rb'))
import glob
import numpy as np
from demo.qa import MemN2N
from util import parse_babi_task
""" Initialize app """
global memn2n, test_story, test_questions, test_qstory
model_file = 'trained_model/memn2n_model.pklz'
# Try to load model
memn2n = MemN2N(data_dir, model_file)
memn2n.load_model()
test_story, test_questions, test_qstory = None, None, None
# Read test data
print("Reading test data from %s ..." % memn2n.data_dir)
test_data_path = glob.glob('%s/qa*_*_test.txt' % memn2n.data_dir)
test_story, test_questions, test_qstory = \
parse_babi_task(test_data_path, memn2n.general_config.dictionary, False)
question_idx = np.random.randint(test_questions.shape[1])
story_idx = test_questions[0, question_idx]
last_sentence_idx = test_questions[1, question_idx]
story_txt, question_txt, correct_answer = memn2n.get_story_texts(
test_story, test_questions, test_qstory, question_idx, story_idx,
last_sentence_idx)
# Format text
story_txt = "\n".join(story_txt)
question_txt += "?"
print('Case study:')
print('Story:', story_txt, '\nQuestion:', question_txt, '\nAnswer:',
correct_answer)
model = tf.keras.models.load_model(model_dir)
predict_answer = model.predict([
vectorize(story_txt, word_idx, 552),
vectorize(question_txt, word_idx, 5)
])
print(list(word_idx.keys())[np.argmax(predict_answer, axis=1)[0] - 1])
train_path = os.path.join(args.data_dir2, 'train')
if not os.path.exists(train_path):
print("'%s' does not exist." % train_path)
sys.exit(1)
test_path = os.path.join(args.data_dir2, 'test')
if not os.path.exists(test_path):
print("'%s' does not exist." % test_path)
sys.exit(1)
args.data_dir = train_path, test_path
if type(args.data_dir) is tuple:
print("Using data from {} and {}".format(args.data_dir[0],
args.data_dir[1]))
else:
print("Using data from %s" % args.data_dir)
if args.test or args.all_tests:
m = MemN2N(args.data_dir, args.model_file)
m.load_model()
if args.all_tests:
run_all_tests(args.data_dir, m)
else:
run_test(args.data_dir, args.task, m)
else:
if args.all_tasks:
run_all_tasks(data_dir)
elif args.joint_tasks:
run_joint_tasks(data_dir)
else:
run_task(data_dir, task_id=args.task)
], [sg.Button('模型一'), sg.Button('模型二')]]
""" Initialize app """
global memn2n, test_story, test_questions, test_qstory
data_dir = 'data/tasks_1-20_v1-2/en'
model_file = 'trained_model/memn2n_model.pklz'
memn2n = None
test_story, test_questions, test_qstory = None, None, None
if __name__ == '__main__':
# Create the Window
window = sg.Window('babi数据集算法演示', layout)
# Event Loop to process "events" and get the "values" of the inputs
# Try to load model
memn2n = MemN2N(data_dir, model_file)
memn2n.load_model()
# Read test data
print("Reading test data from %s ..." % memn2n.data_dir)
test_data_path = glob.glob('%s/qa*_*_test.txt' % memn2n.data_dir)
test_story, test_questions, test_qstory = \
parse_babi_task(test_data_path, memn2n.general_config.dictionary, False)
while True:
event, values = window.read()
if event == sg.WIN_CLOSED:
break
if event == 'getstory':
question_idx = np.random.randint(test_questions.shape[1])
path_to_model_output_dir = './ExperimentA'
'''
Train, test, and save model for three different learning rates each with/without linear start
and 1,2,3 memory modules (hops).
'''
starting_learning_rate = [.005, .1, .01]
linear_start = [True, False]
hops = [1, 2, 3]
config_swtiches = [None, None, None]
for lr in starting_learning_rate:
config_swtiches[0] = lr
for opt in linear_start:
config_swtiches[1] = opt
for num in hops:
config_swtiches[2] = num
model_file_name = 'lr{}_linearstart{}_hops{}'.format(
*config_swtiches)
write_model_to = os.path.join(path_to_model_output_dir,
model_file_name)
message = 'Training and testing model for {}'.format(
model_file_name)
print(message)
model = MemN2N(path_to_babi_data, write_model_to, config_swtiches)
model.train_and_test()
print('\n\n')
from demo.qa import MemN2N
import os
import random
# Default location for data is location on the HPCC
#path_to_babi_data = '/mnt/research/CSE842/FinalProject/Benchmarks/bAbI'
path_to_babi_data = '/home/bbaker/bAbI'
path_to_model_output_dir = './ExperimentA'
'''
Train, test, and save model from original repo configuration
'''
for i in range(10):
seed = random.randint(1, 10000000)
model_file_name = 'default_config_model{}_seed'.format(i)
write_model_to = os.path.join(path_to_model_output_dir, model_file_name)
message = 'Training and testing model for {}'.format(model_file_name)
print(message)
model = MemN2N(path_to_babi_data, write_model_to)
model.train_and_test(seed)
print('\n\n')