def main(option):
if option == 1 or option == 3:
model = unigram.Unigram(read_data())
print('-------------- Baseline -----------------')
print(
f'Accuracy of Unigram on data/dev = {predict(model, filename_dev)*100:.2f} %\n'
)
if option == 2 or option == 3:
five_gram_model = five_gram.Five_gram(read_data())
print('-------- n-gram Language Model ----------')
print(
f'Accuracy of 5-Gram on data/dev = {predict(five_gram_model, filename_dev)*100:.2f} %'
)
print(
f'Accuracy of 5-Gram on data/test = {predict(five_gram_model, filename_test)*100:.2f} %'
)
import argparse
import unigram
import fivegram
import sevengram_smoothed
import operator
unigram_model = unigram.Unigram()
fivegram_model = fivegram.Fivegram()
sevengram_smoothed_model = sevengram_smoothed.Sevengram_Smoothed()
def predict(filename):
"""Predict the text file on the model."""
total = 0.
correct = 0
for line in open(filename):
model.start()
for w in line.rstrip('\n'):
if model.predict() == w:
correct += 1
model.read(w)
total += 1
print("Accuracy = " + str(correct/total*100) + "%")
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('--model')
parser.add_argument('--train', required=True)
parser.add_argument('--predict', required=True)
args = parser.parse_args()
if args.model == 'unigram':
def main():
model = unigram.Unigram(read_data())
print('-------------- Baseline -----------------')
print(f'Accuracy of Unigram on data/dev = {predict(model)*100:.2f} %')
print('\a', end='')
else:
erase()
print(f'\x1b[1D', end='', flush=True)
chars.pop()
states.pop()
else:
print(c, end='', flush=True)
chars.append(c)
states.append(lm.read(states[-1], c))
return ''.join(chars)
if __name__ == "__main__":
import argparse
import unigram
parser = argparse.ArgumentParser()
parser.add_argument(dest='train')
args = parser.parse_args()
data = [list(line.rstrip('\n')) for line in open(args.train)]
##### Replace this line with an instantiation of your model #####
lm = unigram.Unigram(data)
while True:
try:
line = getline('> ')
except EOFError:
break
parser.add_argument(dest='train')
args = parser.parse_args()
train = []
dev= []
#Import Train
with open(args.train, "r") as f:
train = [list(line.rstrip()) for line in f]
#Import Dev
with open("data/dev", "r") as f:
dev = [list(line.rstrip("\n")) + ['<EOS>'] for line in f]
#Train Model
m = unigram.Unigram(train)
q = m.start()
total = 0
correct = 0
#Read Dev Data into the model.
for line in dev:
for index, a in enumerate(line):
predicted_symbol = m.best(q)
if index != len(line) - 1:
gold_symbol = line[index + 1]
if predicted_symbol == gold_symbol:
correct += 1
total+=1