def test_dnn_accuracy(snli, dnn):
classifier = LogisticRegression(n_features=2 * dnn.dim, n_labels=3)
model = CompositeModel(snli, dnn, classifier)
model.train(iters=2, bsize=1, rate=0.01, log_interval=2)
assert len(model.acc) == 2
assert 0 < model.acc[0] < 0.65
def test_bow_accuracy(snli):
dim = 50
encoder = BagOfWords(dim=dim, vocab=snli.vocab)
classifier = LogisticRegression(n_features=2 * dim, n_labels=3)
model = CompositeModel(snli, encoder, classifier)
model.train(iters=2, bsize=1, rate=0.01, log_interval=2)
assert len(model.acc) == 2
assert 0 < model.acc[0] < 0.65
def test_predict(snli, dnn, rnn):
labels = {'entailment', 'contradiction', 'neutral'}
classifier = LogisticRegression(n_features=2 * dnn.dim, n_labels=3)
model = CompositeModel(snli, dnn, classifier)
model.train(iters=2, bsize=1, rate=0.01, log_interval=2)
label = model.predict(s1='The boy ran.', s2='The boy moved.')
assert label in labels
model = CompositeModel(snli, rnn, classifier)
model.train(iters=2, bsize=1, rate=0.01, log_interval=2)
label = model.predict(s1='The boy ran.', s2='The boy moved.')
assert label in labels
import pickle
import time
from pysem.corpora import SNLI
from pysem.utils.ml import MultiLayerPerceptron
from pysem.utils.snli import CompositeModel, BagOfWords
snli = SNLI('/Users/peterblouw/corpora/snli_1.0/')
snli.load_vocab('snli_words.pickle')
dim = 300
pretrained = 'pretrained_snli_embeddings.pickle'
encoder = BagOfWords(dim=dim, vocab=snli.vocab, pretrained=pretrained)
classifier = MultiLayerPerceptron(di=2 * dim, dh=dim, do=3)
start_time = time.time()
model = CompositeModel(snli, encoder, classifier)
model.train(iters=50, bsize=100, rate=0.1, log_interval=10, schedule=12000)
print('Test: ', model.rnn_accuracy(model.test_data))
print('Train: ', model.rnn_accuracy(model.train_data))
print('Dev: ', model.rnn_accuracy(model.dev_data))
print('Total runtime: ', ((time.time() - start_time) / 3600.0))
with open('bow_model', 'wb') as pfile:
pickle.dump(model, pfile)
def test_average(snli):
array = list(range(100))
assert len(CompositeModel.average(array, [], 10)) == 10