def wikipedia():
sentences, word2idx = get_wikipedia_data()
print("finished retrieving data")
print("vocab size:", len(word2idx), "number of sentences:", len(sentences))
rnn = SimpleRNN(20, 15, len(word2idx))
rnn.fit(sentences,
learning_rate=10e-5,
show_fig=True,
activation=T.nnet.relu)
def main():
# sentences, word2idx= get_sentences_with_word2idx_limit_vocab(n_vocab=2000)
sentences, word2idx = get_wikipedia_data(n_files=50, n_vocab=2000)
with open('w2v_word2idx.json', 'w') as f:
json.dump(word2idx, f)
V = len(word2idx)
model = Model(80, V, 10)
model.fit(sentences, learning_rate=10e-4, mu=0, epochs=7)
model.save('w2v_model.npz')
def train_wikipedia(we_file='word_embeddings.npy', w2i_file='wikipedia_word2idx.json', RecurrentUnit=GRU):
# there are 32 files
sentences, word2idx = get_wikipedia_data(n_files=1, n_vocab=2000)
print "finished retrieving data"
print "vocab size:", len(word2idx), "number of sentences:", len(sentences)
rnn = RNN(30, [30], len(word2idx))
rnn.fit(sentences, learning_rate=10e-6, epochs=10, show_fig=True, activation=T.nnet.relu)
np.save(we_file, rnn.We.get_value())
with open(w2i_file, 'w') as f:
json.dump(word2idx, f)
def train_wikipedia(we_file='word_embeddings.npy', w2i_file='wikipedia_word2idx.json', RecurrentUnit=GRU):
# there are 32 files
sentences, word2idx = get_wikipedia_data(n_files=1, n_vocab=2000)
print("finished retrieving data")
print("vocab size:", len(word2idx), "number of sentences:", len(sentences))
rnn = RNN(30, [30], len(word2idx))
rnn.fit(sentences, learning_rate=10e-6, epochs=10, show_fig=True, activation=T.nnet.relu)
np.save(we_file, rnn.We.get_value())
with open(w2i_file, 'w') as f:
json.dump(word2idx, f)
def train_wikipedia(we_file='lstm_word_embeddings.npy', w2i_file='lstm_wikipedia_word2idx.json'):
# there are 32 files
sentences, word2idx = get_wikipedia_data(n_files=100, n_vocab=2000)
print "finished retrieving data"
print "vocab size:", len(word2idx), "number of sentences:", len(sentences)
rnn = RNN(50, [50], len(word2idx))
# todo: next try increas LR
rnn.fit(sentences, learning_rate=10e-6, epochs=10, show_fig=True, activation=T.nnet.relu, RecurrentUnit=LSTM, normalize=False)
np.save(we_file, rnn.We.get_value())
with open(w2i_file, 'w') as f:
json.dump(word2idx, f)
def train_wikipedia(we_file='word_embeddings.npy',
w2i_file='wikipedia_word2idx.json',
RecurrentUnit=GRU):
sentences, word2idx = get_wikipedia_data(n_files=100, n_vocab=2000)
print('Finished retrieving data!')
print(
f'vocab size: {len(word2idx)}, number of sentences: {len(sentences)}')
rnn = RNN(50, [50], len(word2idx))
rnn.fit(sentences,
learning_rate=10e-6,
epochs=10,
show_fig=True,
activation=T.nnet.relu)
np.savez(we_file, rnn.We.get_value())
with open(w2i_file, 'w') as f:
json.dump(word2idx, f)
def train_wikipedia(we_file='lstm_word_embeddings.npy',
w2i_file='lstm_wikipedia_word2idx.json'):
# there are 32 files
sentences, word2idx = get_wikipedia_data(n_files=100, n_vocab=2000)
print "finished retrieving data"
print "vocab size:", len(word2idx), "number of sentences:", len(sentences)
rnn = RNN(50, [50], len(word2idx))
# todo: next try increas LR
rnn.fit(sentences,
learning_rate=10e-6,
epochs=10,
show_fig=True,
activation=T.nnet.relu,
RecurrentUnit=LSTM,
normalize=False)
np.save(we_file, rnn.We.get_value())
with open(w2i_file, 'w') as f:
json.dump(word2idx, f)
def train_wikipedia(we_file='word_embeddings.npy',
w2i_file='wikipedia_word2idx.json',
RecurrentUnit=GRU):
# there are 32 files
# note: you can pick between Wikipedia data and Brown corpus
# just comment one out, and uncomment the other!
sentences, word2idx = get_wikipedia_data(n_files=98, n_vocab=2000)
# sentences, word2idx = get_sentences_with_word2idx_limit_vocab()
print("finished retrieving data")
print("vocab size:", len(word2idx), "number of sentences:", len(sentences))
rnn = RNN(30, [30], len(word2idx))
rnn.fit(sentences,
learning_rate=1e-5,
epochs=10,
show_fig=True,
activation=T.nnet.relu)
np.save(we_file, rnn.We.get_value())
with open(w2i_file, 'w') as f:
json.dump(word2idx, f)
def main(we_file, w2i_file, n_files=50):
cc_matrix = "cc_matrix_%s.npy" % n_files
if os.path.exists(cc_matrix):
with open(w2i_file) as f:
word2idx = json.load(f)
sentences = []
else:
sentences, word2idx = get_wikipedia_data(n_files=n_files, n_vocab=2000)
with open(w2i_file, 'w') as f:
json.dump(word2idx, f)
V = len(word2idx)
model = Glove(80, V, 10)
model.fit(sentences=sentences,
cc_matrix=cc_matrix,
learning_rate=3 * 10e-5,
reg=0.01,
epoches=2000,
gd=True,
use_theano=False)
model.save(we_file)
def wikipedia():
sentences, word2idx = get_wikipedia_data()
print "finished retrieving data"
print "vocab size:", len(word2idx), "number of sentences:", len(sentences)
rnn = SimpleRNN(20, 15, len(word2idx))
rnn.fit(sentences, learning_rate=10e-5, show_fig=True, activation=T.nnet.relu)
def main(we_file, w2i_file, use_brown=True, n_files=100):
if use_brown:
cc_matrix = "cc_matrix_brown.npy"
else:
cc_matrix = "cc_matrix_%s.npy" % n_files
# hacky way of checking if we need to re-load the raw data or not
# remember, only the co-occurrence matrix is needed for training
if os.path.exists(cc_matrix):
with open(w2i_file) as f:
word2idx = json.load(f)
sentences = [] # dummy - we won't actually use it
else:
if use_brown:
keep_words = set([
'king',
'man',
'woman',
'france',
'paris',
'london',
'rome',
'italy',
'britain',
'england',
'french',
'english',
'japan',
'japanese',
'chinese',
'italian',
'australia',
'australian',
'december',
'november',
'june',
'january',
'february',
'march',
'april',
'may',
'july',
'august',
'september',
'october',
])
sentences, word2idx = get_sentences_with_word2idx_limit_vocab(
n_vocab=5000, keep_words=keep_words)
else:
sentences, word2idx = get_wikipedia_data(n_files=n_files,
n_vocab=2000)
with open(w2i_file, 'w') as f:
json.dump(word2idx, f)
V = len(word2idx)
model = Glove(100, V, 10)
# alternating least squares method
model.fit(sentences, cc_matrix=cc_matrix)
model.save(we_file)
from sklearn.decomposition import TruncatedSVD, PCA, KernelPCA
from sklearn.feature_extraction.text import TfidfTransformer
from datetime import datetime
import os
import sys
initial_dir = os.getcwd()
os.chdir('../course_repo/machine_learning_examples/rnn_class')
from util import get_wikipedia_data
from brown import get_sentences_with_word2idx_limit_vocab, get_sentences_with_word2idx
# os.chdir(initial_dir)
sentences, word2idx = get_wikipedia_data(n_files=3,
n_vocab=2000,
by_paragraph=True)
# build term document matrix:
V = len(word2idx)
N = len(sentences)
# create raw counts:
A = np.zeros((V, N))
j = 0
for sentence in sentences:
for i in sentence:
A[i, j] += 1
j += 1
def wikipedia():
sentences, word2idx = get_wikipedia_data()
