for (word, i) in word_index.items():
try:
embedding_vector = word_vec.get_vector(word)
# words not found in embedding index will be all-zeros.
embedding_matrix[i] = embedding_vector
except:
# print('word not in word_vec')
pass
return embedding_matrix
if __name__ == '__main__':
train = pd.read_csv(PATH + 'train.txt', sep='\t')
val = pd.read_csv(PATH + 'valid.txt', sep='\t')
test = pd.read_csv(PATH + 'test.txt', sep='\t')
train_id, train_label, val_id, val_label, test_id, test_label, vocab, word_index = text_encode(
train, val, test, type='seq')
word_vec = models.KeyedVectors.load_word2vec_format(WORD_VEC_PATH +
WORD_VEC_FILE,
binary=True)
embedding_matrix = get_embedding_matrix(word_vec, word_index)
print(embedding_matrix.shape)
model = Sequential()
model.add(
Embedding(len(vocab) + 1,
DW,
weights=[embedding_matrix],
trainable=False))
model.add(SimpleRNN(DH, dropout=0.2, recurrent_dropout=0.1))
import pandas as pd
from keras.layers import Embedding, SimpleRNN, Dense
from keras.models import Sequential
from zzz.chapter09.knock80 import text_encode
PATH = '/Users/zz_zhang/勉強会/100本ノック/100knock2020/zzz/chapter06/'
DW = 300
DH = 50
MAX_LEN = 20
if __name__ == '__main__':
train = pd.read_csv(PATH + 'train.txt', sep='\t')
val = pd.read_csv(PATH + 'valid.txt', sep='\t')
test = pd.read_csv(PATH + 'test.txt', sep='\t')
train_onehot, train_label, val_onehot, val_label, test_onehot, test_label, vocab, word_index = text_encode(
train, val, test, type='onehot')
print(train_onehot[:4])
model = Sequential()
model.add(Embedding(len(vocab) + 1, DW))
model.add(SimpleRNN(DH, dropout=0.2, recurrent_dropout=0.1))
model.add(Dense(4, activation='softmax'))
model.summary()
model.compile(
loss='categorical_crossentropy',
optimizer='sgd',
metrics=["accuracy"],