def greek_to_token(wrd):
input = " ".join(list(clean(basify(wrd))))
sequences = tok.texts_to_sequences([input])
sequences_matrix = sequence.pad_sequences(sequences, maxlen=20)
out = to_categorical(sequences_matrix[0],
num_classes=alpha).reshape(1, 20, alpha)
return out
def get_greek(self):
wd = self.word_string.split("/")[0]
wd = wd.lower()
wd = clean(wd)
return wd
model.compile(loss='categorical_crossentropy',optimizer=RMSprop(),metrics=['categorical_accuracy'])
model.fit(to_categorical(sequences_matrix),encoded,batch_size=128,epochs=5,validation_split=0.1)
model.save('/home/q078011/external/greek_dev/pos_mini.h5')
st = '???t? d? ???? ?????e?, ??a p?????? t? ????? ?p? t?? ?????? d?? t?? p??f?t??, ?????t??'
to_pos = clean(basify(st))
tnt_tot = tnt.TnT()
tnt_tot.train([list(zip(list(all_str),list(y)))])
import pickle
with open('/home/q078011/external/greek_dev/dict_letters.pkl', 'wb') as f:
pickle.dump(dict_g, f)
with open('/home/q078011/external/greek_dev/tokenizer.pkl', 'wb') as g:
pickle.dump(tok, g)