def generate_sonnet_rhyme(num_states):
hmm = train.load(num_states, is_reversed=True)
text = open('data/shakespeare.txt').read()
obs, vocab, inv_vocab = preprocess.get_observations(text)
rhyme_dict = preprocess.build_rhyme_dict(text, vocab)
lengths = preprocess.get_lengths(obs)
punctuation = preprocess.get_punctuation(text)
samples = [None] * 14
for quatrain in range(3):
for line in range(2):
couplet = random.sample(random.choice(rhyme_dict), 2)
length = np.random.choice(list(lengths.keys()),
p=list(lengths.values()))
samples[quatrain * 4 + line] = generate_emission_seeded(
hmm, length, couplet[0])
samples[quatrain * 4 + line + 2] = generate_emission_seeded(
hmm, length, couplet[1])
couplet = random.sample(random.choice(rhyme_dict), 2)
samples[12] = generate_emission_seeded(hmm, length, couplet[0])
samples[13] = generate_emission_seeded(hmm, length, couplet[1])
for i in range(len(samples)):
samples[i].reverse()
return format_sonnet(samples, inv_vocab, punctuation)
def predict_BERT(para, feature="radical", use_bert=True):
para['tag_num'] = len(tags)
para['rad_vocab_size'] = len(rad2id.keys()) + 1
para['radical_vocab_size'] = len(radical2id.keys()) + 1
para['pinyin_vocab_size'] = len(pinyin2id.keys()) + 1
para["word_num"] = len(word2id.keys()) + 1
para["fea_embed"] = None
bool_x_test = creat_bool_x(test_x)
if "img" in feature:
para["img_embed_weight"] = img_embed
para["is_trainable"] = False
model = ModelLib.FGN(para, feature=feature, use_bert=use_bert)
# else:
# model = ModelLib.NORMAL_MODEL(para,feature=feature)
model.load_weights(filepath=para["model_path"])
if feature == "":
pred_y = model.predict([x1_test, x2_test], batch_size=64, verbose=1)
elif feature == "radical":
pred_y = model.predict([x1_test, x2_test, radical_train],
batch_size=64,
verbose=1)
elif feature == "pinyin":
pred_y = model.predict([x1_test, x2_test, pinyin_train],
batch_size=64,
verbose=1)
elif feature == "img":
pred_y = model.predict([x1_test, x2_test, test_x, bool_x_test],
batch_size=64,
verbose=1)
elif feature == "img&radical":
pred_y = model.predict(
[x1_test, x2_test, test_x, bool_x_test, radical_test],
batch_size=64,
verbose=1)
lengths = get_lengths(x1_test)
tag_pred_y = []
tag_val_y = []
for i, y in enumerate(pred_y):
y = [numpy.argmax(dim) for dim in y]
print(lengths[i])
p_y = y[:lengths[i]]
print(p_y)
v_y = y_test[i][:lengths[i]].flatten()
print(v_y)
p_y = [tags[dim] for dim in p_y]
v_y = [tags[dim] for dim in v_y]
tag_pred_y.append(p_y)
tag_val_y.append(v_y)
return tag_pred_y, tag_val_y
def generate_sonnet(num_states):
hmm = train.load(num_states)
text = open('data/shakespeare.txt').read()
obs, vocab, inv_vocab = preprocess.get_observations(text)
lengths = preprocess.get_lengths(obs)
punctuation = preprocess.get_punctuation(text)
samples = []
for _ in range(14):
length = np.random.choice(list(lengths.keys()),
p=list(lengths.values()))
samples.append(generate_emission(hmm, length))
return format_sonnet(samples, inv_vocab, punctuation)
def result_proess(pred_y, x1_test, tags):
lengths = get_lengths(x1_test)
tag_pred_y = []
tag_val_y = []
for i, y in enumerate(pred_y):
y = [numpy.argmax(dim) for dim in y]
# print(lengths[i])
p_y = y[:lengths[i]]
# print(p_y)
v_y = y_test[i][:lengths[i]].flatten()
# print(v_y)
p_y = [tags[dim] for dim in p_y]
v_y = [tags[dim] for dim in v_y]
tag_pred_y.append(p_y)
tag_val_y.append(v_y)
return tag_pred_y, tag_val_y
def predict_bert(para):
para['tag_num'] = len(tags)
model = ModelLib.BERT_MODEL(para)
model.load_weights(filepath=para["model_path"])
bert_val =load_path_bert(para["test_path"],sep=para["sep"])
lengths = get_lengths(val_x)
pred_y = model.predict(bert_val)
tag_pred_y = []
tag_val_y = []
for i, y in enumerate(pred_y):
y = [numpy.argmax(dim) for dim in y]
print(lengths[i])
p_y = y[:lengths[i]]
print(p_y)
v_y = val_y[i][:lengths[i]].flatten()
print(v_y)
p_y = [tags[dim] for dim in p_y]
v_y = [tags[dim] for dim in v_y]
tag_pred_y.append(p_y)
tag_val_y.append(v_y)
return tag_pred_y,tag_val_y