emotion_sentences = load_pickle(config.emojis_path)
# Load infersent embeddings if necessary
infersent_sentences = None
if config.infersent:
print('Loading infersent sentence embeddings...')
infersent_sentences = load_pickle(config.infersent_path)
embedding_size = infersent_sentences[0][0].shape[0]
config.infersent_output_size = embedding_size
data_loader = get_loader(
sentences=load_pickle(config.sentences_path),
conversation_length=load_pickle(config.conversation_length_path),
sentence_length=load_pickle(config.sentence_length_path),
vocab=vocab,
batch_size=config.batch_size,
emojis=emotion_sentences,
infersent=infersent_sentences)
if config.model in VariationalModels:
solver = VariationalSolver(config,
None,
data_loader,
vocab=vocab,
is_train=False)
solver.build()
solver.importance_sample()
else:
solver = Solver(config, None, data_loader, vocab=vocab, is_train=False)
solver.build()
solver.test()
def load_pickle(path):
with open(path, 'rb') as f:
return pickle.load(f)
if __name__ == '__main__':
config = get_config(mode='test')
print('Loading Vocabulary...')
vocab = Vocab()
vocab.load(config.word2id_path, config.id2word_path)
print(f'Vocabulary size: {vocab.vocab_size}')
config.vocab_size = vocab.vocab_size
data_loader = get_loader(
sentences=load_pickle(config.sentences_path),
conversation_length=load_pickle(config.conversation_length_path),
sentence_length=load_pickle(config.sentence_length_path),
vocab=vocab,
batch_size=config.batch_size,
shuffle=False)
if config.model in VariationalModels:
solver = VariationalSolver(config, None, data_loader, vocab=vocab, is_train=False)
else:
solver = Solver(config, None, data_loader, vocab=vocab, is_train=False)
solver.build()
solver.embedding_metric()
return pickle.load(f)
if __name__ == '__main__':
config = get_config(mode='valid')
print('Loading Vocabulary...')
vocab = Vocab()
vocab.load(config.word2id_path, config.id2word_path)
print(f'Vocabulary size: {vocab.vocab_size}')
config.vocab_size = vocab.vocab_size
for i in range(50):
check_file = str(i+1) + '.pkl'
config.checkpoint = os.path.join(config.eval_checkpoint, check_file)
data_loader = get_loader(
sentences=load_pickle(config.sentences_path),
conversation_length=load_pickle(config.conversation_length_path),
sentence_length=load_pickle(config.sentence_length_path),
vocab=vocab,
batch_size=config.batch_size)
if config.model in VariationalModels:
solver = VariationalSolver(config, None, data_loader, vocab=vocab, is_train=False)
solver.build()
solver.evaluate()
else:
solver = Solver(config, None, data_loader, vocab=vocab, is_train=False)
solver.build()
solver.test()
vocab=vocab_freq,
batch_size=test_freq_config.batch_size,
shuffle=False)
rare_data_loader = get_loader(
sentences=load_pickle(test_rare_config.sentences_path),
conversation_length=load_pickle(
test_rare_config.conversation_length_path),
sentence_length=load_pickle(test_rare_config.sentence_length_path),
vocab=vocab_rare,
batch_size=test_rare_config.batch_size,
shuffle=False)
if test_freq_config.model in VariationalModels:
solver_rare = VariationalSolver(test_rare_config,
None,
rare_data_loader,
rare_data_loader,
vocab=vocab_rare,
is_train=False)
else:
solver_rare = Solver(test_rare_config,
None,
rare_data_loader,
rare_data_loader,
vocab=vocab_rare,
is_train=False)
solver_rare.build()
solver_rare.embedding_metric()
if config.infersent:
print('Loading infersent sentence embeddings...')
infersent_sentences = load_pickle(config.infersent_path)
embedding_size = infersent_sentences[0][0].shape[0]
config.infersent_output_size = embedding_size
data_loader = get_loader(
sentences=load_pickle(config.sentences_path),
conversation_length=load_pickle(config.conversation_length_path),
sentence_length=load_pickle(config.sentence_length_path),
vocab=vocab,
batch_size=config.batch_size,
emojis=emotion_sentences,
infersent=infersent_sentences)
if config.model in VariationalModels:
solver = VariationalSolver(config,
None,
data_loader,
vocab=vocab,
is_train=False)
else:
solver = Solver(config, None, data_loader, vocab=vocab, is_train=False)
solver.build()
solver.interact(max_sentence_length=kwargs.max_sentence_length,
max_conversation_length=kwargs.max_conversation_length,
sample_by=kwargs.sample_by,
debug=kwargs.debug,
print_history=True)
# Load infersent embeddings if necessary
infersent_sentences = None
if config.infersent or config.calc_novel_embedding:
print('Loading infersent sentence embeddings...')
infersent_sentences = load_pickle(config.infersent_path)
embedding_size = infersent_sentences[0][0].shape[0]
config.infersent_output_size = embedding_size
data_loader = get_loader(
sentences=load_pickle(config.sentences_path),
conversation_length=load_pickle(config.conversation_length_path),
sentence_length=load_pickle(config.sentence_length_path),
vocab=vocab,
batch_size=config.batch_size,
shuffle=False,
emojis=emotion_sentences,
infersent=infersent_sentences)
if config.model in VariationalModels:
solver = VariationalSolver(config,
None,
data_loader,
vocab=vocab,
is_train=False)
else:
solver = Solver(config, None, data_loader, vocab=vocab, is_train=False)
solver.build()
solver.novel_metrics()
def load_pickle(path):
with open(path, 'rb') as f:
return pickle.load(f)
if __name__ == '__main__':
config = get_config(mode='test')
print('Loading Vocabulary...')
vocab = Vocab()
vocab.load(config.word2id_path, config.id2word_path)
print(f'Vocabulary size: {vocab.vocab_size}')
config.vocab_size = vocab.vocab_size
data_loader = get_loader(
sentences=load_pickle(config.sentences_path),
conversation_length=load_pickle(config.conversation_length_path),
sentence_length=load_pickle(config.sentence_length_path),
vocab=vocab,
batch_size=config.batch_size)
if config.model in VariationalModels:
solver = VariationalSolver(config,
None,
data_loader,
vocab=vocab,
is_train=False)
solver.build()
solver.generate_file()
conversation_length=load_pickle(config.conversation_length_path),
sentence_length=load_pickle(config.sentence_length_path),
vocab=vocab,
batch_size=1,
shuffle=False,
)
dt_string = datetime.now().strftime("%d-%m-%Y_%H:%M:%S")
probs_dict = SqliteDict(
config.save_path + "/eval_{}.pkl".format(dt_string),
encode=my_encode,
decode=my_decode,
journal_mode="OFF",
autocommit=True,
)
if config.model in VariationalModels:
solver = VariationalSolver(
config, None, data_loader, vocab=vocab, is_train=False
)
solver.build()
batch_probs_history = solver.evaluate_convai()
else:
solver = Solver(config, None, data_loader, vocab=vocab, is_train=False)
solver.build()
batch_probs_history = solver.evaluate_convai()
for idx, probs_item in enumerate(batch_probs_history):
probs_dict[str(idx)] = probs_item
sentences=load_pickle(config.sentences_path),
conversation_length=load_pickle(config.conversation_length_path),
sentence_length=load_pickle(config.sentence_length_path),
score=load_pickle(config.score_path),
vocab=vocab,
batch_size=config.batch_size,
shuffle=False)
else:
data_loader = get_loader(
sentences=load_pickle(config.sentences_path),
conversation_length=load_pickle(config.conversation_length_path),
sentence_length=load_pickle(config.sentence_length_path),
vocab=vocab,
batch_size=config.batch_size,
shuffle=False)
if config.model in VariationalModels:
if config.model == "ADEM":
print("train line:69 config.model", config.model)
solver = AdemSolver(config, None, data_loader, vocab=vocab, is_train=False)
else:
solver = VariationalSolver(config, None, data_loader, vocab=vocab, is_train=False)
solver.build()
solver.importance_sample()
else:
solver = Solver(config, None, data_loader, vocab=vocab, is_train=False)
solver.build()
solver.test()
batch_size=test_freq_config.batch_size,
shuffle=False)
rare_data_loader = get_loader(
sentences=load_pickle(test_rare_config.sentences_path),
conversation_length=load_pickle(
test_rare_config.conversation_length_path),
sentence_length=load_pickle(test_rare_config.sentence_length_path),
vocab=vocab_rare,
batch_size=test_rare_config.batch_size,
shuffle=False)
if test_freq_config.model in VariationalModels:
solver_freq = VariationalSolver(test_freq_config,
None,
freq_data_loader,
rare_data_loader,
vocab=vocab_freq,
is_train=False)
solver_freq.build()
solver_freq.importance_sample()
else:
solver_freq = Solver(test_freq_config,
None,
freq_data_loader,
rare_data_loader,
vocab=vocab_freq,
is_train=False)
solver_freq.build()
freq_perplexity, freq_f1 = solver_freq.test()
print('Loading rare Vocabulary...')