from tqdm import tqdm
from python.dataset import AutoComplete
from python.model import PureGRU
from python.summary import ContextAccuracy
tf.logging.set_verbosity(tf.logging.INFO)
train_dataset = AutoComplete(repeat=True, batch_size=64)
test_dataset = AutoComplete(dataset='test', repeat=False, batch_size=16)
model = PureGRU(train_dataset,
name='autocomplete_gru_600',
embedding_size=600,
verbose=True)
context_accuracy = ContextAccuracy()
for output_i, output in enumerate(
tqdm(model.predict(dataset=test_dataset),
total=test_dataset.observations)):
probabilities = output['probabilities']
predict_sorted = np.argsort(probabilities, axis=1)[:, ::-1]
source = test_dataset.decode_source(output['source'])
target = test_dataset.decode_target(output['target'])
predict = test_dataset.decode_target(predict_sorted)
context_accuracy.add(source, predict, target)
print(context_accuracy.summary())
train_dataset = AutoComplete(repeat=True)
test_dataset = AutoCompleteFixed(
"context the formal study of grammar is an important part of education "
"from a young age through advanced learning though the rules taught in "
"schools are not a grammar in the sense most linguists use",
batch_size=1,
use_offsets=True)
model = PureGRU(train_dataset,
name='autocomplete_gru_600',
embedding_size=600,
verbose=True)
data = []
print(f'sequence:')
for output_i, output in enumerate(model.predict(dataset=test_dataset)):
probabilities = output['probabilities']
predict_index_sorted = np.argsort(probabilities, axis=1)[:, ::-1]
predict_value_sorted = np.sort(probabilities, axis=1)[:, ::-1]
source = test_dataset.decode_source(output['source'])
target = test_dataset.decode_target(output['target'])
predict = test_dataset.decode_target(predict_index_sorted)
connectivity = output['connectivity']
char = source[output_i]
words_sorted = predict[output_i]
props_sorted = predict_value_sorted[output_i]
target_word = target[output_i]
print(f' {char} -> {words_sorted[0]}, {words_sorted[1]},'