def test_target_extraction_fit(self, test_data: bool):
model = AllenNLPModel('TE', self.CONFIG_FILE, 'target-tagger')
assert model.model is None
train_data = TargetTextCollection.load_json(
self.TARGET_EXTRACTION_TRAIN_DATA)
val_data = TargetTextCollection.load_json(
self.TARGET_EXTRACTION_TRAIN_DATA)
tokens_in_vocab = [
'at', 'case', 'was', 'the', 'day', 'great', 'cover', 'office',
'another', 'and', 'rubbish', 'laptop', '@@PADDING@@', '@@UNKNOWN@@'
]
if test_data:
tokens_in_vocab = tokens_in_vocab + ['better']
test_data = TargetTextCollection.load_json(
self.TARGET_EXTRACTION_TEST_DATA)
model.fit(train_data, val_data, test_data)
else:
model.fit(train_data, val_data)
token_index = model.model.vocab.get_token_to_index_vocabulary('tokens')
assert len(token_index) == len(tokens_in_vocab)
for token in tokens_in_vocab:
assert token in token_index
# Check attributes have changed.
assert model.model is not None
assert isinstance(model.model, Model)
# Check that it will save to a directory of our choosing
with tempfile.TemporaryDirectory() as save_dir:
saved_model_fp = Path(save_dir, 'model.tar.gz')
assert not saved_model_fp.exists()
model = AllenNLPModel('TE',
self.CONFIG_FILE,
'target-tagger',
save_dir=save_dir)
model.fit(train_data, val_data)
assert saved_model_fp.exists()
for error in returned_errors:
error_id = error['text_id']
del dataset[error_id]
returned_errors = dataset.sequence_labels(return_errors=True)
if returned_errors:
raise ValueError('Sequence label errors are still persisting')
sizes.append(len(dataset))
dataset: TargetTextCollection
target_sizes.append(dataset.number_targets())
print(
f'Lengths Train: {sizes[0]}, Validation: {sizes[1]}, Test: {sizes[2]}'
)
print(f'Number of targets, Train: {target_sizes[0]}, Validation: '
f'{target_sizes[1]}, Test: {target_sizes[2]}')
print('Fitting model')
model.fit(train_data, val_data, test_data)
print('Finished fitting model\nNow Evaluating model:')
else:
test_data.tokenize(spacy_tokenizer())
device = -1
if args.cuda:
device = 0
model.load(cuda_device=device)
print('Finished loading model\nNow Evaluating model:')
for data in test_data.values():
data['tokens'] = data['tokenized_text']
test_iter = iter(test_data.values())
for test_pred in model.predict_sequences(test_data.values(),
batch_size=args.batch_size):
relevant_test = next(test_iter)