Skip to content

GitLab

Unverified Commit 8e67573a authored by Patrick von Platen's avatar Patrick von Platen Committed by GitHub
Browse files

[EncoderDecoder Tests] Improve tests (#4046) 



* Hoist bert model tester for patric

* indent

* make tests work

* Update tests/test_modeling_bert.py
Co-authored-by: default avatarJulien Chaumond <chaumond@gmail.com>
Co-authored-by: default avatarsshleifer <sshleifer@gmail.com>
Co-authored-by: default avatarJulien Chaumond <chaumond@gmail.com>
parent 6af3306a
Hide whitespace changes
Inline Side-by-side
Showing with 362 additions and 368 deletions
tests/test_modeling_bert.py
View file @ 8e67573a
...@@ -38,159 +38,93 @@ if is_torch_available(): ...@@ -38,159 +38,93 @@ if is_torch_available():
from transformers.modeling_bert import BERT_PRETRAINED_MODEL_ARCHIVE_MAP from transformers.modeling_bert import BERT_PRETRAINED_MODEL_ARCHIVE_MAP
@require_torch class BertModelTester:
class BertModelTest(ModelTesterMixin, unittest.TestCase): def __init__(
self,
parent,
batch_size=13,
seq_length=7,
is_training=True,
use_input_mask=True,
use_token_type_ids=True,
use_labels=True,
vocab_size=99,
hidden_size=32,
num_hidden_layers=5,
num_attention_heads=4,
intermediate_size=37,
hidden_act="gelu",
hidden_dropout_prob=0.1,
attention_probs_dropout_prob=0.1,
max_position_embeddings=512,
type_vocab_size=16,
type_sequence_label_size=2,
initializer_range=0.02,
num_labels=3,
num_choices=4,
scope=None,
):
self.parent = parent
self.batch_size = batch_size
self.seq_length = seq_length
self.is_training = is_training
self.use_input_mask = use_input_mask
self.use_token_type_ids = use_token_type_ids
self.use_labels = use_labels
self.vocab_size = vocab_size
self.hidden_size = hidden_size
self.num_hidden_layers = num_hidden_layers
self.num_attention_heads = num_attention_heads
self.intermediate_size = intermediate_size
self.hidden_act = hidden_act
self.hidden_dropout_prob = hidden_dropout_prob
self.attention_probs_dropout_prob = attention_probs_dropout_prob
self.max_position_embeddings = max_position_embeddings
self.type_vocab_size = type_vocab_size
self.type_sequence_label_size = type_sequence_label_size
self.initializer_range = initializer_range
self.num_labels = num_labels
self.num_choices = num_choices
self.scope = scope
def prepare_config_and_inputs(self):
input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size)
all_model_classes = ( input_mask = None
( if self.use_input_mask:
BertModel, input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
BertForMaskedLM,
BertForNextSentencePrediction, token_type_ids = None
BertForPreTraining, if self.use_token_type_ids:
BertForQuestionAnswering, token_type_ids = ids_tensor([self.batch_size, self.seq_length], self.type_vocab_size)
BertForSequenceClassification,
BertForTokenClassification, sequence_labels = None
token_labels = None
choice_labels = None
if self.use_labels:
sequence_labels = ids_tensor([self.batch_size], self.type_sequence_label_size)
token_labels = ids_tensor([self.batch_size, self.seq_length], self.num_labels)
choice_labels = ids_tensor([self.batch_size], self.num_choices)
config = BertConfig(
vocab_size=self.vocab_size,
hidden_size=self.hidden_size,
num_hidden_layers=self.num_hidden_layers,
num_attention_heads=self.num_attention_heads,
intermediate_size=self.intermediate_size,
hidden_act=self.hidden_act,
hidden_dropout_prob=self.hidden_dropout_prob,
attention_probs_dropout_prob=self.attention_probs_dropout_prob,
max_position_embeddings=self.max_position_embeddings,
type_vocab_size=self.type_vocab_size,
is_decoder=False,
initializer_range=self.initializer_range,
) )
if is_torch_available()
else ()
)
class BertModelTester(object): return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def __init__(
self, def prepare_config_and_inputs_for_decoder(self):
parent, (
batch_size=13,
seq_length=7,
is_training=True,
use_input_mask=True,
use_token_type_ids=True,
use_labels=True,
vocab_size=99,
hidden_size=32,
num_hidden_layers=5,
num_attention_heads=4,
intermediate_size=37,
hidden_act="gelu",
hidden_dropout_prob=0.1,
attention_probs_dropout_prob=0.1,
max_position_embeddings=512,
type_vocab_size=16,
type_sequence_label_size=2,
initializer_range=0.02,
num_labels=3,
num_choices=4,
scope=None,
):
self.parent = parent
self.batch_size = batch_size
self.seq_length = seq_length
self.is_training = is_training
self.use_input_mask = use_input_mask
self.use_token_type_ids = use_token_type_ids
self.use_labels = use_labels
self.vocab_size = vocab_size
self.hidden_size = hidden_size
self.num_hidden_layers = num_hidden_layers
self.num_attention_heads = num_attention_heads
self.intermediate_size = intermediate_size
self.hidden_act = hidden_act
self.hidden_dropout_prob = hidden_dropout_prob
self.attention_probs_dropout_prob = attention_probs_dropout_prob
self.max_position_embeddings = max_position_embeddings
self.type_vocab_size = type_vocab_size
self.type_sequence_label_size = type_sequence_label_size
self.initializer_range = initializer_range
self.num_labels = num_labels
self.num_choices = num_choices
self.scope = scope
def prepare_config_and_inputs(self):
input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size)
input_mask = None
if self.use_input_mask:
input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
token_type_ids = None
if self.use_token_type_ids:
token_type_ids = ids_tensor([self.batch_size, self.seq_length], self.type_vocab_size)
sequence_labels = None
token_labels = None
choice_labels = None
if self.use_labels:
sequence_labels = ids_tensor([self.batch_size], self.type_sequence_label_size)
token_labels = ids_tensor([self.batch_size, self.seq_length], self.num_labels)
choice_labels = ids_tensor([self.batch_size], self.num_choices)
config = BertConfig(
vocab_size=self.vocab_size,
hidden_size=self.hidden_size,
num_hidden_layers=self.num_hidden_layers,
num_attention_heads=self.num_attention_heads,
intermediate_size=self.intermediate_size,
hidden_act=self.hidden_act,
hidden_dropout_prob=self.hidden_dropout_prob,
attention_probs_dropout_prob=self.attention_probs_dropout_prob,
max_position_embeddings=self.max_position_embeddings,
type_vocab_size=self.type_vocab_size,
is_decoder=False,
initializer_range=self.initializer_range,
)
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def prepare_config_and_inputs_for_decoder(self):
(
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
) = self.prepare_config_and_inputs()
config.is_decoder = True
encoder_hidden_states = floats_tensor([self.batch_size, self.seq_length, self.hidden_size])
encoder_attention_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
return (
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
encoder_hidden_states,
encoder_attention_mask,
)
def check_loss_output(self, result):
self.parent.assertListEqual(list(result["loss"].size()), [])
def create_and_check_bert_model(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = BertModel(config=config)
model.to(torch_device)
model.eval()
sequence_output, pooled_output = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids)
sequence_output, pooled_output = model(input_ids, token_type_ids=token_type_ids)
sequence_output, pooled_output = model(input_ids)
result = {
"sequence_output": sequence_output,
"pooled_output": pooled_output,
}
self.parent.assertListEqual(
list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size]
)
self.parent.assertListEqual(list(result["pooled_output"].size()), [self.batch_size, self.hidden_size])
def create_and_check_bert_model_as_decoder(
self,
config, config,
input_ids, input_ids,
token_type_ids, token_type_ids,
...@@ -198,56 +132,13 @@ class BertModelTest(ModelTesterMixin, unittest.TestCase): ...@@ -198,56 +132,13 @@ class BertModelTest(ModelTesterMixin, unittest.TestCase):
sequence_labels, sequence_labels,
token_labels, token_labels,
choice_labels, choice_labels,
encoder_hidden_states, ) = self.prepare_config_and_inputs()
encoder_attention_mask,
): config.is_decoder = True
model = BertModel(config) encoder_hidden_states = floats_tensor([self.batch_size, self.seq_length, self.hidden_size])
model.to(torch_device) encoder_attention_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
model.eval()
sequence_output, pooled_output = model( return (
input_ids,
attention_mask=input_mask,
token_type_ids=token_type_ids,
encoder_hidden_states=encoder_hidden_states,
encoder_attention_mask=encoder_attention_mask,
)
sequence_output, pooled_output = model(
input_ids,
attention_mask=input_mask,
token_type_ids=token_type_ids,
encoder_hidden_states=encoder_hidden_states,
)
sequence_output, pooled_output = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids)
result = {
"sequence_output": sequence_output,
"pooled_output": pooled_output,
}
self.parent.assertListEqual(
list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size]
)
self.parent.assertListEqual(list(result["pooled_output"].size()), [self.batch_size, self.hidden_size])
def create_and_check_bert_for_masked_lm(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = BertForMaskedLM(config=config)
model.to(torch_device)
model.eval()
loss, prediction_scores = model(
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, masked_lm_labels=token_labels
)
result = {
"loss": loss,
"prediction_scores": prediction_scores,
}
self.parent.assertListEqual(
list(result["prediction_scores"].size()), [self.batch_size, self.seq_length, self.vocab_size]
)
self.check_loss_output(result)
def create_and_check_bert_model_for_masked_lm_as_decoder(
self,
config, config,
input_ids, input_ids,
token_type_ids, token_type_ids,
...@@ -257,174 +148,277 @@ class BertModelTest(ModelTesterMixin, unittest.TestCase): ...@@ -257,174 +148,277 @@ class BertModelTest(ModelTesterMixin, unittest.TestCase):
choice_labels, choice_labels,
encoder_hidden_states, encoder_hidden_states,
encoder_attention_mask, encoder_attention_mask,
): )
model = BertForMaskedLM(config=config)
model.to(torch_device) def check_loss_output(self, result):
model.eval() self.parent.assertListEqual(list(result["loss"].size()), [])
loss, prediction_scores = model(
input_ids, def create_and_check_bert_model(
attention_mask=input_mask, self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
token_type_ids=token_type_ids, ):
masked_lm_labels=token_labels, model = BertModel(config=config)
encoder_hidden_states=encoder_hidden_states, model.to(torch_device)
encoder_attention_mask=encoder_attention_mask, model.eval()
) sequence_output, pooled_output = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids)
loss, prediction_scores = model( sequence_output, pooled_output = model(input_ids, token_type_ids=token_type_ids)
input_ids, sequence_output, pooled_output = model(input_ids)
attention_mask=input_mask,
token_type_ids=token_type_ids, result = {
masked_lm_labels=token_labels, "sequence_output": sequence_output,
encoder_hidden_states=encoder_hidden_states, "pooled_output": pooled_output,
) }
result = { self.parent.assertListEqual(
"loss": loss, list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size]
"prediction_scores": prediction_scores, )
} self.parent.assertListEqual(list(result["pooled_output"].size()), [self.batch_size, self.hidden_size])
self.parent.assertListEqual(
list(result["prediction_scores"].size()), [self.batch_size, self.seq_length, self.vocab_size] def create_and_check_bert_model_as_decoder(
) self,
self.check_loss_output(result) config,
input_ids,
def create_and_check_bert_for_next_sequence_prediction( token_type_ids,
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels input_mask,
): sequence_labels,
model = BertForNextSentencePrediction(config=config) token_labels,
model.to(torch_device) choice_labels,
model.eval() encoder_hidden_states,
loss, seq_relationship_score = model( encoder_attention_mask,
input_ids, ):
attention_mask=input_mask, model = BertModel(config)
token_type_ids=token_type_ids, model.to(torch_device)
next_sentence_label=sequence_labels, model.eval()
) sequence_output, pooled_output = model(
result = { input_ids,
"loss": loss, attention_mask=input_mask,
"seq_relationship_score": seq_relationship_score, token_type_ids=token_type_ids,
} encoder_hidden_states=encoder_hidden_states,
self.parent.assertListEqual(list(result["seq_relationship_score"].size()), [self.batch_size, 2]) encoder_attention_mask=encoder_attention_mask,
self.check_loss_output(result) )
sequence_output, pooled_output = model(
def create_and_check_bert_for_pretraining( input_ids,
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels attention_mask=input_mask,
): token_type_ids=token_type_ids,
model = BertForPreTraining(config=config) encoder_hidden_states=encoder_hidden_states,
model.to(torch_device) )
model.eval() sequence_output, pooled_output = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids)
loss, prediction_scores, seq_relationship_score = model(
input_ids, result = {
attention_mask=input_mask, "sequence_output": sequence_output,
token_type_ids=token_type_ids, "pooled_output": pooled_output,
masked_lm_labels=token_labels, }
next_sentence_label=sequence_labels, self.parent.assertListEqual(
) list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size]
result = { )
"loss": loss, self.parent.assertListEqual(list(result["pooled_output"].size()), [self.batch_size, self.hidden_size])
"prediction_scores": prediction_scores,
"seq_relationship_score": seq_relationship_score, def create_and_check_bert_for_masked_lm(
} self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
self.parent.assertListEqual( ):
list(result["prediction_scores"].size()), [self.batch_size, self.seq_length, self.vocab_size] model = BertForMaskedLM(config=config)
) model.to(torch_device)
self.parent.assertListEqual(list(result["seq_relationship_score"].size()), [self.batch_size, 2]) model.eval()
self.check_loss_output(result) loss, prediction_scores = model(
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, masked_lm_labels=token_labels
def create_and_check_bert_for_question_answering( )
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels result = {
): "loss": loss,
model = BertForQuestionAnswering(config=config) "prediction_scores": prediction_scores,
model.to(torch_device) }
model.eval() self.parent.assertListEqual(
loss, start_logits, end_logits = model( list(result["prediction_scores"].size()), [self.batch_size, self.seq_length, self.vocab_size]
input_ids, )
attention_mask=input_mask, self.check_loss_output(result)
token_type_ids=token_type_ids,
start_positions=sequence_labels, def create_and_check_bert_model_for_masked_lm_as_decoder(
end_positions=sequence_labels, self,
) config,
result = { input_ids,
"loss": loss, token_type_ids,
"start_logits": start_logits, input_mask,
"end_logits": end_logits, sequence_labels,
} token_labels,
self.parent.assertListEqual(list(result["start_logits"].size()), [self.batch_size, self.seq_length]) choice_labels,
self.parent.assertListEqual(list(result["end_logits"].size()), [self.batch_size, self.seq_length]) encoder_hidden_states,
self.check_loss_output(result) encoder_attention_mask,
):
def create_and_check_bert_for_sequence_classification( model = BertForMaskedLM(config=config)
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels model.to(torch_device)
): model.eval()
config.num_labels = self.num_labels loss, prediction_scores = model(
model = BertForSequenceClassification(config) input_ids,
model.to(torch_device) attention_mask=input_mask,
model.eval() token_type_ids=token_type_ids,
loss, logits = model( masked_lm_labels=token_labels,
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=sequence_labels encoder_hidden_states=encoder_hidden_states,
) encoder_attention_mask=encoder_attention_mask,
result = { )
"loss": loss, loss, prediction_scores = model(
"logits": logits, input_ids,
} attention_mask=input_mask,
self.parent.assertListEqual(list(result["logits"].size()), [self.batch_size, self.num_labels]) token_type_ids=token_type_ids,
self.check_loss_output(result) masked_lm_labels=token_labels,
encoder_hidden_states=encoder_hidden_states,
def create_and_check_bert_for_token_classification( )
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels result = {
): "loss": loss,
config.num_labels = self.num_labels "prediction_scores": prediction_scores,
model = BertForTokenClassification(config=config) }
model.to(torch_device) self.parent.assertListEqual(
model.eval() list(result["prediction_scores"].size()), [self.batch_size, self.seq_length, self.vocab_size]
loss, logits = model( )
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=token_labels self.check_loss_output(result)
)
result = { def create_and_check_bert_for_next_sequence_prediction(
"loss": loss, self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
"logits": logits, ):
} model = BertForNextSentencePrediction(config=config)
self.parent.assertListEqual( model.to(torch_device)
list(result["logits"].size()), [self.batch_size, self.seq_length, self.num_labels] model.eval()
) loss, seq_relationship_score = model(
self.check_loss_output(result) input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, next_sentence_label=sequence_labels,
)
def create_and_check_bert_for_multiple_choice( result = {
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels "loss": loss,
): "seq_relationship_score": seq_relationship_score,
config.num_choices = self.num_choices }
model = BertForMultipleChoice(config=config) self.parent.assertListEqual(list(result["seq_relationship_score"].size()), [self.batch_size, 2])
model.to(torch_device) self.check_loss_output(result)
model.eval()
multiple_choice_inputs_ids = input_ids.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous() def create_and_check_bert_for_pretraining(
multiple_choice_token_type_ids = token_type_ids.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous() self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
multiple_choice_input_mask = input_mask.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous() ):
loss, logits = model( model = BertForPreTraining(config=config)
multiple_choice_inputs_ids, model.to(torch_device)
attention_mask=multiple_choice_input_mask, model.eval()
token_type_ids=multiple_choice_token_type_ids, loss, prediction_scores, seq_relationship_score = model(
labels=choice_labels, input_ids,
) attention_mask=input_mask,
result = { token_type_ids=token_type_ids,
"loss": loss, masked_lm_labels=token_labels,
"logits": logits, next_sentence_label=sequence_labels,
} )
self.parent.assertListEqual(list(result["logits"].size()), [self.batch_size, self.num_choices]) result = {
self.check_loss_output(result) "loss": loss,
"prediction_scores": prediction_scores,
def prepare_config_and_inputs_for_common(self): "seq_relationship_score": seq_relationship_score,
config_and_inputs = self.prepare_config_and_inputs() }
( self.parent.assertListEqual(
config, list(result["prediction_scores"].size()), [self.batch_size, self.seq_length, self.vocab_size]
input_ids, )
token_type_ids, self.parent.assertListEqual(list(result["seq_relationship_score"].size()), [self.batch_size, 2])
input_mask, self.check_loss_output(result)
sequence_labels,
token_labels, def create_and_check_bert_for_question_answering(
choice_labels, self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
) = config_and_inputs ):
inputs_dict = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": input_mask} model = BertForQuestionAnswering(config=config)
return config, inputs_dict model.to(torch_device)
model.eval()
loss, start_logits, end_logits = model(
input_ids,
attention_mask=input_mask,
token_type_ids=token_type_ids,
start_positions=sequence_labels,
end_positions=sequence_labels,
)
result = {
"loss": loss,
"start_logits": start_logits,
"end_logits": end_logits,
}
self.parent.assertListEqual(list(result["start_logits"].size()), [self.batch_size, self.seq_length])
self.parent.assertListEqual(list(result["end_logits"].size()), [self.batch_size, self.seq_length])
self.check_loss_output(result)
def create_and_check_bert_for_sequence_classification(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = BertForSequenceClassification(config)
model.to(torch_device)
model.eval()
loss, logits = model(
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=sequence_labels
)
result = {
"loss": loss,
"logits": logits,
}
self.parent.assertListEqual(list(result["logits"].size()), [self.batch_size, self.num_labels])
self.check_loss_output(result)
def create_and_check_bert_for_token_classification(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = BertForTokenClassification(config=config)
model.to(torch_device)
model.eval()
loss, logits = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=token_labels)
result = {
"loss": loss,
"logits": logits,
}
self.parent.assertListEqual(list(result["logits"].size()), [self.batch_size, self.seq_length, self.num_labels])
self.check_loss_output(result)
def create_and_check_bert_for_multiple_choice(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_choices = self.num_choices
model = BertForMultipleChoice(config=config)
model.to(torch_device)
model.eval()
multiple_choice_inputs_ids = input_ids.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous()
multiple_choice_token_type_ids = token_type_ids.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous()
multiple_choice_input_mask = input_mask.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous()
loss, logits = model(
multiple_choice_inputs_ids,
attention_mask=multiple_choice_input_mask,
token_type_ids=multiple_choice_token_type_ids,
labels=choice_labels,
)
result = {
"loss": loss,
"logits": logits,
}
self.parent.assertListEqual(list(result["logits"].size()), [self.batch_size, self.num_choices])
self.check_loss_output(result)
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
) = config_and_inputs
inputs_dict = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": input_mask}
return config, inputs_dict
@require_torch
class BertModelTest(ModelTesterMixin, unittest.TestCase):
all_model_classes = (
(
BertModel,
BertForMaskedLM,
BertForNextSentencePrediction,
BertForPreTraining,
BertForQuestionAnswering,
BertForSequenceClassification,
BertForTokenClassification,
)
if is_torch_available()
else ()
)
def setUp(self): def setUp(self):
self.model_tester = BertModelTest.BertModelTester(self) self.model_tester = BertModelTester(self)
self.config_tester = ConfigTester(self, config_class=BertConfig, hidden_size=37) self.config_tester = ConfigTester(self, config_class=BertConfig, hidden_size=37)
def test_config(self): def test_config(self):
......
tests/test_modeling_encoder_decoder.py
View file @ 8e67573a
...@@ -21,7 +21,7 @@ from transformers import is_torch_available ...@@ -21,7 +21,7 @@ from transformers import is_torch_available
# TODO(PVP): this line reruns all the tests in BertModelTest; not sure whether this can be prevented # TODO(PVP): this line reruns all the tests in BertModelTest; not sure whether this can be prevented
# for now only run module with pytest tests/test_modeling_encoder_decoder.py::EncoderDecoderModelTest # for now only run module with pytest tests/test_modeling_encoder_decoder.py::EncoderDecoderModelTest
from .test_modeling_bert import BertModelTest from .test_modeling_bert import BertModelTester
from .utils import require_torch, slow, torch_device from .utils import require_torch, slow, torch_device
...@@ -34,7 +34,7 @@ if is_torch_available(): ...@@ -34,7 +34,7 @@ if is_torch_available():
@require_torch @require_torch
class EncoderDecoderModelTest(unittest.TestCase): class EncoderDecoderModelTest(unittest.TestCase):
def prepare_config_and_inputs_bert(self): def prepare_config_and_inputs_bert(self):
bert_model_tester = BertModelTest.BertModelTester(self) bert_model_tester = BertModelTester(self)
encoder_config_and_inputs = bert_model_tester.prepare_config_and_inputs() encoder_config_and_inputs = bert_model_tester.prepare_config_and_inputs()
decoder_config_and_inputs = bert_model_tester.prepare_config_and_inputs_for_decoder() decoder_config_and_inputs = bert_model_tester.prepare_config_and_inputs_for_decoder()
( (
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment