Skip to content

GitLab

"...git@developer.sourcefind.cn:chenpangpang/transformers.git" did not exist on "dc3f6758cfe84a29fa87e9ce05b9b45e10cdb155"
Unverified Commit c852036b authored by Amil Khare's avatar Amil Khare Committed by GitHub
Browse files

[cleanup] Hoist ModelTester objects to top level (#4939) 


Co-authored-by: default avatarSam Shleifer <sshleifer@gmail.com>
parent 0c55a384
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 3360 additions and 3738 deletions
tests/test_modeling_albert.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_ctrl.py
View file @ c852036b
...@@ -27,6 +27,140 @@ if is_torch_available(): ...@@ -27,6 +27,140 @@ if is_torch_available():
from transformers import CTRLConfig, CTRLModel, CTRL_PRETRAINED_MODEL_ARCHIVE_LIST, CTRLLMHeadModel from transformers import CTRLConfig, CTRLModel, CTRL_PRETRAINED_MODEL_ARCHIVE_LIST, CTRLLMHeadModel
class CTRLModelTester:
def __init__(
self, parent,
):
self.parent = parent
self.batch_size = 14
self.seq_length = 7
self.is_training = True
self.use_token_type_ids = True
self.use_input_mask = True
self.use_labels = True
self.use_mc_token_ids = True
self.vocab_size = 99
self.hidden_size = 32
self.num_hidden_layers = 5
self.num_attention_heads = 4
self.intermediate_size = 37
self.hidden_act = "gelu"
self.hidden_dropout_prob = 0.1
self.attention_probs_dropout_prob = 0.1
self.max_position_embeddings = 512
self.type_vocab_size = 16
self.type_sequence_label_size = 2
self.initializer_range = 0.02
self.num_labels = 3
self.num_choices = 4
self.scope = None
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)
mc_token_ids = None
if self.use_mc_token_ids:
mc_token_ids = ids_tensor([self.batch_size, self.num_choices], self.seq_length)
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 = CTRLConfig(
vocab_size=self.vocab_size,
n_embd=self.hidden_size,
n_layer=self.num_hidden_layers,
n_head=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,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range
)
head_mask = ids_tensor([self.num_hidden_layers, self.num_attention_heads], 2)
return (
config,
input_ids,
input_mask,
head_mask,
token_type_ids,
mc_token_ids,
sequence_labels,
token_labels,
choice_labels,
)
def check_loss_output(self, result):
self.parent.assertListEqual(list(result["loss"].size()), [])
def create_and_check_ctrl_model(self, config, input_ids, input_mask, head_mask, token_type_ids, *args):
model = CTRLModel(config=config)
model.to(torch_device)
model.eval()
model(input_ids, token_type_ids=token_type_ids, head_mask=head_mask)
model(input_ids, token_type_ids=token_type_ids)
sequence_output, presents = model(input_ids)
result = {
"sequence_output": sequence_output,
"presents": presents,
}
self.parent.assertListEqual(
list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size]
)
self.parent.assertEqual(len(result["presents"]), config.n_layer)
def create_and_check_lm_head_model(self, config, input_ids, input_mask, head_mask, token_type_ids, *args):
model = CTRLLMHeadModel(config)
model.to(torch_device)
model.eval()
loss, lm_logits, _ = model(input_ids, token_type_ids=token_type_ids, labels=input_ids)
result = {"loss": loss, "lm_logits": lm_logits}
self.parent.assertListEqual(list(result["loss"].size()), [])
self.parent.assertListEqual(
list(result["lm_logits"].size()), [self.batch_size, self.seq_length, self.vocab_size]
)
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(
config,
input_ids,
input_mask,
head_mask,
token_type_ids,
mc_token_ids,
sequence_labels,
token_labels,
choice_labels,
) = config_and_inputs
inputs_dict = {"input_ids": input_ids, "token_type_ids": token_type_ids, "head_mask": head_mask}
return config, inputs_dict
@require_torch @require_torch
class CTRLModelTest(ModelTesterMixin, unittest.TestCase): class CTRLModelTest(ModelTesterMixin, unittest.TestCase):
...@@ -37,164 +171,8 @@ class CTRLModelTest(ModelTesterMixin, unittest.TestCase): ...@@ -37,164 +171,8 @@ class CTRLModelTest(ModelTesterMixin, unittest.TestCase):
test_resize_embeddings = False test_resize_embeddings = False
test_head_masking = False test_head_masking = False
class CTRLModelTester(object):
def __init__(
self,
parent,
batch_size=14,
seq_length=7,
is_training=True,
use_token_type_ids=True,
use_input_mask=True,
use_labels=True,
use_mc_token_ids=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_token_type_ids = use_token_type_ids
self.use_input_mask = use_input_mask
self.use_labels = use_labels
self.use_mc_token_ids = use_mc_token_ids
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)
mc_token_ids = None
if self.use_mc_token_ids:
mc_token_ids = ids_tensor([self.batch_size, self.num_choices], self.seq_length)
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 = CTRLConfig(
vocab_size=self.vocab_size,
n_embd=self.hidden_size,
n_layer=self.num_hidden_layers,
n_head=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,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range
)
head_mask = ids_tensor([self.num_hidden_layers, self.num_attention_heads], 2)
return (
config,
input_ids,
input_mask,
head_mask,
token_type_ids,
mc_token_ids,
sequence_labels,
token_labels,
choice_labels,
)
def check_loss_output(self, result):
self.parent.assertListEqual(list(result["loss"].size()), [])
def create_and_check_ctrl_model(self, config, input_ids, input_mask, head_mask, token_type_ids, *args):
model = CTRLModel(config=config)
model.to(torch_device)
model.eval()
model(input_ids, token_type_ids=token_type_ids, head_mask=head_mask)
model(input_ids, token_type_ids=token_type_ids)
sequence_output, presents = model(input_ids)
result = {
"sequence_output": sequence_output,
"presents": presents,
}
self.parent.assertListEqual(
list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size]
)
self.parent.assertEqual(len(result["presents"]), config.n_layer)
def create_and_check_lm_head_model(self, config, input_ids, input_mask, head_mask, token_type_ids, *args):
model = CTRLLMHeadModel(config)
model.to(torch_device)
model.eval()
loss, lm_logits, _ = model(input_ids, token_type_ids=token_type_ids, labels=input_ids)
result = {"loss": loss, "lm_logits": lm_logits}
self.parent.assertListEqual(list(result["loss"].size()), [])
self.parent.assertListEqual(
list(result["lm_logits"].size()), [self.batch_size, self.seq_length, self.vocab_size]
)
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(
config,
input_ids,
input_mask,
head_mask,
token_type_ids,
mc_token_ids,
sequence_labels,
token_labels,
choice_labels,
) = config_and_inputs
inputs_dict = {"input_ids": input_ids, "token_type_ids": token_type_ids, "head_mask": head_mask}
return config, inputs_dict
def setUp(self): def setUp(self):
self.model_tester = CTRLModelTest.CTRLModelTester(self) self.model_tester = CTRLModelTester(self)
self.config_tester = ConfigTester(self, config_class=CTRLConfig, n_embd=37) self.config_tester = ConfigTester(self, config_class=CTRLConfig, n_embd=37)
def test_config(self): def test_config(self):
......
tests/test_modeling_distilbert.py
View file @ c852036b
...@@ -34,27 +34,6 @@ if is_torch_available(): ...@@ -34,27 +34,6 @@ if is_torch_available():
DistilBertForSequenceClassification, DistilBertForSequenceClassification,
) )
@require_torch
class DistilBertModelTest(ModelTesterMixin, unittest.TestCase):
all_model_classes = (
(
DistilBertModel,
DistilBertForMaskedLM,
DistilBertForMultipleChoice,
DistilBertForQuestionAnswering,
DistilBertForSequenceClassification,
DistilBertForTokenClassification,
)
if is_torch_available()
else None
)
test_pruning = True
test_torchscript = True
test_resize_embeddings = True
test_head_masking = True
class DistilBertModelTester(object): class DistilBertModelTester(object):
def __init__( def __init__(
self, self,
...@@ -245,8 +224,29 @@ class DistilBertModelTest(ModelTesterMixin, unittest.TestCase): ...@@ -245,8 +224,29 @@ class DistilBertModelTest(ModelTesterMixin, unittest.TestCase):
inputs_dict = {"input_ids": input_ids, "attention_mask": input_mask} inputs_dict = {"input_ids": input_ids, "attention_mask": input_mask}
return config, inputs_dict return config, inputs_dict
@require_torch
class DistilBertModelTest(ModelTesterMixin, unittest.TestCase):
all_model_classes = (
(
DistilBertModel,
DistilBertForMaskedLM,
DistilBertForMultipleChoice,
DistilBertForQuestionAnswering,
DistilBertForSequenceClassification,
DistilBertForTokenClassification,
)
if is_torch_available()
else None
)
test_pruning = True
test_torchscript = True
test_resize_embeddings = True
test_head_masking = True
def setUp(self): def setUp(self):
self.model_tester = DistilBertModelTest.DistilBertModelTester(self) self.model_tester = DistilBertModelTester(self)
self.config_tester = ConfigTester(self, config_class=DistilBertConfig, dim=37) self.config_tester = ConfigTester(self, config_class=DistilBertConfig, dim=37)
def test_config(self): def test_config(self):
......
tests/test_modeling_electra.py
View file @ c852036b
...@@ -36,200 +36,69 @@ if is_torch_available(): ...@@ -36,200 +36,69 @@ if is_torch_available():
from transformers.modeling_electra import ELECTRA_PRETRAINED_MODEL_ARCHIVE_LIST from transformers.modeling_electra import ELECTRA_PRETRAINED_MODEL_ARCHIVE_LIST
@require_torch class ElectraModelTester:
class ElectraModelTest(ModelTesterMixin, unittest.TestCase): def __init__(
self, parent,
all_model_classes = ( ):
( self.parent = parent
ElectraModel, self.batch_size = 13
ElectraForPreTraining, self.seq_length = 7
ElectraForMaskedLM, self.is_training = True
ElectraForTokenClassification, self.use_input_mask = True
ElectraForSequenceClassification, self.use_token_type_ids = True
ElectraForQuestionAnswering, self.use_labels = True
self.vocab_size = 99
self.hidden_size = 32
self.num_hidden_layers = 5
self.num_attention_heads = 4
self.intermediate_size = 37
self.hidden_act = "gelu"
self.hidden_dropout_prob = 0.1
self.attention_probs_dropout_prob = 0.1
self.max_position_embeddings = 512
self.type_vocab_size = 16
self.type_sequence_label_size = 2
self.initializer_range = 0.02
self.num_labels = 3
self.num_choices = 4
self.scope = None
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)
fake_token_labels = ids_tensor([self.batch_size, self.seq_length], 1)
config = ElectraConfig(
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 ElectraModelTester(object): return (
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)
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)
fake_token_labels = ids_tensor([self.batch_size, self.seq_length], 1)
config = ElectraConfig(
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,
fake_token_labels,
)
def check_loss_output(self, result):
self.parent.assertListEqual(list(result["loss"].size()), [])
def create_and_check_electra_model(
self,
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
fake_token_labels,
):
model = ElectraModel(config=config)
model.to(torch_device)
model.eval()
(sequence_output,) = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids)
(sequence_output,) = model(input_ids, token_type_ids=token_type_ids)
(sequence_output,) = model(input_ids)
result = {
"sequence_output": sequence_output,
}
self.parent.assertListEqual(
list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size]
)
def create_and_check_electra_for_masked_lm(
self,
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
fake_token_labels,
):
model = ElectraForMaskedLM(config=config)
model.to(torch_device)
model.eval()
loss, prediction_scores = model(
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, 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_electra_for_token_classification(
self,
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
fake_token_labels,
):
config.num_labels = self.num_labels
model = ElectraForTokenClassification(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_electra_for_pretraining(
self,
config, config,
input_ids, input_ids,
token_type_ids, token_type_ids,
...@@ -238,48 +107,168 @@ class ElectraModelTest(ModelTesterMixin, unittest.TestCase): ...@@ -238,48 +107,168 @@ class ElectraModelTest(ModelTesterMixin, unittest.TestCase):
token_labels, token_labels,
choice_labels, choice_labels,
fake_token_labels, fake_token_labels,
): )
config.num_labels = self.num_labels
model = ElectraForPreTraining(config=config) def check_loss_output(self, result):
model.to(torch_device) self.parent.assertListEqual(list(result["loss"].size()), [])
model.eval()
loss, logits = model( def create_and_check_electra_model(
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=fake_token_labels self,
) config,
result = { input_ids,
"loss": loss, token_type_ids,
"logits": logits, input_mask,
} sequence_labels,
self.parent.assertListEqual(list(result["logits"].size()), [self.batch_size, self.seq_length]) token_labels,
self.check_loss_output(result) choice_labels,
fake_token_labels,
def create_and_check_electra_for_sequence_classification( ):
self, model = ElectraModel(config=config)
config, model.to(torch_device)
model.eval()
(sequence_output,) = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids)
(sequence_output,) = model(input_ids, token_type_ids=token_type_ids)
(sequence_output,) = model(input_ids)
result = {
"sequence_output": sequence_output,
}
self.parent.assertListEqual(
list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size]
)
def create_and_check_electra_for_masked_lm(
self,
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
fake_token_labels,
):
model = ElectraForMaskedLM(config=config)
model.to(torch_device)
model.eval()
loss, prediction_scores = model(
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, 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_electra_for_token_classification(
self,
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
fake_token_labels,
):
config.num_labels = self.num_labels
model = ElectraForTokenClassification(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_electra_for_pretraining(
self,
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
fake_token_labels,
):
config.num_labels = self.num_labels
model = ElectraForPreTraining(config=config)
model.to(torch_device)
model.eval()
loss, logits = model(
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=fake_token_labels
)
result = {
"loss": loss,
"logits": logits,
}
self.parent.assertListEqual(list(result["logits"].size()), [self.batch_size, self.seq_length])
self.check_loss_output(result)
def create_and_check_electra_for_sequence_classification(
self,
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
fake_token_labels,
):
config.num_labels = self.num_labels
model = ElectraForSequenceClassification(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_electra_for_question_answering(
self,
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
fake_token_labels,
):
model = ElectraForQuestionAnswering(config=config)
model.to(torch_device)
model.eval()
loss, start_logits, end_logits = model(
input_ids, input_ids,
token_type_ids, attention_mask=input_mask,
input_mask, token_type_ids=token_type_ids,
sequence_labels, start_positions=sequence_labels,
token_labels, end_positions=sequence_labels,
choice_labels, )
fake_token_labels, result = {
): "loss": loss,
config.num_labels = self.num_labels "start_logits": start_logits,
model = ElectraForSequenceClassification(config) "end_logits": end_logits,
model.to(torch_device) }
model.eval() self.parent.assertListEqual(list(result["start_logits"].size()), [self.batch_size, self.seq_length])
loss, logits = model( self.parent.assertListEqual(list(result["end_logits"].size()), [self.batch_size, self.seq_length])
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=sequence_labels self.check_loss_output(result)
)
result = { def prepare_config_and_inputs_for_common(self):
"loss": loss, config_and_inputs = self.prepare_config_and_inputs()
"logits": logits, (
}
self.parent.assertListEqual(list(result["logits"].size()), [self.batch_size, self.num_labels])
self.check_loss_output(result)
def create_and_check_electra_for_question_answering(
self,
config, config,
input_ids, input_ids,
token_type_ids, token_type_ids,
...@@ -288,43 +277,29 @@ class ElectraModelTest(ModelTesterMixin, unittest.TestCase): ...@@ -288,43 +277,29 @@ class ElectraModelTest(ModelTesterMixin, unittest.TestCase):
token_labels, token_labels,
choice_labels, choice_labels,
fake_token_labels, fake_token_labels,
): ) = config_and_inputs
model = ElectraForQuestionAnswering(config=config) inputs_dict = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": input_mask}
model.to(torch_device) return config, inputs_dict
model.eval()
loss, start_logits, end_logits = model(
input_ids, @require_torch
attention_mask=input_mask, class ElectraModelTest(ModelTesterMixin, unittest.TestCase):
token_type_ids=token_type_ids,
start_positions=sequence_labels, all_model_classes = (
end_positions=sequence_labels, (
) ElectraModel,
result = { ElectraForPreTraining,
"loss": loss, ElectraForMaskedLM,
"start_logits": start_logits, ElectraForTokenClassification,
"end_logits": end_logits, ElectraForSequenceClassification,
} ElectraForQuestionAnswering,
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]) if is_torch_available()
self.check_loss_output(result) else ()
)
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,
fake_token_labels,
) = config_and_inputs
inputs_dict = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": input_mask}
return config, inputs_dict
def setUp(self): def setUp(self):
self.model_tester = ElectraModelTest.ElectraModelTester(self) self.model_tester = ElectraModelTester(self)
self.config_tester = ConfigTester(self, config_class=ElectraConfig, hidden_size=37) self.config_tester = ConfigTester(self, config_class=ElectraConfig, hidden_size=37)
def test_config(self): def test_config(self):
......
tests/test_modeling_flaubert.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_gpt2.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_longformer.py
View file @ c852036b
...@@ -36,56 +36,33 @@ if is_torch_available(): ...@@ -36,56 +36,33 @@ if is_torch_available():
) )
class LongformerModelTester(object): class LongformerModelTester:
def __init__( def __init__(
self, self, parent,
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,
attention_window=4,
): ):
self.parent = parent self.parent = parent
self.batch_size = batch_size self.batch_size = 13
self.seq_length = seq_length self.seq_length = 7
self.is_training = is_training self.is_training = True
self.use_input_mask = use_input_mask self.use_input_mask = True
self.use_token_type_ids = use_token_type_ids self.use_token_type_ids = True
self.use_labels = use_labels self.use_labels = True
self.vocab_size = vocab_size self.vocab_size = 99
self.hidden_size = hidden_size self.hidden_size = 32
self.num_hidden_layers = num_hidden_layers self.num_hidden_layers = 5
self.num_attention_heads = num_attention_heads self.num_attention_heads = 4
self.intermediate_size = intermediate_size self.intermediate_size = 37
self.hidden_act = hidden_act self.hidden_act = "gelu"
self.hidden_dropout_prob = hidden_dropout_prob self.hidden_dropout_prob = 0.1
self.attention_probs_dropout_prob = attention_probs_dropout_prob self.attention_probs_dropout_prob = 0.1
self.max_position_embeddings = max_position_embeddings self.max_position_embeddings = 512
self.type_vocab_size = type_vocab_size self.type_vocab_size = 16
self.type_sequence_label_size = type_sequence_label_size self.type_sequence_label_size = 2
self.initializer_range = initializer_range self.initializer_range = 0.02
self.num_labels = num_labels self.num_labels = 3
self.num_choices = num_choices self.num_choices = 4
self.scope = scope self.scope = None
self.attention_window = attention_window self.attention_window = 4
# `ModelTesterMixin.test_attention_outputs` is expecting attention tensors to be of size # `ModelTesterMixin.test_attention_outputs` is expecting attention tensors to be of size
# [num_attention_heads, encoder_seq_length, encoder_key_length], but LongformerSelfAttention # [num_attention_heads, encoder_seq_length, encoder_key_length], but LongformerSelfAttention
......
tests/test_modeling_openai.py
View file @ c852036b
...@@ -34,6 +34,139 @@ if is_torch_available(): ...@@ -34,6 +34,139 @@ if is_torch_available():
) )
class OpenAIGPTModelTester:
def __init__(
self, parent,
):
self.parent = parent
self.batch_size = 13
self.seq_length = 7
self.is_training = True
self.use_token_type_ids = True
self.use_labels = True
self.vocab_size = 99
self.hidden_size = 32
self.num_hidden_layers = 5
self.num_attention_heads = 4
self.intermediate_size = 37
self.hidden_act = "gelu"
self.hidden_dropout_prob = 0.1
self.attention_probs_dropout_prob = 0.1
self.max_position_embeddings = 512
self.type_vocab_size = 16
self.type_sequence_label_size = 2
self.initializer_range = 0.02
self.num_labels = 3
self.num_choices = 4
self.scope = None
def prepare_config_and_inputs(self):
input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size)
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 = OpenAIGPTConfig(
vocab_size=self.vocab_size,
n_embd=self.hidden_size,
n_layer=self.num_hidden_layers,
n_head=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,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range
)
head_mask = ids_tensor([self.num_hidden_layers, self.num_attention_heads], 2)
return (
config,
input_ids,
head_mask,
token_type_ids,
sequence_labels,
token_labels,
choice_labels,
)
def check_loss_output(self, result):
self.parent.assertListEqual(list(result["loss"].size()), [])
def create_and_check_openai_gpt_model(self, config, input_ids, head_mask, token_type_ids, *args):
model = OpenAIGPTModel(config=config)
model.to(torch_device)
model.eval()
model(input_ids, token_type_ids=token_type_ids, head_mask=head_mask)
model(input_ids, token_type_ids=token_type_ids)
(sequence_output,) = model(input_ids)
result = {"sequence_output": sequence_output}
self.parent.assertListEqual(
list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size],
)
def create_and_check_lm_head_model(self, config, input_ids, head_mask, token_type_ids, *args):
model = OpenAIGPTLMHeadModel(config)
model.to(torch_device)
model.eval()
loss, lm_logits = model(input_ids, token_type_ids=token_type_ids, labels=input_ids)
result = {"loss": loss, "lm_logits": lm_logits}
self.parent.assertListEqual(list(result["loss"].size()), [])
self.parent.assertListEqual(
list(result["lm_logits"].size()), [self.batch_size, self.seq_length, self.vocab_size],
)
def create_and_check_double_lm_head_model(self, config, input_ids, head_mask, token_type_ids, *args):
model = OpenAIGPTDoubleHeadsModel(config)
model.to(torch_device)
model.eval()
loss, lm_logits, mc_logits = model(input_ids, token_type_ids=token_type_ids, labels=input_ids)
result = {"loss": loss, "lm_logits": lm_logits}
self.parent.assertListEqual(list(result["loss"].size()), [])
self.parent.assertListEqual(
list(result["lm_logits"].size()), [self.batch_size, self.seq_length, self.vocab_size],
)
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(
config,
input_ids,
head_mask,
token_type_ids,
sequence_labels,
token_labels,
choice_labels,
) = config_and_inputs
inputs_dict = {
"input_ids": input_ids,
"token_type_ids": token_type_ids,
"head_mask": head_mask,
}
return config, inputs_dict
@require_torch @require_torch
class OpenAIGPTModelTest(ModelTesterMixin, unittest.TestCase): class OpenAIGPTModelTest(ModelTesterMixin, unittest.TestCase):
...@@ -44,161 +177,8 @@ class OpenAIGPTModelTest(ModelTesterMixin, unittest.TestCase): ...@@ -44,161 +177,8 @@ class OpenAIGPTModelTest(ModelTesterMixin, unittest.TestCase):
(OpenAIGPTLMHeadModel,) if is_torch_available() else () (OpenAIGPTLMHeadModel,) if is_torch_available() else ()
) # TODO (PVP): Add Double HeadsModel when generate() function is changed accordingly ) # TODO (PVP): Add Double HeadsModel when generate() function is changed accordingly
class OpenAIGPTModelTester(object):
def __init__(
self,
parent,
batch_size=13,
seq_length=7,
is_training=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_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)
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 = OpenAIGPTConfig(
vocab_size=self.vocab_size,
n_embd=self.hidden_size,
n_layer=self.num_hidden_layers,
n_head=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,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range
)
head_mask = ids_tensor([self.num_hidden_layers, self.num_attention_heads], 2)
return (
config,
input_ids,
head_mask,
token_type_ids,
sequence_labels,
token_labels,
choice_labels,
)
def check_loss_output(self, result):
self.parent.assertListEqual(list(result["loss"].size()), [])
def create_and_check_openai_gpt_model(self, config, input_ids, head_mask, token_type_ids, *args):
model = OpenAIGPTModel(config=config)
model.to(torch_device)
model.eval()
model(input_ids, token_type_ids=token_type_ids, head_mask=head_mask)
model(input_ids, token_type_ids=token_type_ids)
(sequence_output,) = model(input_ids)
result = {"sequence_output": sequence_output}
self.parent.assertListEqual(
list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size],
)
def create_and_check_lm_head_model(self, config, input_ids, head_mask, token_type_ids, *args):
model = OpenAIGPTLMHeadModel(config)
model.to(torch_device)
model.eval()
loss, lm_logits = model(input_ids, token_type_ids=token_type_ids, labels=input_ids)
result = {"loss": loss, "lm_logits": lm_logits}
self.parent.assertListEqual(list(result["loss"].size()), [])
self.parent.assertListEqual(
list(result["lm_logits"].size()), [self.batch_size, self.seq_length, self.vocab_size],
)
def create_and_check_double_lm_head_model(self, config, input_ids, head_mask, token_type_ids, *args):
model = OpenAIGPTDoubleHeadsModel(config)
model.to(torch_device)
model.eval()
loss, lm_logits, mc_logits = model(input_ids, token_type_ids=token_type_ids, labels=input_ids)
result = {"loss": loss, "lm_logits": lm_logits}
self.parent.assertListEqual(list(result["loss"].size()), [])
self.parent.assertListEqual(
list(result["lm_logits"].size()), [self.batch_size, self.seq_length, self.vocab_size],
)
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(
config,
input_ids,
head_mask,
token_type_ids,
sequence_labels,
token_labels,
choice_labels,
) = config_and_inputs
inputs_dict = {
"input_ids": input_ids,
"token_type_ids": token_type_ids,
"head_mask": head_mask,
}
return config, inputs_dict
def setUp(self): def setUp(self):
self.model_tester = OpenAIGPTModelTest.OpenAIGPTModelTester(self) self.model_tester = OpenAIGPTModelTester(self)
self.config_tester = ConfigTester(self, config_class=OpenAIGPTConfig, n_embd=37) self.config_tester = ConfigTester(self, config_class=OpenAIGPTConfig, n_embd=37)
def test_config(self): def test_config(self):
......
tests/test_modeling_roberta.py
View file @ c852036b
...@@ -39,6 +39,183 @@ if is_torch_available(): ...@@ -39,6 +39,183 @@ if is_torch_available():
from transformers.modeling_utils import create_position_ids_from_input_ids from transformers.modeling_utils import create_position_ids_from_input_ids
class RobertaModelTester:
def __init__(
self, parent,
):
self.parent = parent
self.batch_size = 13
self.seq_length = 7
self.is_training = True
self.use_input_mask = True
self.use_token_type_ids = True
self.use_labels = True
self.vocab_size = 99
self.hidden_size = 32
self.num_hidden_layers = 5
self.num_attention_heads = 4
self.intermediate_size = 37
self.hidden_act = "gelu"
self.hidden_dropout_prob = 0.1
self.attention_probs_dropout_prob = 0.1
self.max_position_embeddings = 512
self.type_vocab_size = 16
self.type_sequence_label_size = 2
self.initializer_range = 0.02
self.num_labels = 3
self.num_choices = 4
self.scope = None
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 = RobertaConfig(
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,
initializer_range=self.initializer_range,
)
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def check_loss_output(self, result):
self.parent.assertListEqual(list(result["loss"].size()), [])
def create_and_check_roberta_model(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = RobertaModel(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_roberta_for_masked_lm(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = RobertaForMaskedLM(config=config)
model.to(torch_device)
model.eval()
loss, prediction_scores = model(
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, 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_roberta_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 = RobertaForTokenClassification(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_roberta_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 = RobertaForMultipleChoice(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 create_and_check_roberta_for_question_answering(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = RobertaForQuestionAnswering(config=config)
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 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 @require_torch
class RobertaModelTest(ModelTesterMixin, unittest.TestCase): class RobertaModelTest(ModelTesterMixin, unittest.TestCase):
...@@ -55,210 +232,8 @@ class RobertaModelTest(ModelTesterMixin, unittest.TestCase): ...@@ -55,210 +232,8 @@ class RobertaModelTest(ModelTesterMixin, unittest.TestCase):
else () else ()
) )
class RobertaModelTester(object):
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)
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 = RobertaConfig(
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,
initializer_range=self.initializer_range,
)
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def check_loss_output(self, result):
self.parent.assertListEqual(list(result["loss"].size()), [])
def create_and_check_roberta_model(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = RobertaModel(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_roberta_for_masked_lm(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = RobertaForMaskedLM(config=config)
model.to(torch_device)
model.eval()
loss, prediction_scores = model(
input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, 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_roberta_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 = RobertaForTokenClassification(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_roberta_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 = RobertaForMultipleChoice(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 create_and_check_roberta_for_question_answering(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = RobertaForQuestionAnswering(config=config)
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 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
def setUp(self): def setUp(self):
self.model_tester = RobertaModelTest.RobertaModelTester(self) self.model_tester = RobertaModelTester(self)
self.config_tester = ConfigTester(self, config_class=RobertaConfig, hidden_size=37) self.config_tester = ConfigTester(self, config_class=RobertaConfig, hidden_size=37)
def test_config(self): def test_config(self):
......
tests/test_modeling_t5.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_tf_albert.py
View file @ c852036b
...@@ -34,6 +34,186 @@ if is_tf_available(): ...@@ -34,6 +34,186 @@ if is_tf_available():
) )
class TFAlbertModelTester:
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,
embedding_size=16,
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 = 13
self.seq_length = 7
self.is_training = True
self.use_input_mask = True
self.use_token_type_ids = True
self.use_labels = True
self.vocab_size = 99
self.embedding_size = 16
self.hidden_size = 32
self.num_hidden_layers = 5
self.num_attention_heads = 4
self.intermediate_size = 37
self.hidden_act = "gelu"
self.hidden_dropout_prob = 0.1
self.attention_probs_dropout_prob = 0.1
self.max_position_embeddings = 512
self.type_vocab_size = 16
self.type_sequence_label_size = 2
self.initializer_range = 0.02
self.num_labels = 3
self.num_choices = 4
self.scope = None
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 = AlbertConfig(
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,
initializer_range=self.initializer_range,
)
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def create_and_check_albert_model(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFAlbertModel(config=config)
# inputs = {'input_ids': input_ids,
# 'attention_mask': input_mask,
# 'token_type_ids': token_type_ids}
# sequence_output, pooled_output = model(**inputs)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
sequence_output, pooled_output = model(inputs)
inputs = [input_ids, input_mask]
sequence_output, pooled_output = model(inputs)
sequence_output, pooled_output = model(input_ids)
result = {
"sequence_output": sequence_output.numpy(),
"pooled_output": pooled_output.numpy(),
}
self.parent.assertListEqual(
list(result["sequence_output"].shape), [self.batch_size, self.seq_length, self.hidden_size]
)
self.parent.assertListEqual(list(result["pooled_output"].shape), [self.batch_size, self.hidden_size])
def create_and_check_albert_for_pretraining(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = TFAlbertForPreTraining(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
prediction_scores, sop_scores = model(inputs)
result = {
"prediction_scores": prediction_scores.numpy(),
"sop_scores": sop_scores.numpy(),
}
self.parent.assertListEqual(
list(result["prediction_scores"].shape), [self.batch_size, self.seq_length, self.vocab_size]
)
self.parent.assertListEqual(list(result["sop_scores"].shape), [self.batch_size, self.num_labels])
def create_and_check_albert_for_masked_lm(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFAlbertForMaskedLM(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
(prediction_scores,) = model(inputs)
result = {
"prediction_scores": prediction_scores.numpy(),
}
self.parent.assertListEqual(
list(result["prediction_scores"].shape), [self.batch_size, self.seq_length, self.vocab_size]
)
def create_and_check_albert_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 = TFAlbertForSequenceClassification(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
(logits,) = model(inputs)
result = {
"logits": logits.numpy(),
}
self.parent.assertListEqual(list(result["logits"].shape), [self.batch_size, self.num_labels])
def create_and_check_albert_for_question_answering(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFAlbertForQuestionAnswering(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
start_logits, end_logits = model(inputs)
result = {
"start_logits": start_logits.numpy(),
"end_logits": end_logits.numpy(),
}
self.parent.assertListEqual(list(result["start_logits"].shape), [self.batch_size, self.seq_length])
self.parent.assertListEqual(list(result["end_logits"].shape), [self.batch_size, self.seq_length])
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_tf @require_tf
class TFAlbertModelTest(TFModelTesterMixin, unittest.TestCase): class TFAlbertModelTest(TFModelTesterMixin, unittest.TestCase):
...@@ -49,187 +229,8 @@ class TFAlbertModelTest(TFModelTesterMixin, unittest.TestCase): ...@@ -49,187 +229,8 @@ class TFAlbertModelTest(TFModelTesterMixin, unittest.TestCase):
else () else ()
) )
class TFAlbertModelTester(object):
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,
embedding_size=16,
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.embedding_size = embedding_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 = AlbertConfig(
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,
initializer_range=self.initializer_range,
)
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def create_and_check_albert_model(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFAlbertModel(config=config)
# inputs = {'input_ids': input_ids,
# 'attention_mask': input_mask,
# 'token_type_ids': token_type_ids}
# sequence_output, pooled_output = model(**inputs)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
sequence_output, pooled_output = model(inputs)
inputs = [input_ids, input_mask]
sequence_output, pooled_output = model(inputs)
sequence_output, pooled_output = model(input_ids)
result = {
"sequence_output": sequence_output.numpy(),
"pooled_output": pooled_output.numpy(),
}
self.parent.assertListEqual(
list(result["sequence_output"].shape), [self.batch_size, self.seq_length, self.hidden_size]
)
self.parent.assertListEqual(list(result["pooled_output"].shape), [self.batch_size, self.hidden_size])
def create_and_check_albert_for_pretraining(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = TFAlbertForPreTraining(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
prediction_scores, sop_scores = model(inputs)
result = {
"prediction_scores": prediction_scores.numpy(),
"sop_scores": sop_scores.numpy(),
}
self.parent.assertListEqual(
list(result["prediction_scores"].shape), [self.batch_size, self.seq_length, self.vocab_size]
)
self.parent.assertListEqual(list(result["sop_scores"].shape), [self.batch_size, self.num_labels])
def create_and_check_albert_for_masked_lm(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFAlbertForMaskedLM(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
(prediction_scores,) = model(inputs)
result = {
"prediction_scores": prediction_scores.numpy(),
}
self.parent.assertListEqual(
list(result["prediction_scores"].shape), [self.batch_size, self.seq_length, self.vocab_size]
)
def create_and_check_albert_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 = TFAlbertForSequenceClassification(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
(logits,) = model(inputs)
result = {
"logits": logits.numpy(),
}
self.parent.assertListEqual(list(result["logits"].shape), [self.batch_size, self.num_labels])
def create_and_check_albert_for_question_answering(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFAlbertForQuestionAnswering(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
start_logits, end_logits = model(inputs)
result = {
"start_logits": start_logits.numpy(),
"end_logits": end_logits.numpy(),
}
self.parent.assertListEqual(list(result["start_logits"].shape), [self.batch_size, self.seq_length])
self.parent.assertListEqual(list(result["end_logits"].shape), [self.batch_size, self.seq_length])
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
def setUp(self): def setUp(self):
self.model_tester = TFAlbertModelTest.TFAlbertModelTester(self) self.model_tester = TFAlbertModelTester(self)
self.config_tester = ConfigTester(self, config_class=AlbertConfig, hidden_size=37) self.config_tester = ConfigTester(self, config_class=AlbertConfig, hidden_size=37)
def test_config(self): def test_config(self):
......
tests/test_modeling_tf_bert.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_tf_ctrl.py
View file @ c852036b
...@@ -28,163 +28,141 @@ if is_tf_available(): ...@@ -28,163 +28,141 @@ if is_tf_available():
from transformers.modeling_tf_ctrl import TFCTRLModel, TFCTRLLMHeadModel, TF_CTRL_PRETRAINED_MODEL_ARCHIVE_LIST from transformers.modeling_tf_ctrl import TFCTRLModel, TFCTRLLMHeadModel, TF_CTRL_PRETRAINED_MODEL_ARCHIVE_LIST
class TFCTRLModelTester(object):
def __init__(
self, parent,
):
self.parent = parent
self.batch_size = 13
self.seq_length = 7
self.is_training = True
self.use_token_type_ids = True
self.use_input_mask = True
self.use_labels = True
self.use_mc_token_ids = True
self.vocab_size = 99
self.hidden_size = 32
self.num_hidden_layers = 5
self.num_attention_heads = 4
self.intermediate_size = 37
self.hidden_act = "gelu"
self.hidden_dropout_prob = 0.1
self.attention_probs_dropout_prob = 0.1
self.max_position_embeddings = 512
self.type_vocab_size = 16
self.type_sequence_label_size = 2
self.initializer_range = 0.02
self.num_labels = 3
self.num_choices = 4
self.scope = None
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)
mc_token_ids = None
if self.use_mc_token_ids:
mc_token_ids = ids_tensor([self.batch_size, self.num_choices], self.seq_length)
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 = CTRLConfig(
vocab_size=self.vocab_size,
n_embd=self.hidden_size,
n_layer=self.num_hidden_layers,
n_head=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,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range
)
head_mask = ids_tensor([self.num_hidden_layers, self.num_attention_heads], 2)
return (
config,
input_ids,
input_mask,
head_mask,
token_type_ids,
mc_token_ids,
sequence_labels,
token_labels,
choice_labels,
)
def create_and_check_ctrl_model(self, config, input_ids, input_mask, head_mask, token_type_ids, *args):
model = TFCTRLModel(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
sequence_output = model(inputs)[0]
inputs = [input_ids, None, input_mask] # None is the input for 'past'
sequence_output = model(inputs)[0]
sequence_output = model(input_ids)[0]
result = {
"sequence_output": sequence_output.numpy(),
}
self.parent.assertListEqual(
list(result["sequence_output"].shape), [self.batch_size, self.seq_length, self.hidden_size]
)
def create_and_check_ctrl_lm_head(self, config, input_ids, input_mask, head_mask, token_type_ids, *args):
model = TFCTRLLMHeadModel(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
prediction_scores = model(inputs)[0]
result = {
"prediction_scores": prediction_scores.numpy(),
}
self.parent.assertListEqual(
list(result["prediction_scores"].shape), [self.batch_size, self.seq_length, self.vocab_size]
)
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(
config,
input_ids,
input_mask,
head_mask,
token_type_ids,
mc_token_ids,
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_tf @require_tf
class TFCTRLModelTest(TFModelTesterMixin, unittest.TestCase): class TFCTRLModelTest(TFModelTesterMixin, unittest.TestCase):
all_model_classes = (TFCTRLModel, TFCTRLLMHeadModel) if is_tf_available() else () all_model_classes = (TFCTRLModel, TFCTRLLMHeadModel) if is_tf_available() else ()
all_generative_model_classes = (TFCTRLLMHeadModel,) if is_tf_available() else () all_generative_model_classes = (TFCTRLLMHeadModel,) if is_tf_available() else ()
class TFCTRLModelTester(object):
def __init__(
self,
parent,
batch_size=13,
seq_length=7,
is_training=True,
use_token_type_ids=True,
use_input_mask=True,
use_labels=True,
use_mc_token_ids=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_token_type_ids = use_token_type_ids
self.use_input_mask = use_input_mask
self.use_labels = use_labels
self.use_mc_token_ids = use_mc_token_ids
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)
mc_token_ids = None
if self.use_mc_token_ids:
mc_token_ids = ids_tensor([self.batch_size, self.num_choices], self.seq_length)
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 = CTRLConfig(
vocab_size=self.vocab_size,
n_embd=self.hidden_size,
n_layer=self.num_hidden_layers,
n_head=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,
n_positions=self.max_position_embeddings,
n_ctx=self.max_position_embeddings
# type_vocab_size=self.type_vocab_size,
# initializer_range=self.initializer_range
)
head_mask = ids_tensor([self.num_hidden_layers, self.num_attention_heads], 2)
return (
config,
input_ids,
input_mask,
head_mask,
token_type_ids,
mc_token_ids,
sequence_labels,
token_labels,
choice_labels,
)
def create_and_check_ctrl_model(self, config, input_ids, input_mask, head_mask, token_type_ids, *args):
model = TFCTRLModel(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
sequence_output = model(inputs)[0]
inputs = [input_ids, None, input_mask] # None is the input for 'past'
sequence_output = model(inputs)[0]
sequence_output = model(input_ids)[0]
result = {
"sequence_output": sequence_output.numpy(),
}
self.parent.assertListEqual(
list(result["sequence_output"].shape), [self.batch_size, self.seq_length, self.hidden_size]
)
def create_and_check_ctrl_lm_head(self, config, input_ids, input_mask, head_mask, token_type_ids, *args):
model = TFCTRLLMHeadModel(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
prediction_scores = model(inputs)[0]
result = {
"prediction_scores": prediction_scores.numpy(),
}
self.parent.assertListEqual(
list(result["prediction_scores"].shape), [self.batch_size, self.seq_length, self.vocab_size]
)
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(
config,
input_ids,
input_mask,
head_mask,
token_type_ids,
mc_token_ids,
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
def setUp(self): def setUp(self):
self.model_tester = TFCTRLModelTest.TFCTRLModelTester(self) self.model_tester = TFCTRLModelTester(self)
self.config_tester = ConfigTester(self, config_class=CTRLConfig, n_embd=37) self.config_tester = ConfigTester(self, config_class=CTRLConfig, n_embd=37)
def test_config(self): def test_config(self):
......
tests/test_modeling_tf_distilbert.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_tf_electra.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_tf_gpt2.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_tf_openai_gpt.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_tf_roberta.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_tf_t5.py
View file @ c852036b
This diff is collapsed.
tests/test_modeling_tf_transfo_xl.py
View file @ c852036b
This diff is collapsed.
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