SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

[Test refactor 1/5] Per-folder tests reorganization (#15725)

Browse Source 
* Per-folder tests reorganization

Co-authored-by: sgugger <sylvain.gugger@gmail.com>
Co-authored-by: Stas Bekman <stas@stason.org>
This commit is contained in:
Lysandre Debut
2022-02-23 15:46:28 -05:00
committed by GitHub
parent fecb08c2b8
commit 29c10a41d0
438 changed files with 636 additions and 565 deletions
Download Patch File Download Diff File Expand all files Collapse all files

0
tests/mpnet/__init__.py Normal file
View File

252
tests/mpnet/test_modeling_mpnet.py Normal file
View File

@@ -0,0 +1,252 @@
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. team, Microsoft Corporation.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import unittest
from transformers import MPNetConfig, is_torch_available
from transformers.testing_utils import require_torch, slow, torch_device
from ..test_configuration_common import ConfigTester
from ..test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
if is_torch_available():
import torch
from transformers import (
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
class MPNetModelTester:
def __init__(
self,
parent,
batch_size=13,
seq_length=7,
is_training=True,
use_input_mask=True,
use_token_type_ids=False,
use_labels=True,
vocab_size=99,
hidden_size=64,
num_hidden_layers=5,
num_attention_heads=4,
intermediate_size=64,
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 get_large_model_config(self):
return MPNetConfig.from_pretrained("microsoft/mpnet-base")
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 = random_attention_mask([self.batch_size, 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 = self.get_config()
return config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
def get_config(self):
return MPNetConfig(
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,
initializer_range=self.initializer_range,
)
def create_and_check_mpnet_model(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = MPNetModel(config=config)
model.to(torch_device)
model.eval()
result = model(input_ids, input_mask)
result = model(input_ids)
self.parent.assertEqual(result.last_hidden_state.shape, (self.batch_size, self.seq_length, self.hidden_size))
self.parent.assertEqual(result.pooler_output.shape, (self.batch_size, self.hidden_size))
def create_and_check_mpnet_for_question_answering(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = MPNetForQuestionAnswering(config=config)
model.to(torch_device)
model.eval()
result = model(
input_ids,
attention_mask=input_mask,
start_positions=sequence_labels,
end_positions=sequence_labels,
)
self.parent.assertEqual(result.start_logits.shape, (self.batch_size, self.seq_length))
self.parent.assertEqual(result.end_logits.shape, (self.batch_size, self.seq_length))
def create_and_check_mpnet_for_sequence_classification(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = MPNetForSequenceClassification(config)
model.to(torch_device)
model.eval()
result = model(input_ids, attention_mask=input_mask, labels=sequence_labels)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.num_labels))
def create_and_check_mpnet_for_multiple_choice(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_choices = self.num_choices
model = MPNetForMultipleChoice(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_input_mask = input_mask.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous()
result = model(
multiple_choice_inputs_ids,
attention_mask=multiple_choice_input_mask,
labels=choice_labels,
)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.num_choices))
def create_and_check_mpnet_for_token_classification(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = MPNetForTokenClassification(config=config)
model.to(torch_device)
model.eval()
result = model(input_ids, attention_mask=input_mask, labels=token_labels)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.seq_length, self.num_labels))
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(config, input_ids, input_mask, sequence_labels, token_labels, choice_labels) = config_and_inputs
inputs_dict = {"input_ids": input_ids, "attention_mask": input_mask}
return config, inputs_dict
@require_torch
class MPNetModelTest(ModelTesterMixin, unittest.TestCase):
all_model_classes = (
(
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
if is_torch_available()
else ()
)
test_pruning = False
test_torchscript = True
test_resize_embeddings = True
def setUp(self):
self.model_tester = MPNetModelTester(self)
self.config_tester = ConfigTester(self, config_class=MPNetConfig, hidden_size=37)
def test_config(self):
self.config_tester.run_common_tests()
def test_mpnet_model(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_model(*config_and_inputs)
def test_for_sequence_classification(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_sequence_classification(*config_and_inputs)
def test_for_multiple_choice(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_multiple_choice(*config_and_inputs)
def test_for_token_classification(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_token_classification(*config_and_inputs)
def test_for_question_answering(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_question_answering(*config_and_inputs)
@require_torch
class MPNetModelIntegrationTest(unittest.TestCase):
@slow
def test_inference_no_head(self):
model = MPNetModel.from_pretrained("microsoft/mpnet-base")
input_ids = torch.tensor([[0, 345, 232, 328, 740, 140, 1695, 69, 6078, 1588, 2]])
output = model(input_ids)[0]
expected_shape = torch.Size((1, 11, 768))
self.assertEqual(output.shape, expected_shape)
expected_slice = torch.tensor(
[[[-0.0550, 0.1943, -0.0740], [-0.0562, 0.2211, -0.0579], [-0.0437, 0.3337, -0.0641]]]
)
# compare the actual values for a slice.
self.assertTrue(torch.allclose(output[:, :3, :3], expected_slice, atol=1e-4))

262
tests/mpnet/test_modeling_tf_mpnet.py Normal file
View File

@@ -0,0 +1,262 @@
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. team, Microsoft Corporation.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import unittest
from transformers import MPNetConfig, is_tf_available
from transformers.testing_utils import require_tf, slow
from ..test_configuration_common import ConfigTester
from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
if is_tf_available():
import tensorflow as tf
from transformers.models.mpnet.modeling_tf_mpnet import (
TFMPNetForMaskedLM,
TFMPNetForMultipleChoice,
TFMPNetForQuestionAnswering,
TFMPNetForSequenceClassification,
TFMPNetForTokenClassification,
TFMPNetModel,
)
class TFMPNetModelTester:
def __init__(
self,
parent,
batch_size=13,
seq_length=7,
is_training=True,
use_input_mask=True,
use_token_type_ids=False,
use_labels=True,
vocab_size=99,
hidden_size=64,
num_hidden_layers=5,
num_attention_heads=4,
intermediate_size=64,
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)
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 = MPNetConfig(
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,
initializer_range=self.initializer_range,
)
return config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
def create_and_check_mpnet_model(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFMPNetModel(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask}
result = model(inputs)
inputs = [input_ids, input_mask]
result = model(inputs)
self.parent.assertEqual(result.last_hidden_state.shape, (self.batch_size, self.seq_length, self.hidden_size))
def create_and_check_mpnet_for_masked_lm(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFMPNetForMaskedLM(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask}
result = model(inputs)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.seq_length, self.vocab_size))
def create_and_check_mpnet_for_question_answering(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFMPNetForQuestionAnswering(config=config)
inputs = {
"input_ids": input_ids,
"attention_mask": input_mask,
}
result = model(inputs)
self.parent.assertEqual(result.start_logits.shape, (self.batch_size, self.seq_length))
self.parent.assertEqual(result.end_logits.shape, (self.batch_size, self.seq_length))
def create_and_check_mpnet_for_sequence_classification(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = TFMPNetForSequenceClassification(config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask}
result = model(inputs)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.num_labels))
def create_and_check_mpnet_for_multiple_choice(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_choices = self.num_choices
model = TFMPNetForMultipleChoice(config)
multiple_choice_inputs_ids = tf.tile(tf.expand_dims(input_ids, 1), (1, self.num_choices, 1))
multiple_choice_input_mask = tf.tile(tf.expand_dims(input_mask, 1), (1, self.num_choices, 1))
inputs = {
"input_ids": multiple_choice_inputs_ids,
"attention_mask": multiple_choice_input_mask,
}
result = model(inputs)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.num_choices))
def create_and_check_mpnet_for_token_classification(
self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = TFMPNetForTokenClassification(config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask}
result = model(inputs)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.seq_length, self.num_labels))
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(config, input_ids, input_mask, sequence_labels, token_labels, choice_labels) = config_and_inputs
inputs_dict = {"input_ids": input_ids, "attention_mask": input_mask}
return config, inputs_dict
@require_tf
class TFMPNetModelTest(TFModelTesterMixin, unittest.TestCase):
all_model_classes = (
(
TFMPNetForMaskedLM,
TFMPNetForMultipleChoice,
TFMPNetForQuestionAnswering,
TFMPNetForSequenceClassification,
TFMPNetForTokenClassification,
TFMPNetModel,
)
if is_tf_available()
else ()
)
test_head_masking = False
test_onnx = False
def setUp(self):
self.model_tester = TFMPNetModelTester(self)
self.config_tester = ConfigTester(self, config_class=MPNetConfig, hidden_size=37)
def test_config(self):
self.config_tester.run_common_tests()
def test_mpnet_model(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_model(*config_and_inputs)
def test_for_masked_lm(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_masked_lm(*config_and_inputs)
def test_for_question_answering(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_question_answering(*config_and_inputs)
def test_for_sequence_classification(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_sequence_classification(*config_and_inputs)
def test_for_multiple_choice(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_multiple_choice(*config_and_inputs)
def test_for_token_classification(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_token_classification(*config_and_inputs)
@slow
def test_model_from_pretrained(self):
for model_name in ["microsoft/mpnet-base"]:
model = TFMPNetModel.from_pretrained(model_name)
self.assertIsNotNone(model)
@require_tf
class TFMPNetModelIntegrationTest(unittest.TestCase):
@slow
def test_inference_masked_lm(self):
model = TFMPNetModel.from_pretrained("microsoft/mpnet-base")
input_ids = tf.constant([[0, 1, 2, 3, 4, 5]])
output = model(input_ids)[0]
expected_shape = [1, 6, 768]
self.assertEqual(output.shape, expected_shape)
expected_slice = tf.constant(
[
[
[-0.1067172, 0.08216473, 0.0024543],
[-0.03465879, 0.8354118, -0.03252288],
[-0.06569476, -0.12424111, -0.0494436],
]
]
)
tf.debugging.assert_near(output[:, :3, :3], expected_slice, atol=1e-4)

82
tests/mpnet/test_tokenization_mpnet.py Normal file
View File

@@ -0,0 +1,82 @@
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. team, Microsoft Corporation.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import os
import unittest
from transformers import MPNetTokenizerFast
from transformers.models.mpnet.tokenization_mpnet import VOCAB_FILES_NAMES, MPNetTokenizer
from transformers.testing_utils import require_tokenizers, slow
from ..test_tokenization_common import TokenizerTesterMixin
@require_tokenizers
class MPNetTokenizerTest(TokenizerTesterMixin, unittest.TestCase):
tokenizer_class = MPNetTokenizer
rust_tokenizer_class = MPNetTokenizerFast
test_rust_tokenizer = True
space_between_special_tokens = True
def setUp(self):
super().setUp()
vocab_tokens = [
"[UNK]",
"[CLS]",
"[SEP]",
"[PAD]",
"[MASK]",
"want",
"##want",
"##ed",
"wa",
"un",
"runn",
"##ing",
",",
"low",
"lowest",
]
self.vocab_file = os.path.join(self.tmpdirname, VOCAB_FILES_NAMES["vocab_file"])
with open(self.vocab_file, "w", encoding="utf-8") as vocab_writer:
vocab_writer.write("".join([x + "\n" for x in vocab_tokens]))
def get_input_output_texts(self, tokenizer):
input_text = "UNwant\u00E9d,running"
output_text = "unwanted, running"
return input_text, output_text
def test_full_tokenizer(self):
tokenizer = self.tokenizer_class(self.vocab_file)
tokens = tokenizer.tokenize("UNwant\u00E9d,running")
self.assertListEqual(tokens, ["un", "##want", "##ed", ",", "runn", "##ing"])
self.assertListEqual(tokenizer.convert_tokens_to_ids(tokens), [9, 6, 7, 12, 10, 11])
@slow
def test_sequence_builders(self):
tokenizer = self.tokenizer_class.from_pretrained("microsoft/mpnet-base")
text = tokenizer.encode("sequence builders", add_special_tokens=False)
text_2 = tokenizer.encode("multi-sequence build", add_special_tokens=False)
encoded_sentence = tokenizer.build_inputs_with_special_tokens(text)
encoded_pair = tokenizer.build_inputs_with_special_tokens(text, text_2)
assert encoded_sentence == [0] + text + [2]
assert encoded_pair == [0] + text + [2] + [2] + text_2 + [2]