File size: 10,390 Bytes
455a40f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
# coding=utf-8
# Copyright 2020 The HuggingFace Team. All rights reserved.
#
# 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 DistilBertConfig, 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, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin


if is_tf_available():
    import tensorflow as tf

    from transformers.models.distilbert.modeling_tf_distilbert import (
        TF_DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
        TFDistilBertForMaskedLM,
        TFDistilBertForMultipleChoice,
        TFDistilBertForQuestionAnswering,
        TFDistilBertForSequenceClassification,
        TFDistilBertForTokenClassification,
        TFDistilBertModel,
    )


class TFDistilBertModelTester:
    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 = False
        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 = 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 = DistilBertConfig(
            vocab_size=self.vocab_size,
            dim=self.hidden_size,
            n_layers=self.num_hidden_layers,
            n_heads=self.num_attention_heads,
            hidden_dim=self.intermediate_size,
            hidden_act=self.hidden_act,
            dropout=self.hidden_dropout_prob,
            attention_dropout=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_distilbert_model(
        self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
    ):
        model = TFDistilBertModel(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_distilbert_for_masked_lm(
        self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
    ):
        model = TFDistilBertForMaskedLM(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_distilbert_for_question_answering(
        self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
    ):
        model = TFDistilBertForQuestionAnswering(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_distilbert_for_sequence_classification(
        self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
    ):
        config.num_labels = self.num_labels
        model = TFDistilBertForSequenceClassification(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_distilbert_for_multiple_choice(
        self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
    ):
        config.num_choices = self.num_choices
        model = TFDistilBertForMultipleChoice(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_distilbert_for_token_classification(
        self, config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
    ):
        config.num_labels = self.num_labels
        model = TFDistilBertForTokenClassification(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 TFDistilBertModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
    all_model_classes = (
        (
            TFDistilBertModel,
            TFDistilBertForMaskedLM,
            TFDistilBertForQuestionAnswering,
            TFDistilBertForSequenceClassification,
            TFDistilBertForTokenClassification,
            TFDistilBertForMultipleChoice,
        )
        if is_tf_available()
        else None
    )
    pipeline_model_mapping = (
        {
            "feature-extraction": TFDistilBertModel,
            "fill-mask": TFDistilBertForMaskedLM,
            "question-answering": TFDistilBertForQuestionAnswering,
            "text-classification": TFDistilBertForSequenceClassification,
            "token-classification": TFDistilBertForTokenClassification,
            "zero-shot": TFDistilBertForSequenceClassification,
        }
        if is_tf_available()
        else {}
    )
    test_head_masking = False
    test_onnx = False

    def setUp(self):
        self.model_tester = TFDistilBertModelTester(self)
        self.config_tester = ConfigTester(self, config_class=DistilBertConfig, dim=37)

    def test_config(self):
        self.config_tester.run_common_tests()

    def test_distilbert_model(self):
        config_and_inputs = self.model_tester.prepare_config_and_inputs()
        self.model_tester.create_and_check_distilbert_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_distilbert_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_distilbert_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_distilbert_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_distilbert_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_distilbert_for_token_classification(*config_and_inputs)

    @slow
    def test_model_from_pretrained(self):
        for model_name in list(TF_DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]):
            model = TFDistilBertModel.from_pretrained(model_name)
            self.assertIsNotNone(model)


@require_tf
class TFDistilBertModelIntegrationTest(unittest.TestCase):
    @slow
    def test_inference_masked_lm(self):
        model = TFDistilBertModel.from_pretrained("distilbert-base-uncased")
        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(
            [
                [
                    [0.19261885, -0.13732955, 0.4119799],
                    [0.22150156, -0.07422661, 0.39037204],
                    [0.22756018, -0.0896414, 0.3701467],
                ]
            ]
        )
        tf.debugging.assert_near(output[:, :3, :3], expected_slice, atol=1e-4)