rust_transformers 0.2.0

High performance tokenizers for Rust
Documentation 
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

# Copyright 2018 The HuggingFace Inc. team.
# Copyright 2019 Guillaume Becquin
# 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 tempfile
from pathlib import Path
import pytest
from transformers.data.processors.glue import Sst2Processor
from transformers.file_utils import get_from_cache
from transformers.tokenization_bert import BertTokenizer
from transformers.tokenization_distilbert import DistilBertTokenizer
from transformers.tokenization_ctrl import CTRLTokenizer
from transformers.tokenization_gpt2 import GPT2Tokenizer
from transformers.tokenization_roberta import RobertaTokenizer
from transformers.tokenization_openai import OpenAIGPTTokenizer
from rust_transformers import PyBertTokenizer, PyCtrlTokenizer, PyGpt2Tokenizer, PyRobertaTokenizer, \

    PyOpenAiGptTokenizer
import os


@pytest.mark.slow
class TestTokenizationSST2:
    def setup_class(self):
        self.processor = Sst2Processor()
        # Note: these tests do not download automatically test datasets. Please download them manually and update your
        # environment variables accordingly
        self.examples = self.processor.get_train_examples(os.environ["SST2_PATH"])
        self.test_dir = Path(tempfile.mkdtemp())

    def test_tokenization_bert(self):
        # Given
        self.base_tokenizer = BertTokenizer.from_pretrained('bert-base-uncased', do_lower_case=True,
                                                            cache_dir=self.test_dir)
        self.rust_tokenizer = PyBertTokenizer(
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['vocab_file']['bert-base-uncased']))
        output_baseline = []
        for example in self.examples:
            output_baseline.append(self.base_tokenizer.encode_plus(example.text_a,
                                                                   add_special_tokens=True,
                                                                   return_overflowing_tokens=True,
                                                                   return_special_tokens_mask=True,
                                                                   max_length=128))

        # When
        output_rust = self.rust_tokenizer.encode_list([example.text_a for example in self.examples],
                                                      max_len=128,
                                                      truncation_strategy='longest_first',
                                                      stride=0)

        # Then
        for rust, baseline in zip(output_rust, output_baseline):
            assert (rust.token_ids == baseline['input_ids'])
            assert (rust.segment_ids == baseline['token_type_ids'])
            assert (rust.special_tokens_mask == baseline['special_tokens_mask'])

    def test_tokenization_distilbert(self):
        # Given
        self.base_tokenizer = DistilBertTokenizer.from_pretrained('distilbert-base-uncased', do_lower_case=True,
                                                                  cache_dir=self.test_dir)
        self.rust_tokenizer = PyBertTokenizer(
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['vocab_file']['distilbert-base-uncased']))
        output_baseline = []
        for example in self.examples:
            output_baseline.append(self.base_tokenizer.encode_plus(example.text_a,
                                                                   add_special_tokens=True,
                                                                   return_overflowing_tokens=True,
                                                                   return_special_tokens_mask=True,
                                                                   max_length=128))

        # When
        output_rust = self.rust_tokenizer.encode_list([example.text_a for example in self.examples],
                                                      max_len=128,
                                                      truncation_strategy='longest_first',
                                                      stride=0)

        # Then
        for rust, baseline in zip(output_rust, output_baseline):
            assert (rust.token_ids == baseline['input_ids'])
            assert (rust.segment_ids == baseline['token_type_ids'])
            assert (rust.special_tokens_mask == baseline['special_tokens_mask'])

    def test_tokenization_ctrl(self):
        # Given
        self.base_tokenizer = CTRLTokenizer.from_pretrained('ctrl', do_lower_case=True,
                                                            cache_dir=self.test_dir)
        self.rust_tokenizer = PyCtrlTokenizer(
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['vocab_file']['ctrl']),
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['merges_file']['ctrl'])
        )
        output_baseline = []
        for example in self.examples:
            output_baseline.append(self.base_tokenizer.encode_plus(example.text_a,
                                                                   add_special_tokens=True,
                                                                   return_overflowing_tokens=True,
                                                                   return_special_tokens_mask=True,
                                                                   max_length=128))

        # When
        output_rust = self.rust_tokenizer.encode_list([example.text_a for example in self.examples],
                                                      max_len=128,
                                                      truncation_strategy='longest_first',
                                                      stride=0)

        # Then
        for rust, baseline in zip(output_rust, output_baseline):
            assert (rust.token_ids == baseline['input_ids'])
            assert (rust.segment_ids == baseline['token_type_ids'])
            assert (rust.special_tokens_mask == baseline['special_tokens_mask'])

    def test_tokenization_gpt2(self):
        # Given
        self.base_tokenizer = GPT2Tokenizer.from_pretrained('gpt2', do_lower_case=True,
                                                            cache_dir=self.test_dir)
        self.rust_tokenizer = PyGpt2Tokenizer(
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['vocab_file']['gpt2']),
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['merges_file']['gpt2'])
        )
        output_baseline = []
        for example in self.examples:
            output_baseline.append(self.base_tokenizer.encode_plus(example.text_a,
                                                                   add_special_tokens=True,
                                                                   return_overflowing_tokens=True,
                                                                   return_special_tokens_mask=True,
                                                                   max_length=128))

        # When
        output_rust = self.rust_tokenizer.encode_list([example.text_a for example in self.examples],
                                                      max_len=128,
                                                      truncation_strategy='longest_first',
                                                      stride=0)

        # Then
        for rust, baseline in zip(output_rust, output_baseline):
            assert (rust.token_ids == baseline['input_ids'])
            assert (rust.segment_ids == baseline['token_type_ids'])
            assert (rust.special_tokens_mask == baseline['special_tokens_mask'])

    def test_tokenization_roberta(self):
        # Given
        self.base_tokenizer = RobertaTokenizer.from_pretrained('roberta-base', do_lower_case=True,
                                                               cache_dir=self.test_dir)
        self.rust_tokenizer = PyRobertaTokenizer(
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['vocab_file']['roberta-base']),
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['merges_file']['roberta-base'])
        )
        output_baseline = []
        for example in self.examples:
            output_baseline.append(self.base_tokenizer.encode_plus(example.text_a,
                                                                   add_special_tokens=True,
                                                                   return_overflowing_tokens=True,
                                                                   return_special_tokens_mask=True,
                                                                   max_length=128))

        # When
        output_rust = self.rust_tokenizer.encode_list([example.text_a for example in self.examples],
                                                      max_len=128,
                                                      truncation_strategy='longest_first',
                                                      stride=0)

        # Then
        for rust, baseline in zip(output_rust, output_baseline):
            assert (rust.token_ids == baseline['input_ids'])
            assert (rust.segment_ids == baseline['token_type_ids'])
            assert (rust.special_tokens_mask == baseline['special_tokens_mask'])

    def test_tokenization_openai_gpt(self):
        # Given
        self.base_tokenizer = OpenAIGPTTokenizer.from_pretrained('openai-gpt', do_lower_case=True,
                                                                 cache_dir=self.test_dir)
        self.rust_tokenizer = PyOpenAiGptTokenizer(
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['vocab_file']['openai-gpt']),
            get_from_cache(self.base_tokenizer.pretrained_vocab_files_map['merges_file']['openai-gpt'])
        )
        output_baseline = []
        for example in self.examples:
            output_baseline.append(self.base_tokenizer.encode_plus(example.text_a,
                                                                   add_special_tokens=True,
                                                                   return_overflowing_tokens=True,
                                                                   return_special_tokens_mask=True,
                                                                   max_length=128))

        # When
        output_rust = self.rust_tokenizer.encode_list([example.text_a for example in self.examples],
                                                      max_len=128,
                                                      truncation_strategy='longest_first',
                                                      stride=0)

        # Then
        for rust, baseline in zip(output_rust, output_baseline):
            assert (rust.token_ids == baseline['input_ids'])
            assert (rust.segment_ids == baseline['token_type_ids'])
            assert (rust.special_tokens_mask == baseline['special_tokens_mask'])