File size: 7,123 Bytes
ee6e328
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
# 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

import pytest

from transformers import (
    MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING,
    TF_MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING,
    MBart50TokenizerFast,
    MBartConfig,
    MBartForConditionalGeneration,
    TranslationPipeline,
    pipeline,
)
from transformers.testing_utils import is_pipeline_test, require_tf, require_torch, slow

from .test_pipelines_common import ANY


@is_pipeline_test
class TranslationPipelineTests(unittest.TestCase):
    model_mapping = MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING
    tf_model_mapping = TF_MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING

    def get_test_pipeline(self, model, tokenizer, processor):
        if isinstance(model.config, MBartConfig):
            src_lang, tgt_lang = list(tokenizer.lang_code_to_id.keys())[:2]
            translator = TranslationPipeline(model=model, tokenizer=tokenizer, src_lang=src_lang, tgt_lang=tgt_lang)
        else:
            translator = TranslationPipeline(model=model, tokenizer=tokenizer)
        return translator, ["Some string", "Some other text"]

    def run_pipeline_test(self, translator, _):
        outputs = translator("Some string")
        self.assertEqual(outputs, [{"translation_text": ANY(str)}])

        outputs = translator(["Some string"])
        self.assertEqual(outputs, [{"translation_text": ANY(str)}])

        outputs = translator(["Some string", "other string"])
        self.assertEqual(outputs, [{"translation_text": ANY(str)}, {"translation_text": ANY(str)}])

    @require_torch
    def test_small_model_pt(self):
        translator = pipeline("translation_en_to_ro", model="patrickvonplaten/t5-tiny-random", framework="pt")
        outputs = translator("This is a test string", max_length=20)
        self.assertEqual(
            outputs,
            [
                {
                    "translation_text": (
                        "Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide"
                        " Beide Beide"
                    )
                }
            ],
        )

    @require_tf
    def test_small_model_tf(self):
        translator = pipeline("translation_en_to_ro", model="patrickvonplaten/t5-tiny-random", framework="tf")
        outputs = translator("This is a test string", max_length=20)
        self.assertEqual(
            outputs,
            [
                {
                    "translation_text": (
                        "Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide Beide"
                        " Beide Beide"
                    )
                }
            ],
        )

    @require_torch
    def test_en_to_de_pt(self):
        translator = pipeline("translation_en_to_de", model="patrickvonplaten/t5-tiny-random", framework="pt")
        outputs = translator("This is a test string", max_length=20)
        self.assertEqual(
            outputs,
            [
                {
                    "translation_text": (
                        "monoton monoton monoton monoton monoton monoton monoton monoton monoton monoton urine urine"
                        " urine urine urine urine urine urine urine"
                    )
                }
            ],
        )

    @require_tf
    def test_en_to_de_tf(self):
        translator = pipeline("translation_en_to_de", model="patrickvonplaten/t5-tiny-random", framework="tf")
        outputs = translator("This is a test string", max_length=20)
        self.assertEqual(
            outputs,
            [
                {
                    "translation_text": (
                        "monoton monoton monoton monoton monoton monoton monoton monoton monoton monoton urine urine"
                        " urine urine urine urine urine urine urine"
                    )
                }
            ],
        )


class TranslationNewFormatPipelineTests(unittest.TestCase):
    @require_torch
    @slow
    def test_default_translations(self):
        # We don't provide a default for this pair
        with self.assertRaises(ValueError):
            pipeline(task="translation_cn_to_ar")

        # but we do for this one
        translator = pipeline(task="translation_en_to_de")
        self.assertEqual(translator._preprocess_params["src_lang"], "en")
        self.assertEqual(translator._preprocess_params["tgt_lang"], "de")

    @require_torch
    @slow
    def test_multilingual_translation(self):
        model = MBartForConditionalGeneration.from_pretrained("facebook/mbart-large-50-many-to-many-mmt")
        tokenizer = MBart50TokenizerFast.from_pretrained("facebook/mbart-large-50-many-to-many-mmt")

        translator = pipeline(task="translation", model=model, tokenizer=tokenizer)
        # Missing src_lang, tgt_lang
        with self.assertRaises(ValueError):
            translator("This is a test")

        outputs = translator("This is a test", src_lang="en_XX", tgt_lang="ar_AR")
        self.assertEqual(outputs, [{"translation_text": "هذا إختبار"}])

        outputs = translator("This is a test", src_lang="en_XX", tgt_lang="hi_IN")
        self.assertEqual(outputs, [{"translation_text": "यह एक परीक्षण है"}])

        # src_lang, tgt_lang can be defined at pipeline call time
        translator = pipeline(task="translation", model=model, tokenizer=tokenizer, src_lang="en_XX", tgt_lang="ar_AR")
        outputs = translator("This is a test")
        self.assertEqual(outputs, [{"translation_text": "هذا إختبار"}])

    @require_torch
    def test_translation_on_odd_language(self):
        model = "patrickvonplaten/t5-tiny-random"
        translator = pipeline(task="translation_cn_to_ar", model=model)
        self.assertEqual(translator._preprocess_params["src_lang"], "cn")
        self.assertEqual(translator._preprocess_params["tgt_lang"], "ar")

    @require_torch
    def test_translation_default_language_selection(self):
        model = "patrickvonplaten/t5-tiny-random"
        with pytest.warns(UserWarning, match=r".*translation_en_to_de.*"):
            translator = pipeline(task="translation", model=model)
        self.assertEqual(translator.task, "translation_en_to_de")
        self.assertEqual(translator._preprocess_params["src_lang"], "en")
        self.assertEqual(translator._preprocess_params["tgt_lang"], "de")

    @require_torch
    def test_translation_with_no_language_no_model_fails(self):
        with self.assertRaises(ValueError):
            pipeline(task="translation")