[tests|tokenizers] Refactoring pipelines test backbone - Small tokenizers...

[tests|tokenizers] Refactoring pipelines test backbone - Small tokenizers improvements - General tests speedups (#7970)

* WIP refactoring pipeline tests - switching to fast tokenizers

* fix dialog pipeline and fill-mask

* refactoring pipeline tests backbone

* make large tests slow

* fix tests (tf Bart inactive for now)

* fix doc...

* clean up for merge

* fixing tests - remove bart from summarization until there is TF

* fix quality and RAG

* Add new translation pipeline tests - fix JAX tests

* only slow for dialog

* Fixing the missing TF-BART imports in modeling_tf_auto

* spin out pipeline tests in separate CI job

* adding pipeline test to CI YAML

* add slow pipeline tests

* speed up tf and pt join test to avoid redoing all the standalone pt and tf tests

* Update src/transformers/tokenization_utils_base.py
Co-authored-by: Sam Shleifer <sshleifer@gmail.com>

* Update src/transformers/pipelines.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/pipelines.py
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

* Update src/transformers/testing_utils.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* add require_torch and require_tf in is_pt_tf_cross_test
Co-authored-by: Sam Shleifer <sshleifer@gmail.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

tests/test_pipelines_dialog.py

+import unittest
+
+from transformers.pipelines import Conversation, Pipeline
+
+from .test_pipelines_common import CustomInputPipelineCommonMixin
+
+
+class DialoguePipelineTests(CustomInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "conversational"
+    small_models = []  # Default model - Models tested without the @slow decorator
+    large_models = ["microsoft/DialoGPT-medium"]  # Models tested with the @slow decorator
+
+    def _test_pipeline(self, nlp: Pipeline):
+        valid_inputs = [Conversation("Hi there!"), [Conversation("Hi there!"), Conversation("How are you?")]]
+        invalid_inputs = ["Hi there!", Conversation()]
+        self.assertIsNotNone(nlp)
+
+        mono_result = nlp(valid_inputs[0])
+        self.assertIsInstance(mono_result, Conversation)
+
+        multi_result = nlp(valid_inputs[1])
+        self.assertIsInstance(multi_result, list)
+        self.assertIsInstance(multi_result[0], Conversation)
+        # Inactive conversations passed to the pipeline raise a ValueError
+        self.assertRaises(ValueError, nlp, valid_inputs[1])
+
+        for bad_input in invalid_inputs:
+            self.assertRaises(Exception, nlp, bad_input)
+        self.assertRaises(Exception, nlp, invalid_inputs)

tests/test_pipelines_feature_extraction.py

+import unittest
+
+from .test_pipelines_common import MonoInputPipelineCommonMixin
+
+
+class FeatureExtractionPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "feature-extraction"
+    small_models = [
+        "sshleifer/tiny-distilbert-base-cased"
+    ]  # Default model - Models tested without the @slow decorator
+    large_models = [None]  # Models tested with the @slow decorator
+    mandatory_keys = {}  # Keys which should be in the output

tests/test_pipelines_fill_mask.py

+import unittest
+
+from transformers import pipeline
+from transformers.testing_utils import require_tf, require_torch, slow
+
+from .test_pipelines_common import MonoInputPipelineCommonMixin
+
+
+EXPECTED_FILL_MASK_RESULT = [
+    [
+        {"sequence": "<s>My name is John</s>", "score": 0.00782308354973793, "token": 610, "token_str": "ĠJohn"},
+        {"sequence": "<s>My name is Chris</s>", "score": 0.007475061342120171, "token": 1573, "token_str": "ĠChris"},
+    ],
+    [
+        {"sequence": "<s>The largest city in France is Paris</s>", "score": 0.3185044229030609, "token": 2201},
+        {"sequence": "<s>The largest city in France is Lyon</s>", "score": 0.21112334728240967, "token": 12790},
+    ],
+]
+
+EXPECTED_FILL_MASK_TARGET_RESULT = [
+    [
+        {
+            "sequence": "<s>My name is Patrick</s>",
+            "score": 0.004992353264242411,
+            "token": 3499,
+            "token_str": "ĠPatrick",
+        },
+        {
+            "sequence": "<s>My name is Clara</s>",
+            "score": 0.00019297805556561798,
+            "token": 13606,
+            "token_str": "ĠClara",
+        },
+    ]
+]
+
+
+class FillMaskPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "fill-mask"
+    pipeline_loading_kwargs = {"topk": 2}
+    small_models = ["sshleifer/tiny-distilroberta-base"]  # Models tested without the @slow decorator
+    large_models = ["distilroberta-base"]  # Models tested with the @slow decorator
+    mandatory_keys = {"sequence", "score", "token"}
+    valid_inputs = [
+        "My name is <mask>",
+        "The largest city in France is <mask>",
+    ]
+    invalid_inputs = [
+        "This is <mask> <mask>"  # More than 1 mask_token in the input is not supported
+        "This is"  # No mask_token is not supported
+    ]
+    expected_check_keys = ["sequence"]
+
+    @require_torch
+    def test_torch_fill_mask_with_targets(self):
+        valid_inputs = ["My name is <mask>"]
+        valid_targets = [[" Teven", " Patrick", " Clara"], [" Sam"]]
+        invalid_targets = [[], [""], ""]
+        for model_name in self.small_models:
+            nlp = pipeline(task="fill-mask", model=model_name, tokenizer=model_name, framework="pt")
+            for targets in valid_targets:
+                outputs = nlp(valid_inputs, targets=targets)
+                self.assertIsInstance(outputs, list)
+                self.assertEqual(len(outputs), len(targets))
+            for targets in invalid_targets:
+                self.assertRaises(ValueError, nlp, valid_inputs, targets=targets)
+
+    @require_tf
+    def test_tf_fill_mask_with_targets(self):
+        valid_inputs = ["My name is <mask>"]
+        valid_targets = [[" Teven", " Patrick", " Clara"], [" Sam"]]
+        invalid_targets = [[], [""], ""]
+        for model_name in self.small_models:
+            nlp = pipeline(task="fill-mask", model=model_name, tokenizer=model_name, framework="tf")
+            for targets in valid_targets:
+                outputs = nlp(valid_inputs, targets=targets)
+                self.assertIsInstance(outputs, list)
+                self.assertEqual(len(outputs), len(targets))
+            for targets in invalid_targets:
+                self.assertRaises(ValueError, nlp, valid_inputs, targets=targets)
+
+    @require_torch
+    @slow
+    def test_torch_fill_mask_results(self):
+        mandatory_keys = {"sequence", "score", "token"}
+        valid_inputs = [
+            "My name is <mask>",
+            "The largest city in France is <mask>",
+        ]
+        valid_targets = [" Patrick", " Clara"]
+        for model_name in self.large_models:
+            nlp = pipeline(
+                task="fill-mask",
+                model=model_name,
+                tokenizer=model_name,
+                framework="pt",
+                topk=2,
+            )
+            self._test_mono_column_pipeline(
+                nlp,
+                valid_inputs,
+                mandatory_keys,
+                expected_multi_result=EXPECTED_FILL_MASK_RESULT,
+                expected_check_keys=["sequence"],
+            )
+            self._test_mono_column_pipeline(
+                nlp,
+                valid_inputs[:1],
+                mandatory_keys,
+                expected_multi_result=EXPECTED_FILL_MASK_TARGET_RESULT,
+                expected_check_keys=["sequence"],
+                targets=valid_targets,
+            )
+
+    @require_tf
+    @slow
+    def test_tf_fill_mask_results(self):
+        mandatory_keys = {"sequence", "score", "token"}
+        valid_inputs = [
+            "My name is <mask>",
+            "The largest city in France is <mask>",
+        ]
+        valid_targets = [" Patrick", " Clara"]
+        for model_name in self.large_models:
+            nlp = pipeline(task="fill-mask", model=model_name, tokenizer=model_name, framework="tf", topk=2)
+            self._test_mono_column_pipeline(
+                nlp,
+                valid_inputs,
+                mandatory_keys,
+                expected_multi_result=EXPECTED_FILL_MASK_RESULT,
+                expected_check_keys=["sequence"],
+            )
+            self._test_mono_column_pipeline(
+                nlp,
+                valid_inputs[:1],
+                mandatory_keys,
+                expected_multi_result=EXPECTED_FILL_MASK_TARGET_RESULT,
+                expected_check_keys=["sequence"],
+                targets=valid_targets,
+            )

tests/test_pipelines_ner.py

+import unittest
+
+from transformers import pipeline
+from transformers.pipelines import Pipeline
+from transformers.testing_utils import require_tf
+
+from .test_pipelines_common import CustomInputPipelineCommonMixin
+
+
+VALID_INPUTS = ["A simple string", ["list of strings"]]
+
+
+class NerPipelineTests(CustomInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "ner"
+    small_models = [
+        "sshleifer/tiny-dbmdz-bert-large-cased-finetuned-conll03-english"
+    ]  # Default model - Models tested without the @slow decorator
+    large_models = []  # Models tested with the @slow decorator
+
+    def _test_pipeline(self, nlp: Pipeline):
+        output_keys = {"entity", "word", "score"}
+
+        ungrouped_ner_inputs = [
+            [
+                {"entity": "B-PER", "index": 1, "score": 0.9994944930076599, "word": "Cons"},
+                {"entity": "B-PER", "index": 2, "score": 0.8025449514389038, "word": "##uelo"},
+                {"entity": "I-PER", "index": 3, "score": 0.9993102550506592, "word": "Ara"},
+                {"entity": "I-PER", "index": 4, "score": 0.9993743896484375, "word": "##új"},
+                {"entity": "I-PER", "index": 5, "score": 0.9992871880531311, "word": "##o"},
+                {"entity": "I-PER", "index": 6, "score": 0.9993029236793518, "word": "No"},
+                {"entity": "I-PER", "index": 7, "score": 0.9981776475906372, "word": "##guera"},
+                {"entity": "B-PER", "index": 15, "score": 0.9998136162757874, "word": "Andrés"},
+                {"entity": "I-PER", "index": 16, "score": 0.999740719795227, "word": "Pas"},
+                {"entity": "I-PER", "index": 17, "score": 0.9997414350509644, "word": "##tran"},
+                {"entity": "I-PER", "index": 18, "score": 0.9996136426925659, "word": "##a"},
+                {"entity": "B-ORG", "index": 28, "score": 0.9989739060401917, "word": "Far"},
+                {"entity": "I-ORG", "index": 29, "score": 0.7188422083854675, "word": "##c"},
+            ],
+            [
+                {"entity": "I-PER", "index": 1, "score": 0.9968166351318359, "word": "En"},
+                {"entity": "I-PER", "index": 2, "score": 0.9957635998725891, "word": "##zo"},
+                {"entity": "I-ORG", "index": 7, "score": 0.9986497163772583, "word": "UN"},
+            ],
+        ]
+        expected_grouped_ner_results = [
+            [
+                {"entity_group": "B-PER", "score": 0.9710702640669686, "word": "Consuelo Araújo Noguera"},
+                {"entity_group": "B-PER", "score": 0.9997273534536362, "word": "Andrés Pastrana"},
+                {"entity_group": "B-ORG", "score": 0.8589080572128296, "word": "Farc"},
+            ],
+            [
+                {"entity_group": "I-PER", "score": 0.9962901175022125, "word": "Enzo"},
+                {"entity_group": "I-ORG", "score": 0.9986497163772583, "word": "UN"},
+            ],
+        ]
+
+        self.assertIsNotNone(nlp)
+
+        mono_result = nlp(VALID_INPUTS[0])
+        self.assertIsInstance(mono_result, list)
+        self.assertIsInstance(mono_result[0], (dict, list))
+
+        if isinstance(mono_result[0], list):
+            mono_result = mono_result[0]
+
+        for key in output_keys:
+            self.assertIn(key, mono_result[0])
+
+        multi_result = [nlp(input) for input in VALID_INPUTS]
+        self.assertIsInstance(multi_result, list)
+        self.assertIsInstance(multi_result[0], (dict, list))
+
+        if isinstance(multi_result[0], list):
+            multi_result = multi_result[0]
+
+        for result in multi_result:
+            for key in output_keys:
+                self.assertIn(key, result)
+
+        for ungrouped_input, grouped_result in zip(ungrouped_ner_inputs, expected_grouped_ner_results):
+            self.assertEqual(nlp.group_entities(ungrouped_input), grouped_result)
+
+    @require_tf
+    def test_tf_only(self):
+        model_name = "Narsil/small"  # This model only has a TensorFlow version
+        # We test that if we don't specificy framework='tf', it gets detected automatically
+        nlp = pipeline(task="ner", model=model_name, tokenizer=model_name)
+        self._test_pipeline(nlp)

tests/test_pipelines_question_answering.py

+import unittest
+
+from transformers.pipelines import Pipeline
+
+from .test_pipelines_common import CustomInputPipelineCommonMixin
+
+
+class QAPipelineTests(CustomInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "question-answering"
+    small_models = [
+        "sshleifer/tiny-distilbert-base-cased-distilled-squad"
+    ]  # Models tested without the @slow decorator
+    large_models = []  # Models tested with the @slow decorator
+
+    def _test_pipeline(self, nlp: Pipeline):
+        output_keys = {"score", "answer", "start", "end"}
+        valid_inputs = [
+            {"question": "Where was HuggingFace founded ?", "context": "HuggingFace was founded in Paris."},
+            {
+                "question": "In what field is HuggingFace working ?",
+                "context": "HuggingFace is a startup based in New-York founded in Paris which is trying to solve NLP.",
+            },
+        ]
+        invalid_inputs = [
+            {"question": "", "context": "This is a test to try empty question edge case"},
+            {"question": None, "context": "This is a test to try empty question edge case"},
+            {"question": "What is does with empty context ?", "context": ""},
+            {"question": "What is does with empty context ?", "context": None},
+        ]
+        self.assertIsNotNone(nlp)
+
+        mono_result = nlp(valid_inputs[0])
+        self.assertIsInstance(mono_result, dict)
+
+        for key in output_keys:
+            self.assertIn(key, mono_result)
+
+        multi_result = nlp(valid_inputs)
+        self.assertIsInstance(multi_result, list)
+        self.assertIsInstance(multi_result[0], dict)
+
+        for result in multi_result:
+            for key in output_keys:
+                self.assertIn(key, result)
+        for bad_input in invalid_inputs:
+            self.assertRaises(Exception, nlp, bad_input)
+        self.assertRaises(Exception, nlp, invalid_inputs)

tests/test_pipelines_sentiment_analysis.py

+import unittest
+
+from .test_pipelines_common import MonoInputPipelineCommonMixin
+
+
+class SentimentAnalysisPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "sentiment-analysis"
+    small_models = [
+        "sshleifer/tiny-distilbert-base-uncased-finetuned-sst-2-english"
+    ]  # Default model - Models tested without the @slow decorator
+    large_models = [None]  # Models tested with the @slow decorator
+    mandatory_keys = {"label", "score"}  # Keys which should be in the output

tests/test_pipelines_summarization.py

+import unittest
+
+from transformers import pipeline
+from transformers.testing_utils import require_torch, slow, torch_device
+
+from .test_pipelines_common import MonoInputPipelineCommonMixin
+
+
+DEFAULT_DEVICE_NUM = -1 if torch_device == "cpu" else 0
+
+
+class SummarizationPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "summarization"
+    pipeline_running_kwargs = {"num_beams": 2, "min_length": 2, "max_length": 5}
+    small_models = [
+        "patrickvonplaten/t5-tiny-random",
+        "sshleifer/bart-tiny-random",
+    ]  # Models tested without the @slow decorator
+    large_models = []  # Models tested with the @slow decorator
+    invalid_inputs = [4, "<mask>"]
+    mandatory_keys = ["summary_text"]
+
+    @require_torch
+    @slow
+    def test_integration_torch_summarization(self):
+        nlp = pipeline(task="summarization", device=DEFAULT_DEVICE_NUM)
+        cnn_article = ' (CNN)The Palestinian Authority officially became the 123rd member of the International Criminal Court on Wednesday, a step that gives the court jurisdiction over alleged crimes in Palestinian territories. The formal accession was marked with a ceremony at The Hague, in the Netherlands, where the court is based. The Palestinians signed the ICC\'s founding Rome Statute in January, when they also accepted its jurisdiction over alleged crimes committed "in the occupied Palestinian territory, including East Jerusalem, since June 13, 2014." Later that month, the ICC opened a preliminary examination into the situation in Palestinian territories, paving the way for possible war crimes investigations against Israelis. As members of the court, Palestinians may be subject to counter-charges as well. Israel and the United States, neither of which is an ICC member, opposed the Palestinians\' efforts to join the body. But Palestinian Foreign Minister Riad al-Malki, speaking at Wednesday\'s ceremony, said it was a move toward greater justice. "As Palestine formally becomes a State Party to the Rome Statute today, the world is also a step closer to ending a long era of impunity and injustice," he said, according to an ICC news release. "Indeed, today brings us closer to our shared goals of justice and peace." Judge Kuniko Ozaki, a vice president of the ICC, said acceding to the treaty was just the first step for the Palestinians. "As the Rome Statute today enters into force for the State of Palestine, Palestine acquires all the rights as well as responsibilities that come with being a State Party to the Statute. These are substantive commitments, which cannot be taken lightly," she said. Rights group Human Rights Watch welcomed the development. "Governments seeking to penalize Palestine for joining the ICC should immediately end their pressure, and countries that support universal acceptance of the court\'s treaty should speak out to welcome its membership," said Balkees Jarrah, international justice counsel for the group. "What\'s objectionable is the attempts to undermine international justice, not Palestine\'s decision to join a treaty to which over 100 countries around the world are members." In January, when the preliminary ICC examination was opened, Israeli Prime Minister Benjamin Netanyahu described it as an outrage, saying the court was overstepping its boundaries. The United States also said it "strongly" disagreed with the court\'s decision. "As we have said repeatedly, we do not believe that Palestine is a state and therefore we do not believe that it is eligible to join the ICC," the State Department said in a statement. It urged the warring sides to resolve their differences through direct negotiations. "We will continue to oppose actions against Israel at the ICC as counterproductive to the cause of peace," it said. But the ICC begs to differ with the definition of a state for its purposes and refers to the territories as "Palestine." While a preliminary examination is not a formal investigation, it allows the court to review evidence and determine whether to investigate suspects on both sides. Prosecutor Fatou Bensouda said her office would "conduct its analysis in full independence and impartiality." The war between Israel and Hamas militants in Gaza last summer left more than 2,000 people dead. The inquiry will include alleged war crimes committed since June. The International Criminal Court was set up in 2002 to prosecute genocide, crimes against humanity and war crimes. CNN\'s Vasco Cotovio, Kareem Khadder and Faith Karimi contributed to this report.'
+        expected_cnn_summary = " The Palestinian Authority becomes the 123rd member of the International Criminal Court . The move gives the court jurisdiction over alleged crimes in Palestinian territories . Israel and the United States opposed the Palestinians' efforts to join the court . Rights group Human Rights Watch welcomes the move, says governments seeking to penalize Palestine should end pressure ."
+        result = nlp(cnn_article)
+        self.assertEqual(result[0]["summary_text"], expected_cnn_summary)

tests/test_pipelines_text2text_generation.py

+import unittest
+
+from .test_pipelines_common import MonoInputPipelineCommonMixin
+
+
+class Text2TextGenerationPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "text2text-generation"
+    small_models = ["patrickvonplaten/t5-tiny-random"]  # Default model - Models tested without the @slow decorator
+    large_models = []  # Models tested with the @slow decorator
+    invalid_inputs = [4, "<mask>"]
+    mandatory_keys = ["generated_text"]

tests/test_pipelines_text_generation.py

+import unittest
+
+from .test_pipelines_common import MonoInputPipelineCommonMixin
+
+
+class TextGenerationPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "text-generation"
+    pipeline_running_kwargs = {"prefix": "This is "}
+    small_models = ["sshleifer/tiny-ctrl"]  # Models tested without the @slow decorator
+    large_models = []  # Models tested with the @slow decorator

tests/test_pipelines_translation.py

+import unittest
+
+import pytest
+
+from transformers import pipeline
+from transformers.testing_utils import is_pipeline_test, require_torch, slow
+
+from .test_pipelines_common import MonoInputPipelineCommonMixin
+
+
+class TranslationEnToDePipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "translation_en_to_de"
+    small_models = ["patrickvonplaten/t5-tiny-random"]  # Default model - Models tested without the @slow decorator
+    large_models = [None]  # Models tested with the @slow decorator
+    invalid_inputs = [4, "<mask>"]
+    mandatory_keys = ["translation_text"]
+
+
+class TranslationEnToRoPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "translation_en_to_ro"
+    small_models = ["patrickvonplaten/t5-tiny-random"]  # Default model - Models tested without the @slow decorator
+    large_models = [None]  # Models tested with the @slow decorator
+    invalid_inputs = [4, "<mask>"]
+    mandatory_keys = ["translation_text"]
+
+
+@is_pipeline_test
+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
+        pipeline(task="translation_en_to_de")
+
+    @require_torch
+    def test_translation_on_odd_language(self):
+        model = "patrickvonplaten/t5-tiny-random"
+        pipeline(task="translation_cn_to_ar", model=model)
+
+    @require_torch
+    def test_translation_default_language_selection(self):
+        model = "patrickvonplaten/t5-tiny-random"
+        with pytest.warns(UserWarning, match=r".*translation_en_to_de.*"):
+            nlp = pipeline(task="translation", model=model)
+        self.assertEqual(nlp.task, "translation_en_to_de")
+
+    @require_torch
+    def test_translation_with_no_language_no_model_fails(self):
+        with self.assertRaises(ValueError):
+            pipeline(task="translation")

tests/test_pipelines_zero_shot.py

+import unittest
+
+from transformers.pipelines import Pipeline
+
+from .test_pipelines_common import CustomInputPipelineCommonMixin
+
+
+class ZeroShotClassificationPipelineTests(CustomInputPipelineCommonMixin, unittest.TestCase):
+    pipeline_task = "zero-shot-classification"
+    small_models = [
+        "sshleifer/tiny-distilbert-base-uncased-finetuned-sst-2-english"
+    ]  # Models tested without the @slow decorator
+    large_models = ["roberta-large-mnli"]  # Models tested with the @slow decorator
+
+    def _test_scores_sum_to_one(self, result):
+        sum = 0.0
+        for score in result["scores"]:
+            sum += score
+        self.assertAlmostEqual(sum, 1.0)
+
+    def _test_pipeline(self, nlp: Pipeline):
+        output_keys = {"sequence", "labels", "scores"}
+        valid_mono_inputs = [
+            {"sequences": "Who are you voting for in 2020?", "candidate_labels": "politics"},
+            {"sequences": "Who are you voting for in 2020?", "candidate_labels": ["politics"]},
+            {"sequences": "Who are you voting for in 2020?", "candidate_labels": "politics, public health"},
+            {"sequences": "Who are you voting for in 2020?", "candidate_labels": ["politics", "public health"]},
+            {"sequences": ["Who are you voting for in 2020?"], "candidate_labels": "politics"},
+            {
+                "sequences": "Who are you voting for in 2020?",
+                "candidate_labels": "politics",
+                "hypothesis_template": "This text is about {}",
+            },
+        ]
+        valid_multi_input = {
+            "sequences": ["Who are you voting for in 2020?", "What is the capital of Spain?"],
+            "candidate_labels": "politics",
+        }
+        invalid_inputs = [
+            {"sequences": None, "candidate_labels": "politics"},
+            {"sequences": "", "candidate_labels": "politics"},
+            {"sequences": "Who are you voting for in 2020?", "candidate_labels": None},
+            {"sequences": "Who are you voting for in 2020?", "candidate_labels": ""},
+            {
+                "sequences": "Who are you voting for in 2020?",
+                "candidate_labels": "politics",
+                "hypothesis_template": None,
+            },
+            {
+                "sequences": "Who are you voting for in 2020?",
+                "candidate_labels": "politics",
+                "hypothesis_template": "",
+            },
+            {
+                "sequences": "Who are you voting for in 2020?",
+                "candidate_labels": "politics",
+                "hypothesis_template": "Template without formatting syntax.",
+            },
+        ]
+        self.assertIsNotNone(nlp)
+
+        for mono_input in valid_mono_inputs:
+            mono_result = nlp(**mono_input)
+            self.assertIsInstance(mono_result, dict)
+            if len(mono_result["labels"]) > 1:
+                self._test_scores_sum_to_one(mono_result)
+
+            for key in output_keys:
+                self.assertIn(key, mono_result)
+
+        multi_result = nlp(**valid_multi_input)
+        self.assertIsInstance(multi_result, list)
+        self.assertIsInstance(multi_result[0], dict)
+        self.assertEqual(len(multi_result), len(valid_multi_input["sequences"]))
+
+        for result in multi_result:
+            for key in output_keys:
+                self.assertIn(key, result)
+
+            if len(result["labels"]) > 1:
+                self._test_scores_sum_to_one(result)
+
+        for bad_input in invalid_inputs:
+            self.assertRaises(Exception, nlp, **bad_input)
+
+        if nlp.model.name_or_path in self.large_models:
+            # We also check the outputs for the large models
+            inputs = [
+                {
+                    "sequences": "Who are you voting for in 2020?",
+                    "candidate_labels": ["politics", "public health", "science"],
+                },
+                {
+                    "sequences": "The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. Experiments on two machine translation tasks show these models to be superior in quality while being more parallelizable and requiring significantly less time to train. Our model achieves 28.4 BLEU on the WMT 2014 English-to-German translation task, improving over the existing best results, including ensembles by over 2 BLEU. On the WMT 2014 English-to-French translation task, our model establishes a new single-model state-of-the-art BLEU score of 41.8 after training for 3.5 days on eight GPUs, a small fraction of the training costs of the best models from the literature. We show that the Transformer generalizes well to other tasks by applying it successfully to English constituency parsing both with large and limited training data.",
+                    "candidate_labels": ["machine learning", "statistics", "translation", "vision"],
+                    "multi_class": True,
+                },
+            ]
+
+            expected_outputs = [
+                {
+                    "sequence": "Who are you voting for in 2020?",
+                    "labels": ["politics", "public health", "science"],
+                    "scores": [0.975, 0.015, 0.008],
+                },
+                {
+                    "sequence": "The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. Experiments on two machine translation tasks show these models to be superior in quality while being more parallelizable and requiring significantly less time to train. Our model achieves 28.4 BLEU on the WMT 2014 English-to-German translation task, improving over the existing best results, including ensembles by over 2 BLEU. On the WMT 2014 English-to-French translation task, our model establishes a new single-model state-of-the-art BLEU score of 41.8 after training for 3.5 days on eight GPUs, a small fraction of the training costs of the best models from the literature. We show that the Transformer generalizes well to other tasks by applying it successfully to English constituency parsing both with large and limited training data.",
+                    "labels": ["translation", "machine learning", "vision", "statistics"],
+                    "scores": [0.817, 0.712, 0.018, 0.017],
+                },
+            ]
+
+            for input, expected_output in zip(inputs, expected_outputs):
+                output = nlp(**input)
+                for key in output:
+                    if key == "scores":
+                        for output_score, expected_score in zip(output[key], expected_output[key]):
+                            self.assertAlmostEqual(output_score, expected_score, places=2)
+                    else:
+                        self.assertEqual(output[key], expected_output[key])

tests/test_tokenization_common.py

@@ -25,7 +25,14 @@ from itertools import takewhile
 from typing import TYPE_CHECKING, Dict, List, Tuple, Union
 
 from transformers import PreTrainedTokenizer, PreTrainedTokenizerBase, PreTrainedTokenizerFast, is_torch_available
-from transformers.testing_utils import get_tests_dir, require_tf, require_tokenizers, require_torch, slow
+from transformers.testing_utils import (
+    get_tests_dir,
+    is_pt_tf_cross_test,
+    require_tf,
+    require_tokenizers,
+    require_torch,
+    slow,
+)
 from transformers.tokenization_utils import AddedToken
 
 
@@ -1517,8 +1524,7 @@ class TokenizerTesterMixin:
                 string_sequences, return_overflowing_tokens=True, truncation=True, padding=True, max_length=3
             )
 
-    @require_torch
-    @require_tf
+    @is_pt_tf_cross_test
     def test_batch_encode_plus_tensors(self):
         tokenizers = self.get_tokenizers(do_lower_case=False)
         for tokenizer in tokenizers: