[BC] Fixing usage of text pairs (#17324)

* [BC] Fixing usage of text pairs

The BC is actually preventing users from misusing the pipeline since
users could have been willing to send text pairs and the pipeline would
instead understand the thing as a batch returning bogus results.

The correct usage of text pairs is preserved in this PR even when that
makes the code clunky.

Adds support for {"text":..,, "text_pair": ...} inputs for both dataset
iteration and more explicit usage to pairs.

* Updating the doc.

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

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

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

* quality.
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

src/transformers/pipelines/text_classification.py

@@ -94,8 +94,9 @@ class TextClassificationPipeline(Pipeline):
         Classify the text(s) given as inputs.
 
         Args:
-            args (`str` or `List[str]`):
-                One or several texts (or one list of prompts) to classify.
+            args (`str` or `List[str]` or `Dict[str]`, or `List[Dict[str]]`):
+                One or several texts to classify. In order to use text pairs for your classification, you can send a
+                dictionnary containing `{"text", "text_pair"}` keys, or a list of those.
             return_all_scores (`bool`, *optional*, defaults to `False`):
                 Whether to return scores for all labels.
             function_to_apply (`str`, *optional*, defaults to `"default"`):
@@ -131,6 +132,19 @@ class TextClassificationPipeline(Pipeline):
 
     def preprocess(self, inputs, **tokenizer_kwargs) -> Dict[str, GenericTensor]:
         return_tensors = self.framework
+        if isinstance(inputs, dict):
+            return self.tokenizer(**inputs, return_tensors=return_tensors, **tokenizer_kwargs)
+        elif isinstance(inputs, list) and len(inputs) == 1 and isinstance(inputs[0], list) and len(inputs[0]) == 2:
+            # It used to be valid to use a list of list of list for text pairs, keeping this path for BC
+            return self.tokenizer(
+                text=inputs[0][0], text_pair=inputs[0][1], return_tensors=return_tensors, **tokenizer_kwargs
+            )
+        elif isinstance(inputs, list):
+            # This is likely an invalid usage of the pipeline attempting to pass text pairs.
+            raise ValueError(
+                "The pipeline received invalid inputs, if you are trying to send text pairs, you can try to send a"
+                ' dictionnary `{"text": "My text", "text_pair": "My pair"}` in order to send a text pair.'
+            )
         return self.tokenizer(inputs, return_tensors=return_tensors, **tokenizer_kwargs)
 
     def _forward(self, model_inputs):

tests/pipelines/test_pipelines_text_classification.py

@@ -107,3 +107,28 @@ class TextClassificationPipelineTests(unittest.TestCase, metaclass=PipelineTestC
         )
         self.assertTrue(outputs[0]["label"] in model.config.id2label.values())
         self.assertTrue(outputs[1]["label"] in model.config.id2label.values())
+
+        valid_inputs = {"text": "HuggingFace is in ", "text_pair": "Paris is in France"}
+        outputs = text_classifier(valid_inputs)
+        self.assertEqual(
+            nested_simplify(outputs),
+            {"label": ANY(str), "score": ANY(float)},
+        )
+        self.assertTrue(outputs["label"] in model.config.id2label.values())
+
+        # This might be used a text pair, but tokenizer + pipe interaction
+        # makes it hard to understand that it's not using the pair properly
+        # https://github.com/huggingface/transformers/issues/17305
+        # We disabled this usage instead as it was outputting wrong outputs.
+        invalid_input = [["HuggingFace is in ", "Paris is in France"]]
+        with self.assertRaises(ValueError):
+            text_classifier(invalid_input)
+
+        # This used to be valid for doing text pairs
+        # We're keeping it working because of backward compatibility
+        outputs = text_classifier([[["HuggingFace is in ", "Paris is in France"]]])
+        self.assertEqual(
+            nested_simplify(outputs),
+            [{"label": ANY(str), "score": ANY(float)}],
+        )
+        self.assertTrue(outputs[0]["label"] in model.config.id2label.values())