From d771567aae6ab314794d90060ffdc23bd9a85c35 Mon Sep 17 00:00:00 2001
From: martinbrose <13284268+martinbrose@users.noreply.github.com>
Date: Sun, 20 Aug 2023 19:19:13 +0100
Subject: [PATCH 1/6] Add question-answering to inference client

---
 docs/source/guides/inference.md               |  2 +-
 src/huggingface_hub/inference/_client.py      | 55 ++++++++++++++++++
 .../inference/_generated/_async_client.py     | 56 +++++++++++++++++++
 src/huggingface_hub/inference/_types.py       | 19 +++++++
 ...ClientVCRTest.test_question_answering.yaml | 48 ++++++++++++++++
 tests/test_inference_client.py                | 10 ++++
 6 files changed, 189 insertions(+), 1 deletion(-)
 create mode 100644 tests/cassettes/InferenceClientVCRTest.test_question_answering.yaml

diff --git a/docs/source/guides/inference.md b/docs/source/guides/inference.md
index 5ee7d7a114..e12f72045a 100644
--- a/docs/source/guides/inference.md
+++ b/docs/source/guides/inference.md
@@ -135,7 +135,7 @@ has a simple API that supports the most common tasks. Here is a list of the curr
 | NLP | [Conversational](https://huggingface.co/tasks/conversational)                 | ✅ | [`~InferenceClient.conversational`] |
 | | [Feature Extraction](https://huggingface.co/tasks/feature-extraction)             | ✅ | [`~InferenceClient.feature_extraction`] |
 | | [Fill Mask](https://huggingface.co/tasks/fill-mask)                      |              |  |
-| | [Question Answering](https://huggingface.co/tasks/question-answering)             |              |  |
+| | [Question Answering](https://huggingface.co/tasks/question-answering)             | ✅ | [`~InferenceClient.question-answering`]
 | | [Sentence Similarity](https://huggingface.co/tasks/sentence-similarity)            | ✅ | [`~InferenceClient.sentence_similarity`] |
 | | [Summarization](https://huggingface.co/tasks/summarization)                  | ✅ | [`~InferenceClient.summarization`] |
 | | [Table Question Answering](https://huggingface.co/tasks/table-question-answering)       | |  |
diff --git a/src/huggingface_hub/inference/_client.py b/src/huggingface_hub/inference/_client.py
index 15d0c4edff..536a9e645a 100644
--- a/src/huggingface_hub/inference/_client.py
+++ b/src/huggingface_hub/inference/_client.py
@@ -80,6 +80,7 @@
     ConversationalOutput,
     ImageSegmentationOutput,
     ObjectDetectionOutput,
+    QuestionAnsweringOutput,
 )
 from huggingface_hub.utils import (
     BadRequestError,
@@ -673,6 +674,60 @@ def object_detection(
             raise ValueError(f"Server output must be a list. Got {type(output)}: {str(output)[:200]}...")
         return output
 
+    def question_answering(
+        self, question: str, context: str, model: str, *, parameters: Optional[Dict[str, Any]] = None
+    ) -> List[QuestionAnsweringOutput]:
+        """
+        Perform sentiment-analysis on the given text.
+
+        Args:
+            question (`str`):
+                Question to be answered.
+            context (`str`):
+                The context of the question.
+            model (`str`):
+                The model to use for the text question-answering task. Can be a model ID hosted on the Hugging Face Hub or a URL to
+                a deployed Inference Endpoint.
+            parameters (`Dict[str, Any]`, *optional*):
+                Additional parameters for the text classification task. Defaults to None. For more details about the available
+                parameters, please refer to [this page](https://huggingface.co/docs/api-inference/detailed_parameters#text-classification-task)
+
+
+        Returns:
+            `Dict`: a dictionary containing:
+            - answer:	A string that’s the answer within the text.
+            - score:	A float that represents how likely that the answer is correct
+            - start:	The index (string wise) of the start of the answer within context.
+            - stop:	    The index (string wise) of the stop of the answer within context.
+
+        Raises:
+            [`InferenceTimeoutError`]:
+                If the model is unavailable or the request times out.
+            `HTTPError`:
+                If the request fails with an HTTP error status code other than HTTP 503.
+
+        Example:
+        ```py
+        >>> from huggingface_hub import InferenceClient
+        >>> client = InferenceClient()
+        >>> output = client.question_answering(question="What's my name?", context="My name is Clara and I live in Berkeley.")
+        >>> output
+        {'score': 0.9326562285423279, 'start': 11, 'end': 16, 'answer': 'Clara'}
+        ```
+        """
+        if model is None:
+            raise ValueError("You must specify a model. Task question-answering has no recommended standard model.")
+
+        payload: Dict[str, Any] = {"question": question, "context": context}
+        if parameters is not None:
+            payload["parameters"] = parameters
+        response = self.post(
+            json=payload,
+            model=model,
+            task="question_answering",
+        )
+        return _bytes_to_dict(response)
+
     def sentence_similarity(
         self, sentence: str, other_sentences: List[str], *, model: Optional[str] = None
     ) -> List[float]:
diff --git a/src/huggingface_hub/inference/_generated/_async_client.py b/src/huggingface_hub/inference/_generated/_async_client.py
index f9df5e2644..bf405aac13 100644
--- a/src/huggingface_hub/inference/_generated/_async_client.py
+++ b/src/huggingface_hub/inference/_generated/_async_client.py
@@ -64,6 +64,7 @@
     ConversationalOutput,
     ImageSegmentationOutput,
     ObjectDetectionOutput,
+    QuestionAnsweringOutput,
 )
 from huggingface_hub.utils import (
     build_hf_headers,
@@ -678,6 +679,61 @@ async def object_detection(
             raise ValueError(f"Server output must be a list. Got {type(output)}: {str(output)[:200]}...")
         return output
 
+    async def question_answering(
+        self, question: str, context: str, model: str, *, parameters: Optional[Dict[str, Any]] = None
+    ) -> List[QuestionAnsweringOutput]:
+        """
+        Perform sentiment-analysis on the given text.
+
+        Args:
+            question (`str`):
+                Question to be answered.
+            context (`str`):
+                The context of the question.
+            model (`str`):
+                The model to use for the text question-answering task. Can be a model ID hosted on the Hugging Face Hub or a URL to
+                a deployed Inference Endpoint.
+            parameters (`Dict[str, Any]`, *optional*):
+                Additional parameters for the text classification task. Defaults to None. For more details about the available
+                parameters, please refer to [this page](https://huggingface.co/docs/api-inference/detailed_parameters#text-classification-task)
+
+
+        Returns:
+            `Dict`: a dictionary containing:
+            - answer:	A string that’s the answer within the text.
+            - score:	A float that represents how likely that the answer is correct
+            - start:	The index (string wise) of the start of the answer within context.
+            - stop:	    The index (string wise) of the stop of the answer within context.
+
+        Raises:
+            [`InferenceTimeoutError`]:
+                If the model is unavailable or the request times out.
+            `aiohttp.ClientResponseError`:
+                If the request fails with an HTTP error status code other than HTTP 503.
+
+        Example:
+        ```py
+        # Must be run in an async context
+        >>> from huggingface_hub import AsyncInferenceClient
+        >>> client = AsyncInferenceClient()
+        >>> output = await client.question_answering(question="What's my name?", context="My name is Clara and I live in Berkeley.")
+        >>> output
+        {'score': 0.9326562285423279, 'start': 11, 'end': 16, 'answer': 'Clara'}
+        ```
+        """
+        if model is None:
+            raise ValueError("You must specify a model. Task question-answering has no recommended standard model.")
+
+        payload: Dict[str, Any] = {"question": question, "context": context}
+        if parameters is not None:
+            payload["parameters"] = parameters
+        response = await self.post(
+            json=payload,
+            model=model,
+            task="question_answering",
+        )
+        return _bytes_to_dict(response)
+
     async def sentence_similarity(
         self, sentence: str, other_sentences: List[str], *, model: Optional[str] = None
     ) -> List[float]:
diff --git a/src/huggingface_hub/inference/_types.py b/src/huggingface_hub/inference/_types.py
index 3af2717aec..6e364e4a19 100644
--- a/src/huggingface_hub/inference/_types.py
+++ b/src/huggingface_hub/inference/_types.py
@@ -98,3 +98,22 @@ class ObjectDetectionOutput(TypedDict):
     label: str
     box: dict
     score: float
+
+
+class QuestionAnsweringOutput(TypedDict):
+    """Dictionary containing information about a [`~InferenceClient.question_answering`] task.
+
+    Args:
+        label (`str`):
+            The label corresponding to the detected object.
+        box (`dict`):
+            A dict response of bounding box coordinates of
+            the detected object: xmin, ymin, xmax, ymax
+        score (`float`):
+            The score corresponding to the detected object.
+    """
+
+    score: float
+    start: int
+    end: int
+    answer: str
diff --git a/tests/cassettes/InferenceClientVCRTest.test_question_answering.yaml b/tests/cassettes/InferenceClientVCRTest.test_question_answering.yaml
new file mode 100644
index 0000000000..5e6ee3c0d6
--- /dev/null
+++ b/tests/cassettes/InferenceClientVCRTest.test_question_answering.yaml
@@ -0,0 +1,48 @@
+interactions:
+- request:
+    body: '{"question": "What is the meaning of life?", "context": "42"}'
+    headers:
+      Accept:
+      - '*/*'
+      Accept-Encoding:
+      - gzip, deflate, br
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '61'
+      Content-Type:
+      - application/json
+      X-Amzn-Trace-Id:
+      - e3071dca-bb69-47b9-b9c7-f0ce58a69927
+      user-agent:
+      - unknown/None; hf_hub/0.17.0.dev0; python/3.10.12
+    method: POST
+    uri: https://api-inference.huggingface.co/models/deepset/roberta-base-squad2
+  response:
+    body:
+      string: '{"score":1.4291124728060822e-08,"start":0,"end":2,"answer":"42"}'
+    headers:
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '64'
+      Content-Type:
+      - application/json
+      Date:
+      - Sun, 20 Aug 2023 18:17:17 GMT
+      access-control-allow-credentials:
+      - 'true'
+      vary:
+      - Origin, Access-Control-Request-Method, Access-Control-Request-Headers
+      x-compute-time:
+      - '0.094'
+      x-compute-type:
+      - cache
+      x-request-id:
+      - vY1d3zhYMs71Bmhh1OI5N
+      x-sha:
+      - e09df911dd96d8b052d2665dfbb309e9398a9d70
+    status:
+      code: 200
+      message: OK
+version: 1
diff --git a/tests/test_inference_client.py b/tests/test_inference_client.py
index 2fe2b9bdb9..f73c3017dd 100644
--- a/tests/test_inference_client.py
+++ b/tests/test_inference_client.py
@@ -171,6 +171,16 @@ def test_object_detection(self) -> None:
             self.assertIn("xmax", item["box"])
             self.assertIn("ymax", item["box"])
 
+    def test_question_answering(self) -> None:
+        model = "deepset/roberta-base-squad2"
+        output = self.client.question_answering(question="What is the meaning of life?", context="42", model=model)
+        self.assertIsInstance(output, dict)
+        self.assertGreater(len(output), 0)
+        self.assertIsInstance(output["score"], float)
+        self.assertIsInstance(output["start"], int)
+        self.assertIsInstance(output["end"], int)
+        self.assertIsInstance(output["answer"], str)
+
     def test_sentence_similarity(self) -> None:
         scores = self.client.sentence_similarity(
             "Machine learning is so easy.",

From 1cbdbff42bbc561677355ae0829a01b94eb350ff Mon Sep 17 00:00:00 2001
From: martinbrose <13284268+martinbrose@users.noreply.github.com>
Date: Mon, 21 Aug 2023 14:14:38 +0100
Subject: [PATCH 2/6] Corrected doc string.

---
 src/huggingface_hub/inference/_client.py                  | 2 +-
 src/huggingface_hub/inference/_generated/_async_client.py | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/src/huggingface_hub/inference/_client.py b/src/huggingface_hub/inference/_client.py
index 536a9e645a..9db1dec3d6 100644
--- a/src/huggingface_hub/inference/_client.py
+++ b/src/huggingface_hub/inference/_client.py
@@ -678,7 +678,7 @@ def question_answering(
         self, question: str, context: str, model: str, *, parameters: Optional[Dict[str, Any]] = None
     ) -> List[QuestionAnsweringOutput]:
         """
-        Perform sentiment-analysis on the given text.
+        Retrieve the answer to a question from a given text.
 
         Args:
             question (`str`):
diff --git a/src/huggingface_hub/inference/_generated/_async_client.py b/src/huggingface_hub/inference/_generated/_async_client.py
index bf405aac13..ba736720fd 100644
--- a/src/huggingface_hub/inference/_generated/_async_client.py
+++ b/src/huggingface_hub/inference/_generated/_async_client.py
@@ -683,7 +683,7 @@ async def question_answering(
         self, question: str, context: str, model: str, *, parameters: Optional[Dict[str, Any]] = None
     ) -> List[QuestionAnsweringOutput]:
         """
-        Perform sentiment-analysis on the given text.
+        Retrieve the answer to a question from a given text.
 
         Args:
             question (`str`):

From 186da5d52083de7ca691094df9dbdd24048a2e64 Mon Sep 17 00:00:00 2001
From: martinbrose <13284268+martinbrose@users.noreply.github.com>
Date: Mon, 21 Aug 2023 16:13:10 +0100
Subject: [PATCH 3/6] Corrected typos

---
 src/huggingface_hub/inference/_client.py                  | 8 ++++----
 src/huggingface_hub/inference/_generated/_async_client.py | 8 ++++----
 2 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/src/huggingface_hub/inference/_client.py b/src/huggingface_hub/inference/_client.py
index 9db1dec3d6..7657b7c47e 100644
--- a/src/huggingface_hub/inference/_client.py
+++ b/src/huggingface_hub/inference/_client.py
@@ -686,11 +686,11 @@ def question_answering(
             context (`str`):
                 The context of the question.
             model (`str`):
-                The model to use for the text question-answering task. Can be a model ID hosted on the Hugging Face Hub or a URL to
+                The model to use for the question answering task. Can be a model ID hosted on the Hugging Face Hub or a URL to
                 a deployed Inference Endpoint.
             parameters (`Dict[str, Any]`, *optional*):
-                Additional parameters for the text classification task. Defaults to None. For more details about the available
-                parameters, please refer to [this page](https://huggingface.co/docs/api-inference/detailed_parameters#text-classification-task)
+                Additional parameters for the question answering task. Defaults to None. For more details about the available
+                parameters, please refer to [this page](https://huggingface.co/docs/api-inference/detailed_parameters#question-answering-task)
 
 
         Returns:
@@ -724,7 +724,7 @@ def question_answering(
         response = self.post(
             json=payload,
             model=model,
-            task="question_answering",
+            task="question-answering",
         )
         return _bytes_to_dict(response)
 
diff --git a/src/huggingface_hub/inference/_generated/_async_client.py b/src/huggingface_hub/inference/_generated/_async_client.py
index ba736720fd..a1f6217227 100644
--- a/src/huggingface_hub/inference/_generated/_async_client.py
+++ b/src/huggingface_hub/inference/_generated/_async_client.py
@@ -691,11 +691,11 @@ async def question_answering(
             context (`str`):
                 The context of the question.
             model (`str`):
-                The model to use for the text question-answering task. Can be a model ID hosted on the Hugging Face Hub or a URL to
+                The model to use for the question answering task. Can be a model ID hosted on the Hugging Face Hub or a URL to
                 a deployed Inference Endpoint.
             parameters (`Dict[str, Any]`, *optional*):
-                Additional parameters for the text classification task. Defaults to None. For more details about the available
-                parameters, please refer to [this page](https://huggingface.co/docs/api-inference/detailed_parameters#text-classification-task)
+                Additional parameters for the question answering task. Defaults to None. For more details about the available
+                parameters, please refer to [this page](https://huggingface.co/docs/api-inference/detailed_parameters#question-answering-task)
 
 
         Returns:
@@ -730,7 +730,7 @@ async def question_answering(
         response = await self.post(
             json=payload,
             model=model,
-            task="question_answering",
+            task="question-answering",
         )
         return _bytes_to_dict(response)
 

From 56b314cd35a01c10ae0a3deaf2bd6b3e7af24046 Mon Sep 17 00:00:00 2001
From: martinbrose <13284268+martinbrose@users.noreply.github.com>
Date: Tue, 5 Sep 2023 22:03:14 +0100
Subject: [PATCH 4/6] Change in line with review points on other PRs

---
 src/huggingface_hub/inference/_client.py      | 20 ++++---------------
 .../inference/_generated/_async_client.py     | 20 ++++---------------
 src/huggingface_hub/inference/_types.py       | 13 ++++++------
 3 files changed, 15 insertions(+), 38 deletions(-)

diff --git a/src/huggingface_hub/inference/_client.py b/src/huggingface_hub/inference/_client.py
index 933c614d4b..858ee199fc 100644
--- a/src/huggingface_hub/inference/_client.py
+++ b/src/huggingface_hub/inference/_client.py
@@ -677,8 +677,8 @@ def object_detection(
         return output
 
     def question_answering(
-        self, question: str, context: str, model: str, *, parameters: Optional[Dict[str, Any]] = None
-    ) -> List[QuestionAnsweringOutput]:
+        self, question: str, context: str, *, model: Optional[str] = None
+    ) -> QuestionAnsweringOutput:
         """
         Retrieve the answer to a question from a given text.
 
@@ -690,17 +690,9 @@ def question_answering(
             model (`str`):
                 The model to use for the question answering task. Can be a model ID hosted on the Hugging Face Hub or a URL to
                 a deployed Inference Endpoint.
-            parameters (`Dict[str, Any]`, *optional*):
-                Additional parameters for the question answering task. Defaults to None. For more details about the available
-                parameters, please refer to [this page](https://huggingface.co/docs/api-inference/detailed_parameters#question-answering-task)
-
 
         Returns:
-            `Dict`: a dictionary containing:
-            - answer:	A string that’s the answer within the text.
-            - score:	A float that represents how likely that the answer is correct
-            - start:	The index (string wise) of the start of the answer within context.
-            - stop:	    The index (string wise) of the stop of the answer within context.
+            `Dict`: a dictionary of question answering output containing the score, start index, end index, and answer.
 
         Raises:
             [`InferenceTimeoutError`]:
@@ -717,18 +709,14 @@ def question_answering(
         {'score': 0.9326562285423279, 'start': 11, 'end': 16, 'answer': 'Clara'}
         ```
         """
-        if model is None:
-            raise ValueError("You must specify a model. Task question-answering has no recommended standard model.")
 
         payload: Dict[str, Any] = {"question": question, "context": context}
-        if parameters is not None:
-            payload["parameters"] = parameters
         response = self.post(
             json=payload,
             model=model,
             task="question-answering",
         )
-        return _bytes_to_dict(response)
+        return _bytes_to_dict(response)  # type: ignore
 
     def sentence_similarity(
         self, sentence: str, other_sentences: List[str], *, model: Optional[str] = None
diff --git a/src/huggingface_hub/inference/_generated/_async_client.py b/src/huggingface_hub/inference/_generated/_async_client.py
index 36b5ce4afe..c997ae6e9d 100644
--- a/src/huggingface_hub/inference/_generated/_async_client.py
+++ b/src/huggingface_hub/inference/_generated/_async_client.py
@@ -682,8 +682,8 @@ async def object_detection(
         return output
 
     async def question_answering(
-        self, question: str, context: str, model: str, *, parameters: Optional[Dict[str, Any]] = None
-    ) -> List[QuestionAnsweringOutput]:
+        self, question: str, context: str, *, model: Optional[str] = None
+    ) -> QuestionAnsweringOutput:
         """
         Retrieve the answer to a question from a given text.
 
@@ -695,17 +695,9 @@ async def question_answering(
             model (`str`):
                 The model to use for the question answering task. Can be a model ID hosted on the Hugging Face Hub or a URL to
                 a deployed Inference Endpoint.
-            parameters (`Dict[str, Any]`, *optional*):
-                Additional parameters for the question answering task. Defaults to None. For more details about the available
-                parameters, please refer to [this page](https://huggingface.co/docs/api-inference/detailed_parameters#question-answering-task)
-
 
         Returns:
-            `Dict`: a dictionary containing:
-            - answer:	A string that’s the answer within the text.
-            - score:	A float that represents how likely that the answer is correct
-            - start:	The index (string wise) of the start of the answer within context.
-            - stop:	    The index (string wise) of the stop of the answer within context.
+            `Dict`: a dictionary of question answering output containing the score, start index, end index, and answer.
 
         Raises:
             [`InferenceTimeoutError`]:
@@ -723,18 +715,14 @@ async def question_answering(
         {'score': 0.9326562285423279, 'start': 11, 'end': 16, 'answer': 'Clara'}
         ```
         """
-        if model is None:
-            raise ValueError("You must specify a model. Task question-answering has no recommended standard model.")
 
         payload: Dict[str, Any] = {"question": question, "context": context}
-        if parameters is not None:
-            payload["parameters"] = parameters
         response = await self.post(
             json=payload,
             model=model,
             task="question-answering",
         )
-        return _bytes_to_dict(response)
+        return _bytes_to_dict(response)  # type: ignore
 
     async def sentence_similarity(
         self, sentence: str, other_sentences: List[str], *, model: Optional[str] = None
diff --git a/src/huggingface_hub/inference/_types.py b/src/huggingface_hub/inference/_types.py
index 6e364e4a19..0597b7d1b5 100644
--- a/src/huggingface_hub/inference/_types.py
+++ b/src/huggingface_hub/inference/_types.py
@@ -104,13 +104,14 @@ class QuestionAnsweringOutput(TypedDict):
     """Dictionary containing information about a [`~InferenceClient.question_answering`] task.
 
     Args:
-        label (`str`):
-            The label corresponding to the detected object.
-        box (`dict`):
-            A dict response of bounding box coordinates of
-            the detected object: xmin, ymin, xmax, ymax
         score (`float`):
-            The score corresponding to the detected object.
+            A float that represents how likely that the answer is correct.
+        start (`int`):
+            The index (string wise) of the start of the answer within context.
+        end (`int`):
+            The index (string wise) of the end of the answer within context.
+        answer (`str`):
+            A string that is the answer within the text.
     """
 
     score: float

From 839a3a0a9cb829806f3d2dd4db565ad77ed18c1c Mon Sep 17 00:00:00 2001
From: Lucain <lucainp@gmail.com>
Date: Wed, 6 Sep 2023 18:04:50 +0200
Subject: [PATCH 5/6] Apply suggestions from code review

---
 docs/source/en/guides/inference.md                        | 2 +-
 src/huggingface_hub/inference/_client.py                  | 3 +--
 src/huggingface_hub/inference/_generated/_async_client.py | 3 +--
 tests/test_inference_client.py                            | 2 +-
 4 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/docs/source/en/guides/inference.md b/docs/source/en/guides/inference.md
index 44b367dd86..7e766a44a8 100644
--- a/docs/source/en/guides/inference.md
+++ b/docs/source/en/guides/inference.md
@@ -135,7 +135,7 @@ has a simple API that supports the most common tasks. Here is a list of the curr
 | NLP | [Conversational](https://huggingface.co/tasks/conversational)                 | ✅ | [`~InferenceClient.conversational`] |
 | | [Feature Extraction](https://huggingface.co/tasks/feature-extraction)             | ✅ | [`~InferenceClient.feature_extraction`] |
 | | [Fill Mask](https://huggingface.co/tasks/fill-mask)                      |              |  |
-| | [Question Answering](https://huggingface.co/tasks/question-answering)             | ✅ | [`~InferenceClient.question-answering`]
+| | [Question Answering](https://huggingface.co/tasks/question-answering)             | ✅ | [`~InferenceClient.question_answering`]
 | | [Sentence Similarity](https://huggingface.co/tasks/sentence-similarity)            | ✅ | [`~InferenceClient.sentence_similarity`] |
 | | [Summarization](https://huggingface.co/tasks/summarization)                  | ✅ | [`~InferenceClient.summarization`] |
 | | [Table Question Answering](https://huggingface.co/tasks/table-question-answering)       | |  |
diff --git a/src/huggingface_hub/inference/_client.py b/src/huggingface_hub/inference/_client.py
index f7217862a3..bae967b126 100644
--- a/src/huggingface_hub/inference/_client.py
+++ b/src/huggingface_hub/inference/_client.py
@@ -705,8 +705,7 @@ def question_answering(
         ```py
         >>> from huggingface_hub import InferenceClient
         >>> client = InferenceClient()
-        >>> output = client.question_answering(question="What's my name?", context="My name is Clara and I live in Berkeley.")
-        >>> output
+        >>> client.question_answering(question="What's my name?", context="My name is Clara and I live in Berkeley.")
         {'score': 0.9326562285423279, 'start': 11, 'end': 16, 'answer': 'Clara'}
         ```
         """
diff --git a/src/huggingface_hub/inference/_generated/_async_client.py b/src/huggingface_hub/inference/_generated/_async_client.py
index 0f9df19470..777dd26bc3 100644
--- a/src/huggingface_hub/inference/_generated/_async_client.py
+++ b/src/huggingface_hub/inference/_generated/_async_client.py
@@ -711,8 +711,7 @@ async def question_answering(
         # Must be run in an async context
         >>> from huggingface_hub import AsyncInferenceClient
         >>> client = AsyncInferenceClient()
-        >>> output = await client.question_answering(question="What's my name?", context="My name is Clara and I live in Berkeley.")
-        >>> output
+        >>> await client.question_answering(question="What's my name?", context="My name is Clara and I live in Berkeley.")
         {'score': 0.9326562285423279, 'start': 11, 'end': 16, 'answer': 'Clara'}
         ```
         """
diff --git a/tests/test_inference_client.py b/tests/test_inference_client.py
index a12df250df..52509eba40 100644
--- a/tests/test_inference_client.py
+++ b/tests/test_inference_client.py
@@ -179,7 +179,7 @@ def test_question_answering(self) -> None:
         self.assertIsInstance(output["score"], float)
         self.assertIsInstance(output["start"], int)
         self.assertIsInstance(output["end"], int)
-        self.assertIsInstance(output["answer"], str)
+        self.assertEqual(output["answer"], "Clara")
 
     def test_sentence_similarity(self) -> None:
         scores = self.client.sentence_similarity(

From 20e1ba45025333495f786b816ccc8231a6f81e8d Mon Sep 17 00:00:00 2001
From: Lucain Pouget <lucainp@gmail.com>
Date: Wed, 6 Sep 2023 18:25:23 +0200
Subject: [PATCH 6/6] fix

---
 tests/test_inference_client.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tests/test_inference_client.py b/tests/test_inference_client.py
index 52509eba40..61dc5852eb 100644
--- a/tests/test_inference_client.py
+++ b/tests/test_inference_client.py
@@ -179,7 +179,7 @@ def test_question_answering(self) -> None:
         self.assertIsInstance(output["score"], float)
         self.assertIsInstance(output["start"], int)
         self.assertIsInstance(output["end"], int)
-        self.assertEqual(output["answer"], "Clara")
+        self.assertEqual(output["answer"], "42")
 
     def test_sentence_similarity(self) -> None:
         scores = self.client.sentence_similarity(