Fix question_answering (#3134)

* fix question_answering

* fix raising 400 HTTP error while calling question answering models

* fix question_answering

remove extra_payload parameter values

* Apply style fixes

---------

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>

Author: 2 people

src/huggingface_hub/inference/_client.py

@@ -1502,7 +1502,7 @@ def question_answering(
         model_id = model or self.model
         provider_helper = get_provider_helper(self.provider, task="question-answering", model=model_id)
         request_parameters = provider_helper.prepare_request(
-            inputs=None,
+            inputs={"question": question, "context": context},
             parameters={
                 "align_to_words": align_to_words,
                 "doc_stride": doc_stride,
@@ -1512,7 +1512,6 @@ def question_answering(
                 "max_seq_len": max_seq_len,
                 "top_k": top_k,
             },
-            extra_payload={"question": question, "context": context},
             headers=self.headers,
             model=model_id,
             api_key=self.token,

src/huggingface_hub/inference/_generated/_async_client.py

@@ -1551,7 +1551,7 @@ async def question_answering(
         model_id = model or self.model
         provider_helper = get_provider_helper(self.provider, task="question-answering", model=model_id)
         request_parameters = provider_helper.prepare_request(
-            inputs=None,
+            inputs={"question": question, "context": context},
             parameters={
                 "align_to_words": align_to_words,
                 "doc_stride": doc_stride,
@@ -1561,7 +1561,6 @@ async def question_answering(
                 "max_seq_len": max_seq_len,
                 "top_k": top_k,
             },
-            extra_payload={"question": question, "context": context},
             headers=self.headers,
             model=model_id,
             api_key=self.token,