feat: Add EuroBertForQuestionAnswering

This is copied directly from the [210m modeling script](https://huggingface.co/EuroBERT/EuroBERT-210m/edit/main/modeling_eurobert.py).

modeling_eurobert.py

@@ -30,7 +30,7 @@ from transformers.activations import ACT2FN
 from transformers.cache_utils import Cache, StaticCache
 from transformers.modeling_attn_mask_utils import AttentionMaskConverter
 from transformers.modeling_flash_attention_utils import FlashAttentionKwargs
-from transformers.modeling_outputs import BaseModelOutput, BaseModelOutputWithPast, MaskedLMOutput, SequenceClassifierOutput, TokenClassifierOutput
+from transformers.modeling_outputs import BaseModelOutput, BaseModelOutputWithPast, MaskedLMOutput, QuestionAnsweringModelOutput, SequenceClassifierOutput, TokenClassifierOutput
 from transformers.modeling_rope_utils import ROPE_INIT_FUNCTIONS
 from transformers.modeling_utils import ALL_ATTENTION_FUNCTIONS, PreTrainedModel
 from transformers.processing_utils import Unpack
@@ -951,10 +951,107 @@ class EuroBertForTokenClassification(EuroBertPreTrainedModel):
         )
 
 
+@add_start_docstrings(
+    """
+    The EuroBert Model with a span classification head on top for extractive question-answering tasks 
+    like SQuAD (a linear layers on top of the hidden-states output to compute span start logits 
+    and span end logits).
+    """,
+    EUROBERT_START_DOCSTRING,
+)
+class EuroBertForQuestionAnswering(EuroBertPreTrainedModel):
+    def __init__(self, config: EuroBertConfig):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+        self.model = EuroBertModel(config)
+
+        self.qa_outputs = nn.Linear(config.hidden_size, config.num_labels)
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.model.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.model.embed_tokens = value
+
+    @add_start_docstrings_to_model_forward(EUROBERT_INPUTS_DOCSTRING)
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        use_cache: Optional[bool] = None,
+        start_positions: Optional[torch.Tensor] = None,
+        end_positions: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple[torch.Tensor], QuestionAnsweringModelOutput]:
+        r"""
+        start_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for position (index) of the start of the labelled span for computing the token classification loss.
+            Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence
+            are not taken into account for computing the loss.
+        end_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for position (index) of the end of the labelled span for computing the token classification loss.
+            Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence
+            are not taken into account for computing the loss.
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs = self.model(
+            input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        sequence_output = outputs[0]
+
+        logits = self.qa_outputs(sequence_output)
+        start_logits, end_logits = logits.split(1, dim=-1)
+        start_logits = start_logits.squeeze(-1).contiguous()
+        end_logits = end_logits.squeeze(-1).contiguous()
+
+        total_loss = None
+        if start_positions is not None and end_positions is not None:
+            # If we are on multi-GPU, split add a dimension
+            if len(start_positions.size()) > 1:
+                start_positions = start_positions.squeeze(-1)
+            if len(end_positions.size()) > 1:
+                end_positions = end_positions.squeeze(-1)
+            # sometimes the start/end positions are outside our model inputs, we ignore these terms
+            ignored_index = start_logits.size(1)
+            start_positions = start_positions.clamp(0, ignored_index)
+            end_positions = end_positions.clamp(0, ignored_index)
+
+            loss_fct = CrossEntropyLoss(ignore_index=ignored_index)
+            start_loss = loss_fct(start_logits, start_positions)
+            end_loss = loss_fct(end_logits, end_positions)
+            total_loss = (start_loss + end_loss) / 2
+
+        if not return_dict:
+            output = (start_logits, end_logits) + outputs[2:]
+            return ((total_loss,) + output) if total_loss is not None else output
+
+        return QuestionAnsweringModelOutput(
+            loss=total_loss,
+            start_logits=start_logits,
+            end_logits=end_logits,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+
 __all__ = [
     "EuroBertPreTrainedModel",
     "EuroBertModel",
     "EuroBertForMaskedLM",
     "EuroBertForSequenceClassification",
     "EuroBertForTokenClassification",
+    "EuroBertForQuestionAnswering",
 ]