mcqa_bert.py

# mcqa_bert.py
# --------------------------------------------------
# Plain BertModel + single‑unit classification head
# --------------------------------------------------
import torch
import torch.nn as nn
from transformers import BertModel


class MCQABERT(nn.Module):
    def __init__(self, ckpt: str = "bert-base-uncased"):
        super().__init__()
        self.encoder = BertModel.from_pretrained(ckpt)
        self.head    = nn.Linear(self.encoder.config.hidden_size, 1)

    # --------------------------------------------------

    def forward(self, input_ids, attention_mask):
        out   = self.encoder(
            input_ids=input_ids,
            attention_mask=attention_mask,
            return_dict=True,
        )
        cls_vec = out.last_hidden_state[:, 0]          # [CLS]
        logits  = self.head(cls_vec).squeeze(-1)       # (B)
        return logits