import torch.nn as nn

<<<<<<< HEAD
from .bert import BERT
=======
from bert import BERT
>>>>>>> bffd3381ccb717f802fe651d4111ec0a268e3896


class BERTSM(nn.Module):
    """
    BERT Sequence Model
    Masked Sequence Model
    """

    def __init__(self, bert: BERT, vocab_size):
        """
        :param bert: BERT model which should be trained
        :param vocab_size: total vocab size for masked_lm
        """

        super().__init__()
        self.bert = bert
        self.mask_lm = MaskedSequenceModel(self.bert.hidden, vocab_size)
<<<<<<< HEAD
        
    def forward(self, x, segment_label):
        x = self.bert(x, segment_label)
        return self.mask_lm(x), x[:, 0]
=======
        self.same_student = SameStudentPrediction(self.bert.hidden)
        
    def forward(self, x, segment_label, pred=False):
        x = self.bert(x, segment_label)
        # torch.Size([32, 200, 512])
        # print("????????????   ",x.shape)
        if pred:
            return x[:, 0], self.mask_lm(x), self.same_student(x)
        else:
            return x[:, 0], self.mask_lm(x)
>>>>>>> bffd3381ccb717f802fe651d4111ec0a268e3896

    
class MaskedSequenceModel(nn.Module):
    """
    predicting origin token from masked input sequence
    n-class classification problem, n-class = vocab_size
    """

    def __init__(self, hidden, vocab_size):
        """
        :param hidden: output size of BERT model
        :param vocab_size: total vocab size
        """
        super().__init__()
        self.linear = nn.Linear(hidden, vocab_size)
        self.softmax = nn.LogSoftmax(dim=-1)

    def forward(self, x):
<<<<<<< HEAD
        return self.softmax(self.linear(x))
=======
        return self.softmax(self.linear(x))

    
class SameStudentPrediction(nn.Module):

    def __init__(self, hidden):
        """
        :param hidden: BERT model output size
        """
        super().__init__()
        self.linear = nn.Linear(hidden, 2)
        self.softmax = nn.LogSoftmax(dim=-1)

    def forward(self, x):
        return self.softmax(self.linear(x[:, 0]))

>>>>>>> bffd3381ccb717f802fe651d4111ec0a268e3896