distilbertmod.py

from transformers.modeling_distilbert import DistilBertPreTrainedModel, DistilBertForSequenceClassification
from transformers import DistilBertModel
from torch import nn, tensor


class MyDistilBert(DistilBertForSequenceClassification):
    r"""
        **labels**: (`optional`) ``torch.LongTensor`` of shape ``(batch_size,)``:
            Labels for computing the sequence classification/regression loss.
            Indices should be in ``[0, ..., config.num_labels - 1]``.
            If ``config.num_labels == 1`` a regression loss is computed (Mean-Square loss),
            If ``config.num_labels > 1`` a classification loss is computed (Cross-Entropy).

    Outputs: `Tuple` comprising various elements depending on the configuration (config) and inputs:
        **loss**: (`optional`, returned when ``labels`` is provided) ``torch.FloatTensor`` of shape ``(1,)``:
            Classification (or regression if config.num_labels==1) loss.
        **logits**: ``torch.FloatTensor`` of shape ``(batch_size, config.num_labels)``
            Classification (or regression if config.num_labels==1) scores (before SoftMax).
        **hidden_states**: (`optional`, returned when ``config.output_hidden_states=True``)
            list of ``torch.FloatTensor`` (one for the output of each layer + the output of the embeddings)
            of shape ``(batch_size, sequence_length, hidden_size)``:
            Hidden-states of the model at the output of each layer plus the initial embedding outputs.
        **attentions**: (`optional`, returned when ``config.output_attentions=True``)
            list of ``torch.FloatTensor`` (one for each layer) of shape ``(batch_size, num_heads, sequence_length, sequence_length)``:
            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

    Examples::

        tokenizer = DistilBertTokenizer.from_pretrained('distilbert-base-uncased')
        model = DistilBertForSequenceClassification.from_pretrained('distilbert-base-uncased')
        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
        labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
        outputs = model(input_ids, labels=labels)
        loss, logits = outputs[:2]

    """
    def __init__(self, config):
        DistilBertForSequenceClassification.__init__(self, config)
        self.num_labels = config.num_labels
        print(config)
        self.distilbert = DistilBertModel(config)
        for child in self.distilbert.children():
            for param in child.parameters():
                param.requires_grad = False
#        self.bottom_classifier = nn.Linear(config.dim, config.dim)
#        self.pre_classifier = nn.LSTM(config.dim, config.dim, config.dim)
#        self.pre_dropout = nn.Dropout(config.seq_classif_dropout)
#        self.pre_pre_classifier = nn.Linear(config.dim, config.dim)
        self.pre_classifier = nn.Linear(config.dim, config.dim)
        self.classifier = nn.Linear(config.dim, config.num_labels)
        self.dropout = nn.Dropout(config.seq_classif_dropout)

        self.init_weights()

    def forward(self, input_ids,  attention_mask=None, head_mask=None, labels=None):
#        input_ids = self.bottom_classifier(input_ids)
#        input_ids = nn.ReLU()(input_ids)
        distilbert_output = self.distilbert(input_ids=input_ids,
                                            attention_mask=attention_mask,
                                            head_mask=head_mask)
        hidden_state = distilbert_output[0]                    # (bs, seq_len, dim)
        pooled_output = hidden_state[:, 0]                    # (bs, dim)
#        pooled_output = self.pre_pre_classifier(pooled_output)
#        pooled_output = self.pre_dropout(pooled_output)
#        pooled_output = nn.ReLU()(pooled_output)
        pooled_output = self.pre_classifier(pooled_output)   # (bs, dim)
        pooled_output = nn.ReLU()(pooled_output)             # (bs, dim)
        pooled_output = self.dropout(pooled_output)         # (bs, dim)
        logits = self.classifier(pooled_output)              # (bs, dim)

        outputs = (logits,) + distilbert_output[1:]
        if labels is not None:
            if self.num_labels == 1:
                loss_fct = nn.MSELoss()
                loss = loss_fct(logits.view(-1), labels.view(-1))
            else:
                loss_fct = nn.CrossEntropyLoss()
                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
            outputs = (loss,) + outputs

        return outputs  # (loss), logits, (hidden_states), (attentions)