Python modeling.BertModel方法代碼示例

本文整理匯總了Python中pytorch_pretrained_bert.modeling.BertModel方法的典型用法代碼示例。如果您正苦於以下問題：Python modeling.BertModel方法的具體用法？Python modeling.BertModel怎麽用？Python modeling.BertModel使用的例子？那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在類pytorch_pretrained_bert.modeling的用法示例。

在下文中一共展示了modeling.BertModel方法的15個代碼示例，這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚，您的評價將有助於係統推薦出更棒的Python代碼示例。

示例1: forward

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def forward(self, input_ids,
                token_type_ids=None,
                attention_mask=None,
                output_all_encoded_layers=True,
                token_subword_index=None):
        """
        :param input_ids: same as it in BertModel
        :param token_type_ids: same as it in BertModel
        :param attention_mask: same as it in BertModel
        :param output_all_encoded_layers: same as it in BertModel
        :param token_subword_index: [batch_size, num_tokens, num_subwords]
        :return:
        """
        # encoded_layers: [batch_size, num_subword_pieces, hidden_size]
        encoded_layers, pooled_output = super(Seq2SeqBertEncoder, self).forward(
            input_ids, token_type_ids, attention_mask, output_all_encoded_layers)
        if token_subword_index is None:
            return encoded_layers, pooled_output
        else:
            return self.average_pooling(encoded_layers, token_subword_index), pooled_output

開發者ID:jcyk，項目名稱:gtos，代碼行數:22，代碼來源:seq2seq_bert_encoder.py

示例2: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config):
        super(BertCNNForTripletNet, self).__init__(config)

        filters = [3, 4, 5]

        self.bert = BertModel(config)
        self.embedding_dropout = SpatialDropout1D(config.hidden_dropout_prob)

        self.conv_layers = nn.ModuleList()
        for filter_size in filters:
            conv_block = nn.Sequential(
                nn.Conv1d(
                    config.hidden_size,
                    CHANNEL_UNITS,
                    kernel_size=filter_size,
                    padding=1,
                ),
                # nn.BatchNorm1d(CHANNEL_UNITS),
                # nn.ReLU(inplace=True),
            )
            self.conv_layers.append(conv_block)

        self.apply(self.init_bert_weights)

開發者ID:GuidoPaul，項目名稱:CAIL2019，代碼行數:25，代碼來源:net.py

示例3: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config, answer_verification=True, hidden_dropout_prob=0.3):
        super(CailModel, self).__init__(config)
        self.bert = BertModel(config)
        # TODO check with Google if it's normal there is no dropout on the token classifier of SQuAD in the TF version
        # self.qa_dropout = nn.Dropout(config.hidden_dropout_prob)
        self.qa_outputs = nn.Linear(config.hidden_size*4, 2)
        self.apply(self.init_bert_weights)
        self.answer_verification = answer_verification
        if self.answer_verification:
            self.retionale_outputs = nn.Linear(config.hidden_size*4, 1)
            self.unk_ouputs = nn.Linear(config.hidden_size, 1)
            self.doc_att = nn.Linear(config.hidden_size*4, 1)
            self.yes_no_ouputs = nn.Linear(config.hidden_size*4, 2)
            self.ouputs_cls_3 = nn.Linear(config.hidden_size*4, 3)

            self.beta = 100
        else:
            # self.unk_yes_no_outputs_dropout = nn.Dropout(config.hidden_dropout_prob)
            self.unk_yes_no_outputs = nn.Linear(config.hidden_size, 3)

開發者ID:NoneWait，項目名稱:cail2019，代碼行數:21，代碼來源:CailModel.py

示例4: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config):
        super(BertForQuestionAnswering, self).__init__(config)
        self.bert = BertModel(config)
        self.qa_outputs = nn.Linear(config.hidden_size, 1)
        self.apply(self.init_bert_weights)

開發者ID:ymcui，項目名稱:cmrc2019，代碼行數:7，代碼來源:run_cmrc2019_baseline.py

示例5: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config):
        super(BertQA, self).__init__(config)
        self.bert = BertModel(config)

        self.qa_outputs = nn.Linear(config.hidden_size, 2)
        self.apply(self.init_bert_weights)

開發者ID:pranciskus，項目名稱:mrc-for-flat-nested-ner，代碼行數:8，代碼來源:bert_qa.py

示例6: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config, num_entity_labels):
        super(BertForBasicNER, self).__init__(config)
        self.bert = BertModel(config)

        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(config.hidden_size, num_entity_labels)
        self.apply(self.init_bert_weights)

        self.num_entity_labels = num_entity_labels

開發者ID:dolphin-zs，項目名稱:Doc2EDAG，代碼行數:11，代碼來源:ner_model.py

示例7: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config):
        super(BertForUtteranceEncoding, self).__init__(config)

        self.config = config
        self.bert = BertModel(config)

開發者ID:ConvLab，項目名稱:ConvLab，代碼行數:7，代碼來源:BeliefTrackerSlotQueryMultiSlot.py

示例8: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config):
        super(BERT_classifer, self).__init__(config)

        self.num_labels = NUM_EMO
        self.bert = BertModel(config)
        self.dropout = nn.Dropout(0.1)
        self.apply(self.init_bert_weights)
        self.bert_out_dim = None
        self.out2label = None
        self.out2binary = None
        self.out2emo = None

開發者ID:chenyangh，項目名稱:SemEval2019Task3，代碼行數:13，代碼來源:bert.py

示例9: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config, num_labels):
        super().__init__(config)
        self.num_labels = num_labels
        self.bert = BertModel(config)

        self.pooler = BertPooler(config)

        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(1*config.hidden_size, num_labels)
        self.apply(self.init_bert_weights)

開發者ID:sattree，項目名稱:gap，代碼行數:12，代碼來源:probert.py

示例10: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config, num_labels):
        super().__init__(config)
        self.num_labels = num_labels
        self.bert = BertModel(config)

        self.pooler = BertPooler(config)

        self.evidence_pooler_p = EvidencePooler(config)

        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(2 * config.hidden_size, num_labels)

        torch.nn.init.xavier_uniform_(self.classifier.weight)

        self.apply(self.init_bert_weights)

開發者ID:sattree，項目名稱:gap，代碼行數:17，代碼來源:grep.py

示例11: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, opt, bert_config=None):
        super(SANBertNetwork, self).__init__()
        self.dropout_list = nn.ModuleList()
        self.bert_config = BertConfig.from_dict(opt)
        self.bert = BertModel(self.bert_config)
        if opt['update_bert_opt'] > 0:
            for p in self.bert.parameters():
                p.requires_grad = False
        mem_size = self.bert_config.hidden_size
        self.decoder_opt = opt['answer_opt']
        self.scoring_list = nn.ModuleList()
        labels = [int(ls) for ls in opt['label_size'].split(',')]
        task_dropout_p = opt['tasks_dropout_p']
        self.bert_pooler = None

        for task, lab in enumerate(labels):
            decoder_opt = self.decoder_opt[task]
            dropout = DropoutWrapper(task_dropout_p[task], opt['vb_dropout'])
            self.dropout_list.append(dropout)
            if decoder_opt == 1:
                out_proj = SANClassifier(mem_size, mem_size, lab, opt, prefix='answer', dropout=dropout)
                self.scoring_list.append(out_proj)
            else:
                out_proj = nn.Linear(self.bert_config.hidden_size, lab)
                self.scoring_list.append(out_proj)

        self.opt = opt
        self._my_init()
        self.set_embed(opt)

開發者ID:RTIInternational，項目名稱:gobbli，代碼行數:31，代碼來源:matcher.py

示例12: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config, trainable=False):
        super(BertForLabelEncoding, self).__init__(config)

        self.config = config
        self.bert = BertModel(config)
        #self.apply(self.init_bert_weights)     # don't need to perform due to pre-trained params loading

        if not trainable:
            for p in self.bert.parameters():
                p.requires_grad = False

開發者ID:SKTBrain，項目名稱:SUMBT，代碼行數:12，代碼來源:BertForLabelEncoding.py

示例13: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config):
        super(BertForUtteranceEncoding, self).__init__(config)

        self.config = config
        self.bert = BertModel(config)
        #self.apply(self.init_bert_weights)     # don't need to perform due to pre-trained params loading

開發者ID:SKTBrain，項目名稱:SUMBT，代碼行數:8，代碼來源:BertForUtteranceEncoding.py

示例14: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self,config):
        super(DMBERT_Encoder,self).__init__(config)
        self.bert=BertModel(config)
        self.dropout=nn.Dropout(p=keepProb)
        #self.M=nn.Linear(EncodedDim,dimE)
        self.maxpooling=nn.MaxPool1d(SenLen)

開發者ID:thunlp，項目名稱:Adv-ED，代碼行數:8，代碼來源:models.py

示例15: init

# 需要導入模塊: from pytorch_pretrained_bert import modeling [as 別名]
# 或者: from pytorch_pretrained_bert.modeling import BertModel [as 別名]
def __init__(self, config, num_labels=3):
        super(BertForSequenceLabeling, self).__init__(config)
        self.num_labels = num_labels
        self.bert = BertModel(config)
        self.dropout = torch.nn.Dropout(config.hidden_dropout_prob)
        self.classifier = torch.nn.Linear(config.hidden_size, num_labels)
        self.apply(self.init_bert_weights)

開發者ID:howardhsu，項目名稱:BERT-for-RRC-ABSA，代碼行數:9，代碼來源:run_ae.py

注：本文中的pytorch_pretrained_bert.modeling.BertModel方法示例由純淨天空整理自Github/MSDocs等開源代碼及文檔管理平台，相關代碼片段篩選自各路編程大神貢獻的開源項目，源碼版權歸原作者所有，傳播和使用請參考對應項目的License；未經允許，請勿轉載。

示例1: forward

示例2: __init__

示例3: __init__

示例4: __init__

示例5: __init__

示例6: __init__

示例7: __init__

示例8: __init__

示例9: __init__

示例10: __init__

示例11: __init__

示例12: __init__

示例13: __init__

示例14: __init__

示例15: __init__