modeling.py

# coding=utf-8
# Copyright 2018 The Google AI Language Team Authors and The HugginFace Inc. team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""PyTorch BERT model."""

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import copy
import json
import math
import six
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.nn import CrossEntropyLoss, BCEWithLogitsLoss, BCELoss
from transformers import BertModel, BertTokenizer
import random

device = "cuda"

def gelu(x):
    """Implementation of the gelu activation function.
        For information: OpenAI GPT's gelu is slightly different (and gives slightly different results):
        0.5 * x * (1 + torch.tanh(math.sqrt(2 / math.pi) * (x + 0.044715 * torch.pow(x, 3))))
    """
    return x * 0.5 * (1.0 + torch.erf(x / math.sqrt(2.0)))


class BertConfig(object):
    """Configuration class to store the configuration of a `BertModel`.
    """
    def __init__(self,
                vocab_size,
                hidden_size=768,
                num_hidden_layers=12,
                num_attention_heads=12,
                intermediate_size=3072,
                hidden_act="gelu",
                hidden_dropout_prob=0.1,
                attention_probs_dropout_prob=0.1,
                max_position_embeddings=512,
                type_vocab_size=16,
                initializer_range=0.02):
        """Constructs BertConfig.

        Args:
            vocab_size: Vocabulary size of `inputs_ids` in `BertModel`.
            hidden_size: Size of the encoder layers and the pooler layer.
            num_hidden_layers: Number of hidden layers in the Transformer encoder.
            num_attention_heads: Number of attention heads for each attention layer in
                the Transformer encoder.
            intermediate_size: The size of the "intermediate" (i.e., feed-forward)
                layer in the Transformer encoder.
            hidden_act: The non-linear activation function (function or string) in the
                encoder and pooler.
            hidden_dropout_prob: The dropout probabilitiy for all fully connected
                layers in the embeddings, encoder, and pooler.
            attention_probs_dropout_prob: The dropout ratio for the attention
                probabilities.
            max_position_embeddings: The maximum sequence length that this model might
                ever be used with. Typically set this to something large just in case
                (e.g., 512 or 1024 or 2048).
            type_vocab_size: The vocabulary size of the `token_type_ids` passed into
                `BertModel`.
            initializer_range: The sttdev of the truncated_normal_initializer for
                initializing all weight matrices.
        """
        self.vocab_size = vocab_size
        self.hidden_size = hidden_size
        self.num_hidden_layers = num_hidden_layers
        self.num_attention_heads = num_attention_heads
        self.hidden_act = hidden_act
        self.intermediate_size = intermediate_size
        self.hidden_dropout_prob = hidden_dropout_prob
        self.attention_probs_dropout_prob = attention_probs_dropout_prob
        self.max_position_embeddings = max_position_embeddings
        self.type_vocab_size = type_vocab_size
        self.initializer_range = initializer_range

    @classmethod
    def from_dict(cls, json_object):
        """Constructs a `BertConfig` from a Python dictionary of parameters."""
        config = BertConfig(vocab_size=None)
        for (key, value) in six.iteritems(json_object):
            config.__dict__[key] = value
        return config

    @classmethod
    def from_json_file(cls, json_file):
        """Constructs a `BertConfig` from a json file of parameters."""
        with open(json_file, "r") as reader:
            text = reader.read()
        return cls.from_dict(json.loads(text))

    def to_dict(self):
        """Serializes this instance to a Python dictionary."""
        output = copy.deepcopy(self.__dict__)
        return output

    def to_json_string(self):
        """Serializes this instance to a JSON string."""
        return json.dumps(self.to_dict(), indent=2, sort_keys=True) + "\n"


class BERTLayerNorm(nn.Module):
    def __init__(self, config, variance_epsilon=1e-12):
        """Construct a layernorm module in the TF style (epsilon inside the square root).
        """
        super(BERTLayerNorm, self).__init__()
        self.gamma = nn.Parameter(torch.ones(config.hidden_size))
        self.beta = nn.Parameter(torch.zeros(config.hidden_size))
        self.variance_epsilon = variance_epsilon

    def forward(self, x):
        u = x.mean(-1, keepdim=True)
        s = (x - u).pow(2).mean(-1, keepdim=True)
        x = (x - u) / torch.sqrt(s + self.variance_epsilon)
        return self.gamma * x + self.beta

class BERTEmbeddings(nn.Module):
    def __init__(self, config):
        super(BERTEmbeddings, self).__init__()
        """Construct the embedding module from word, position and token_type embeddings.
        """
        self.word_embeddings = nn.Embedding(config.vocab_size, config.hidden_size)
        self.position_embeddings = nn.Embedding(config.max_position_embeddings, config.hidden_size)
        self.token_type_embeddings = nn.Embedding(config.type_vocab_size, config.hidden_size)

        # self.LayerNorm is not snake-cased to stick with TensorFlow model variable name and be able to load
        # any TensorFlow checkpoint file
        self.LayerNorm = BERTLayerNorm(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)            

    def forward(self, input_ids, token_type_ids=None):
        seq_length = input_ids.size(1)
        position_ids = torch.arange(seq_length, dtype=torch.long, device=input_ids.device)
        position_ids = position_ids.unsqueeze(0).expand_as(input_ids)
        if token_type_ids is None:
            token_type_ids = torch.zeros_like(input_ids)

        words_embeddings = self.word_embeddings(input_ids)
        position_embeddings = self.position_embeddings(position_ids)
        token_type_embeddings = self.token_type_embeddings(token_type_ids)

        embeddings = words_embeddings + position_embeddings + token_type_embeddings
        embeddings = self.LayerNorm(embeddings)
        embeddings = self.dropout(embeddings)
        return embeddings


class BERTSelfAttention(nn.Module):
    def __init__(self, config):
        super(BERTSelfAttention, self).__init__()
        if config.hidden_size % config.num_attention_heads != 0:
            raise ValueError(
                "The hidden size (%d) is not a multiple of the number of attention "
                "heads (%d)" % (config.hidden_size, config.num_attention_heads))
        self.num_attention_heads = config.num_attention_heads
        self.attention_head_size = int(config.hidden_size / config.num_attention_heads)
        self.all_head_size = self.num_attention_heads * self.attention_head_size

        self.query = nn.Linear(config.hidden_size, self.all_head_size)
        self.key = nn.Linear(config.hidden_size, self.all_head_size)
        self.value = nn.Linear(config.hidden_size, self.all_head_size)

        self.dropout = nn.Dropout(config.attention_probs_dropout_prob)

    def transpose_for_scores(self, x):
        new_x_shape = x.size()[:-1] + (self.num_attention_heads, self.attention_head_size)
        x = x.view(*new_x_shape)
        return x.permute(0, 2, 1, 3)

    def forward(self, hidden_states, attention_mask):
        mixed_query_layer = self.query(hidden_states)
        mixed_key_layer = self.key(hidden_states)
        mixed_value_layer = self.value(hidden_states)

        query_layer = self.transpose_for_scores(mixed_query_layer)
        key_layer = self.transpose_for_scores(mixed_key_layer)
        value_layer = self.transpose_for_scores(mixed_value_layer)

        # Take the dot product between "query" and "key" to get the raw attention scores.
        attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))
        attention_scores = attention_scores / math.sqrt(self.attention_head_size)
        # Apply the attention mask is (precomputed for all layers in BertModel forward() function)
        attention_scores = attention_scores + attention_mask

        # Normalize the attention scores to probabilities.
        attention_probs = nn.Softmax(dim=-1)(attention_scores)

        # This is actually dropping out entire tokens to attend to, which might
        # seem a bit unusual, but is taken from the original Transformer paper.
        attention_probs = self.dropout(attention_probs)

        context_layer = torch.matmul(attention_probs, value_layer)
        context_layer = context_layer.permute(0, 2, 1, 3).contiguous()
        new_context_layer_shape = context_layer.size()[:-2] + (self.all_head_size,)
        context_layer = context_layer.view(*new_context_layer_shape)
        return context_layer


class BERTSelfOutput(nn.Module):
    def __init__(self, config):
        super(BERTSelfOutput, self).__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.LayerNorm = BERTLayerNorm(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)

    def forward(self, hidden_states, input_tensor):
        hidden_states = self.dense(hidden_states)
        hidden_states = self.dropout(hidden_states)
        hidden_states = self.LayerNorm(hidden_states + input_tensor)
        return hidden_states


class BERTAttention(nn.Module):
    def __init__(self, config):
        super(BERTAttention, self).__init__()
        self.self = BERTSelfAttention(config)
        self.output = BERTSelfOutput(config)

    def forward(self, input_tensor, attention_mask):
        self_output = self.self(input_tensor, attention_mask)
        attention_output = self.output(self_output, input_tensor)
        return attention_output


class BERTIntermediate(nn.Module):
    def __init__(self, config):
        super(BERTIntermediate, self).__init__()
        self.dense = nn.Linear(config.hidden_size, config.intermediate_size)
        self.intermediate_act_fn = gelu

    def forward(self, hidden_states):
        hidden_states = self.dense(hidden_states)
        hidden_states = self.intermediate_act_fn(hidden_states)
        return hidden_states


class BERTOutput(nn.Module):
    def __init__(self, config):
        super(BERTOutput, self).__init__()
        self.dense = nn.Linear(config.intermediate_size, config.hidden_size)
        self.LayerNorm = BERTLayerNorm(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)

    def forward(self, hidden_states, input_tensor):
        hidden_states = self.dense(hidden_states)
        hidden_states = self.dropout(hidden_states)
        hidden_states = self.LayerNorm(hidden_states + input_tensor)
        return hidden_states


class BERTLayer(nn.Module):
    def __init__(self, config):
        super(BERTLayer, self).__init__()
        self.attention = BERTAttention(config)
        self.intermediate = BERTIntermediate(config)
        self.output = BERTOutput(config)

    def forward(self, hidden_states, attention_mask):
        attention_output = self.attention(hidden_states, attention_mask)
        intermediate_output = self.intermediate(attention_output)
        layer_output = self.output(intermediate_output, attention_output)
        return layer_output


class BERTEncoder(nn.Module):
    def __init__(self, config):
        super(BERTEncoder, self).__init__()
        layer = BERTLayer(config)
        self.layer = nn.ModuleList([copy.deepcopy(layer) for _ in range(config.num_hidden_layers)])    

    def forward(self, hidden_states, attention_mask):
        all_encoder_layers = []
        for layer_module in self.layer:
            hidden_states = layer_module(hidden_states, attention_mask)
            all_encoder_layers.append(hidden_states)
        return all_encoder_layers


class BERTPooler(nn.Module):
    def __init__(self, config):
        super(BERTPooler, self).__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.activation = nn.Tanh()

    def forward(self, hidden_states):
        # We "pool" the model by simply taking the hidden state corresponding
        # to the first token.
        first_token_tensor = hidden_states[:, 0]
        pooled_output = self.dense(first_token_tensor)
        pooled_output = self.activation(pooled_output)
        return pooled_output


class BertModel(nn.Module):
    """BERT model ("Bidirectional Embedding Representations from a Transformer").

    Example usage:
    ```python
    # Already been converted into WordPiece token ids
    input_ids = torch.LongTensor([[31, 51, 99], [15, 5, 0]])
    input_mask = torch.LongTensor([[1, 1, 1], [1, 1, 0]])
    token_type_ids = torch.LongTensor([[0, 0, 1], [0, 2, 0]])

    config = modeling.BertConfig(vocab_size=32000, hidden_size=512,
        num_hidden_layers=8, num_attention_heads=6, intermediate_size=1024)

    model = modeling.BertModel(config=config)
    all_encoder_layers, pooled_output = model(input_ids, token_type_ids, input_mask)
    ```
    """
    def __init__(self, config: BertConfig):
        """Constructor for BertModel.

        Args:
            config: `BertConfig` instance.
        """
        super(BertModel, self).__init__()
        self.embeddings = BERTEmbeddings(config)
        self.encoder = BERTEncoder(config)
        self.pooler = BERTPooler(config)

    def forward(self, input_ids, token_type_ids=None, attention_mask=None):
        if attention_mask is None:
            attention_mask = torch.ones_like(input_ids)
        if token_type_ids is None:
            token_type_ids = torch.zeros_like(input_ids)

        # We create a 3D attention mask from a 2D tensor mask.
        # Sizes are [batch_size, 1, 1, to_seq_length]
        # So we can broadcast to [batch_size, num_heads, from_seq_length, to_seq_length]
        # this attention mask is more simple than the triangular masking of causal attention
        # used in OpenAI GPT, we just need to prepare the broadcast dimension here.
        extended_attention_mask = attention_mask.unsqueeze(1).unsqueeze(2)

        # Since attention_mask is 1.0 for positions we want to attend and 0.0 for
        # masked positions, this operation will create a tensor which is 0.0 for
        # positions we want to attend and -10000.0 for masked positions.
        # Since we are adding it to the raw scores before the softmax, this is
        # effectively the same as removing these entirely.
        #extended_attention_mask = extended_attention_mask.float()
        extended_attention_mask = extended_attention_mask.to(dtype=next(self.parameters()).dtype) # fp16 compatibility
        extended_attention_mask = (1.0 - extended_attention_mask) * -10000.0

        embedding_output = self.embeddings(input_ids, token_type_ids)
        all_encoder_layers = self.encoder(embedding_output, extended_attention_mask)
        sequence_output = all_encoder_layers[-1]
        pooled_output = self.pooler(sequence_output)
        return all_encoder_layers, pooled_output

class BertForSequenceClassification(nn.Module):
    def __init__(self, config, num_labels):
        super(BertForSequenceClassification, self).__init__()
        self.bert = BertModel(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        # self.classifier = nn.Linear(config.hidden_size, num_labels * 36)
        self.classifier = nn.Linear(2*config.hidden_size, num_labels)
        self.proj_trigger = nn.Linear(config.hidden_size, 2)

        def init_weights(module):
            if isinstance(module, (nn.Linear, nn.Embedding)):
                # Slightly different from the TF version which uses truncated_normal for initialization
                # cf https://github.com/pytorch/pytorch/pull/5617
                module.weight.data.normal_(mean=0.0, std=config.initializer_range)
            elif isinstance(module, BERTLayerNorm):
                module.beta.data.normal_(mean=0.0, std=config.initializer_range)
                module.gamma.data.normal_(mean=0.0, std=config.initializer_range)
            if isinstance(module, nn.Linear):
                module.bias.data.zero_()
        self.apply(init_weights)

    def forward(self, input_ids, token_type_ids, attention_mask, labels=None, n_class=1, pos_weight=None, input_t_idx=None, input_x_idx=None):
        seq_length = input_ids.size(2)
        all_encoder_layers, pooled_output = self.bert(input_ids.view(-1,seq_length),
                                     token_type_ids.view(-1,seq_length),
                                     attention_mask.view(-1,seq_length))
        
        last_hidden_states = all_encoder_layers[-1] # [6, 512, 768]
        # print(pooled_output.shape) # [6,768]
        # pooled_output = self.dropout(pooled_output)




        ##############################################################################
        start_end_logit = self.proj_trigger(last_hidden_states)
        input_x_idx = input_x_idx.view(-1,2)
        
        ids = []
        for x_idx in input_x_idx:
            ids.append([0, x_idx[0]-1])
        masked_start_end_logit = self.get_masked(
            start_end_logit,
            torch.tensor(ids),
            mask_val=float('-inf')
        )

        ids = self.get_triggers_ids(masked_start_end_logit)

        #--------------------------------------------------------------------------------
        # tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
        # p_trigs, gt_trigs = [], []
        # for i in range(len(input_ids.view(-1,seq_length))):
        #     p_trig = tokenizer.decode(input_ids.view(-1,seq_length)[i][ids[i][0] : ids[i][1]])
        #     p_trigs.append(p_trig)
        #     gt_trig = tokenizer.decode(input_ids.view(-1,seq_length)[i]\
        #                          [input_t_idx.view(-1,2)[i][0] : input_t_idx.view(-1,2)[i][1]])
        #     gt_trigs.append(gt_trig)
            
        #     cls_tri = input_ids.view(-1,seq_length)[i][input_t_idx.view(-1,2)[i][0] : input_t_idx.view(-1,2)[i][1]]
            # print(gt_trig)
            # print(input_t_idx)
            # print()
        #--------------------------------------------------------------------------------
        # print(input_t_idx.shape)
        # print(input_t_idx.view(-1, 2).shape)
        # print(len(input_t_idx.view(-1,2)))
        # print(input_t_idx.view(-1,2))

        cls_tri = []
        for i in range(len(input_t_idx.view(-1,2))):
            cls_tri.append([0, 1])
        cls_tris = torch.tensor(cls_tri, dtype=torch.long).to(device)
        # print(cls_tris.shape)
        # print(cls_tris.view(-1,2).shape)
        # print(cls_tris)
        # exit()
        #--------------------------------------------------------------------------------

        # Training
        # teacher_train = 0
        # if random.randint(0,100)>70:
        #     teacher_train = 1
        #     trigger = self.attention(last_hidden_states, input_t_idx.view(-1, 2))
        # else:
        #     trigger = self.attention(last_hidden_states, torch.tensor(ids))
        


        # Inference
        # trigger = self.attention(last_hidden_states, torch.tensor(ids))
        # ground truth
        # trigger = self.attention(last_hidden_states, input_t_idx.view(-1, 2))
        # [CLS] 
        trigger = self.attention(last_hidden_states, cls_tris)






        x = []
        for b_idx in range(len(last_hidden_states)):
            x.append(last_hidden_states[b_idx][input_x_idx[b_idx][1]+1, :])
        x = torch.vstack(x)


        concat_hid = torch.hstack((trigger, last_hidden_states[:, 0, :]))
        
        
        logits = self.classifier(concat_hid)
        # binary_logit = self.proj_binary(x)
        ##############################################################################


        logits = logits.view(-1)
        if labels is not None:
            loss_fct = BCEWithLogitsLoss(pos_weight=pos_weight)
            loss_fn = nn.CrossEntropyLoss()
            labels = labels.view(-1)
            
            loss = loss_fct(logits, labels)
            
            
            
            # -------------------------------------------------------
            # Training
            # if teacher_train==1:
            #     trigger_loss = 0
            # else:
            #     trigger_loss = loss_fn(start_end_logit, input_t_idx.view(-1,2))
            # -------------------------------------------------------

            # Inference
            trigger_loss = loss_fn(start_end_logit, input_t_idx.view(-1,2))
            
            
            
            
            # total_loss = loss
            total_loss = loss + 0.3*trigger_loss
            # print("total")
            # print(" total loss:{:.2f} relation loss:{:.2f} trigger loss:{:.2f}".format(total_loss, loss, trigger_loss))

            # return total_loss, logits, p_trigs, gt_trigs
            return total_loss, logits
        else:
            return logits
            # return logits, p_trigs, gt_trigs


    def get_triggers_ids(self, masked_start_end_logit, tri_len=None):
        ids = []

        if tri_len is None:
            for batch_idx, sample in enumerate(masked_start_end_logit):
                start = sample[:,0] # start: shape (512)
                end = sample[:,1]
                start_candidates = torch.topk(start, k=30)
                end_candidates = torch.topk(end, k=30)
                ans_candidates = [(0, 1)]
                scores = [-100]
                start_logits = F.softmax(start_candidates[0])
                end_logits = F.softmax(end_candidates[0])
                for i, s in enumerate(start_candidates[1]):
                    for j, e in enumerate(end_candidates[1]):
                        if s == 0:
                            ans_candidates.append((s, s+1))
                            scores.append(start_logits[i] * end_logits[j])
                        if s<e and e-s <= 10:
                            ans_candidates.append((s, e))
                            scores.append(start_logits[i] * end_logits[j])
                results = list(zip(scores, ans_candidates))
                results.sort()
                results.reverse()

                ids.append([int(results[0][1][0]), int(results[0][1][1])])
            return ids
        else:
            for batch_idx, sample in enumerate(masked_start_end_logit):
                start = sample[:,0] # start: shape (512)
                end = sample[:,1]
                start_logits = F.softmax(start)
                end_logits = F.softmax(end)
                max_score = float('-inf')
                cand = None
                for i in range(len(start_logits)-tri_len[batch_idx]):
                    cur_score = start_logits[i] + end_logits[i+tri_len[batch_idx]]
                    if cur_score > max_score:
                        max_score = cur_score
                        cand = [i, i+tri_len[batch_idx]]
                ids.append(cand)
            return ids
    
    def attention(self, mat, ids):
        triggers = []
        batch_size, _, _ = mat.shape
        for b_id in range(batch_size):
            trigger = mat[b_id][ids[b_id][0] : ids[b_id][1]][:]
            score = []
            cls = mat[b_id, 0, :]
            for j in range(len(trigger)):
                score.append(torch.dot(cls, trigger[j]))
            score = torch.tensor(score, device=device)
            score = F.softmax(score)
            triggers.append(torch.matmul(trigger.T, score))
        return torch.vstack(triggers)

    def get_masked(self, mat, ids, mask_val=0):
        batch_size, seq_len, cls = mat.shape
        mask = torch.ones(batch_size, seq_len, cls)
        for i in range(batch_size):
            mask[i, ids[i][0]:ids[i][1], :] = 0
        mask = mask.bool()
        return mat.masked_fill(mask.to(device), mask_val)