modeling.py

# coding=utf-8
# Copyright (c) 2021, Habana Labs Ltd.  All rights reserved.
# Copyright (c) 2019 NVIDIA CORPORATION. All rights reserved.
# 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, division, print_function, unicode_literals

import copy
import json
import logging
import math
import os
import shutil
import tarfile
import tempfile
import sys
from io import open

import torch
from torch import nn
from torch.nn import CrossEntropyLoss
from deepspeed.runtime.activation_checkpointing.checkpointing import checkpoint

sys.path.append('/workspace/bert/')

from torch.nn import Module
from torch.nn.parameter import Parameter
import torch.nn.functional as F
import torch.nn.init as init

logger = logging.getLogger(__name__)

PRETRAINED_MODEL_ARCHIVE_MAP = {
    'bert-base-uncased': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-uncased.tar.gz",
    'bert-large-uncased': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-uncased.tar.gz",
    'bert-base-cased': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-cased.tar.gz",
    'bert-large-cased': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-large-cased.tar.gz",
    'bert-base-multilingual-uncased': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-multilingual-uncased.tar.gz",
    'bert-base-multilingual-cased': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-multilingual-cased.tar.gz",
    'bert-base-chinese': "https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-chinese.tar.gz",
}
CONFIG_NAME = 'bert_config.json'
WEIGHTS_NAME = 'pytorch_model.bin'
TF_WEIGHTS_NAME = 'model.ckpt'

def load_tf_weights_in_bert(model, tf_checkpoint_path):
    """ Load tf checkpoints in a pytorch model
    """
    try:
        import re
        import numpy as np
        import tensorflow as tf
    except ImportError:
        print("Loading a TensorFlow models in PyTorch, requires TensorFlow to be installed. Please see "
            "https://www.tensorflow.org/install/ for installation instructions.")
        raise
    tf_path = os.path.abspath(tf_checkpoint_path)
    print("Converting TensorFlow checkpoint from {}".format(tf_path))
    # Load weights from TF model
    init_vars = tf.train.list_variables(tf_path)
    names = []
    arrays = []
    for name, shape in init_vars:
        print("Loading TF weight {} with shape {}".format(name, shape))
        array = tf.train.load_variable(tf_path, name)
        names.append(name)
        arrays.append(array)

    for name, array in zip(names, arrays):
        name = name.split('/')
        # adam_v and adam_m are variables used in AdamWeightDecayOptimizer to calculated m and v
        # which are not required for using pretrained model
        if any(n in ["adam_v", "adam_m"] for n in name):
            print("Skipping {}".format("/".join(name)))
            continue
        pointer = model
        for m_name in name:
            if re.fullmatch(r'[A-Za-z]+_\d+', m_name):
                l = re.split(r'_(\d+)', m_name)
            else:
                l = [m_name]
            if l[0] == 'kernel' or l[0] == 'gamma':
                pointer = getattr(pointer, 'weight')
            elif l[0] == 'output_bias' or l[0] == 'beta':
                pointer = getattr(pointer, 'bias')
            elif l[0] == 'output_weights':
                pointer = getattr(pointer, 'weight')
            else:
                pointer = getattr(pointer, l[0])
            if len(l) >= 2:
                num = int(l[1])
                pointer = pointer[num]
        if m_name[-11:] == '_embeddings':
            pointer = getattr(pointer, 'weight')
        elif m_name == 'kernel':
            array = np.ascontiguousarray(np.transpose(array))
        try:
            assert pointer.shape == array.shape
        except AssertionError as e:
            e.args += (pointer.shape, array.shape)
            raise
        print("Initialize PyTorch weight {}".format(name))
        pointer.data = torch.from_numpy(array)
    return model

def gelu(x):
    if torch.cuda.is_available():
        return x * 0.5 * (1.0 + torch.erf(x / 1.41421))
    else:
        return F.gelu(x)

#used only for triton inference
def bias_gelu(bias, y):
    x = bias + y
    return x * 0.5 * (1.0 + torch.erf(x / 1.41421))

# used specifically for training since torch.nn.functional.gelu breaks ONNX export
def bias_gelu_training(bias, y):
    x = bias + y
    return torch.nn.functional.gelu(x) # Breaks ONNX export

def bias_tanh(bias, y):
    x = bias + y
    return torch.tanh(x)

def swish(x):
    return x * torch.sigmoid(x)

def tanh(x):
    return  torch.tanh(x)

#torch.nn.functional.gelu(x) # Breaks ONNX export
ACT2FN = {"gelu": gelu, "bias_gelu": bias_gelu, "bias_tanh": bias_tanh, "relu": torch.nn.functional.relu, "swish": swish, "tanh": tanh}

class LinearActivation(Module):
    r"""Fused Linear and activation Module.
    """
    __constants__ = ['bias']

    def __init__(self, in_features, out_features, act='gelu', bias=True):
        super(LinearActivation, self).__init__()
        self.in_features = in_features
        self.out_features = out_features
        # setting act_fn to nn.Identity caused issues when re-assigning to gelu.Hence set to None
        #self.act_fn = nn.Identity()                                                        #
        self.act_fn = None                                                                  #
        self.biased_act_fn = None                                                           #
        self.bias = None                                                                    #
        if isinstance(act, str) or (sys.version_info[0] == 2 and isinstance(act, unicode)): # For TorchScript
            if bias and not 'bias' in act and torch.cuda.is_available():                    # compatibility
                act = 'bias_' + act                                                         #
                self.biased_act_fn = ACT2FN[act]                                            #

            else:
                self.act_fn = ACT2FN[act]
        else:
            self.act_fn = act
        self.weight = Parameter(torch.Tensor(out_features, in_features))
        if bias:
            self.bias = Parameter(torch.Tensor(out_features))
        else:
            self.register_parameter('bias', None)
        self.reset_parameters()

    def reset_parameters(self):
        init.kaiming_uniform_(self.weight, a=math.sqrt(5))
        if self.bias is not None:
            fan_in, _ = init._calculate_fan_in_and_fan_out(self.weight)
            bound = 1 / math.sqrt(fan_in)
            init.uniform_(self.bias, -bound, bound)

    def forward(self, input):
        if not self.bias is None and torch.cuda.is_available():
            return self.biased_act_fn(self.bias, F.linear(input, self.weight, None))
        else:
            return self.act_fn(F.linear(input, self.weight, self.bias))

    def extra_repr(self):
        return 'in_features={}, out_features={}, bias={}'.format(
            self.in_features, self.out_features, self.bias is not None
        )


class BertConfig(object):
    """Configuration class to store the configuration of a `BertModel`.
    """
    def __init__(self,
                 vocab_size_or_config_json_file,
                 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=2,
                 initializer_range=0.02,
                 output_all_encoded_layers=False,
                 layer_norm_large_model=False,
                 gated_layers=()):
        """Constructs BertConfig.

        Args:
            vocab_size_or_config_json_file: 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. If string, "gelu", "relu" and "swish" are supported.
            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.
            gated_layers: iterable with layer numbers that should be gated, i.e. BertGatedLayer
        """
        if isinstance(vocab_size_or_config_json_file, str) or (sys.version_info[0] == 2
                        and isinstance(vocab_size_or_config_json_file, unicode)):
            with open(vocab_size_or_config_json_file, "r", encoding='utf-8') as reader:
                json_config = json.loads(reader.read())
            for key, value in json_config.items():
                self.__dict__[key] = value
        elif isinstance(vocab_size_or_config_json_file, int):
            self.vocab_size = vocab_size_or_config_json_file
            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
            self.output_all_encoded_layers = output_all_encoded_layers
            self.layer_norm_large_model = layer_norm_large_model
            self.gated_layers = gated_layers
        else:
            raise ValueError("First argument must be either a vocabulary size (int)"
                             "or the path to a pretrained model config file (str)")

    @classmethod
    def from_dict(cls, json_object):
        """Constructs a `BertConfig` from a Python dictionary of parameters."""
        config = BertConfig(vocab_size_or_config_json_file=-1)
        for key, value in json_object.items():
            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", encoding='utf-8') as reader:
            text = reader.read()
        return cls.from_dict(json.loads(text))

    def __repr__(self):
        return str(self.to_json_string())

    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 BertNonFusedLayerNorm(nn.Module):
    def __init__(self, hidden_size, eps=1e-12):
        """Construct a layernorm module in the TF style (epsilon inside the square root).
        """
        super(BertNonFusedLayerNorm, self).__init__()
        self.weight = nn.Parameter(torch.ones(hidden_size))
        self.bias = nn.Parameter(torch.zeros(hidden_size))
        self.variance_epsilon = eps

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

try:
    import apex
    #apex.amp.register_half_function(apex.normalization.fused_layer_norm, 'FusedLayerNorm')
    import apex.normalization
    from apex.normalization.fused_layer_norm import FusedLayerNormAffineFunction
    #apex.amp.register_float_function(apex.normalization.FusedLayerNorm, 'forward')
    #BertLayerNorm = apex.normalization.FusedLayerNorm
    APEX_IS_AVAILABLE = True
except ImportError:
    if torch.cuda.is_available():
        print("Better speed can be achieved with apex installed from https://www.github.com/nvidia/apex.")
    #BertLayerNorm = BertNonFusedLayerNorm
    APEX_IS_AVAILABLE = False

if torch.cuda.is_available():
    class BertLayerNorm(Module):
        def __init__(self, hidden_size, eps=1e-12):
            super(BertLayerNorm, self).__init__()
            self.shape = torch.Size((hidden_size,))
            self.eps = eps
            self.weight = nn.Parameter(torch.ones(hidden_size))
            self.bias = nn.Parameter(torch.zeros(hidden_size))
            self.apex_enabled = APEX_IS_AVAILABLE

        @torch.jit.unused
        def fused_layer_norm(self, x):
            return FusedLayerNormAffineFunction.apply(
                        x, self.weight, self.bias, self.shape, self.eps)


        def forward(self, x):
            if self.apex_enabled and not torch.jit.is_scripting():
                x = self.fused_layer_norm(x)
            else:
                u = x.mean(-1, keepdim=True)
                s = (x - u)
                s = s * s
                s = s.mean(-1, keepdim=True)
                x = (x - u) / torch.sqrt(s + self.eps)
                x = self.weight * x + self.bias
            return x

else:
    BertLayerNorm = torch.nn.LayerNorm

class BertEmbeddings(nn.Module):
    """Construct the embeddings from word, position and token_type embeddings.
    """
    def __init__(self, config):
        super(BertEmbeddings, self).__init__()
        self.word_embeddings = nn.Embedding(config.vocab_size, config.hidden_size, padding_idx=0)
        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.hidden_size, eps=1e-12)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)

    def forward(self, input_ids, token_type_ids):
        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)

        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 = torch.reshape(x, new_x_shape)
        return x.permute(0, 2, 1, 3)

    def transpose_key_for_scores(self, x):
        new_x_shape = x.size()[:-1] + (self.num_attention_heads, self.attention_head_size)
        x = torch.reshape(x, new_x_shape)
        return x.permute(0, 2, 3, 1)

    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_key_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)
        attention_scores = attention_scores / math.sqrt(self.attention_head_size)
        # Apply the attention mask is (precomputed for all layers in BertModel forward() function)
        # make sure attention_scores has the correct dtype
        attention_scores = attention_scores + attention_mask

        # Normalize the attention scores to probabilities.
        attention_probs = F.softmax(attention_scores, dim=-1)

        # 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)
        context_layer = context_layer.flatten(start_dim=2, end_dim=3)
        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.hidden_size, eps=1e-12)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.layer_norm_large_model = config.layer_norm_large_model

    def forward(self, hidden_states, output_layer_norm_input, input_tensor):
        hidden_states = self.dense(hidden_states)
        hidden_states = self.dropout(hidden_states)
        if not self.layer_norm_large_model:
            hidden_states = self.LayerNorm(hidden_states+input_tensor)
            attention_layer_norm_input = 0
        else:
            attention_layer_norm_input = hidden_states+output_layer_norm_input
            hidden_states = self.LayerNorm(attention_layer_norm_input)
        return hidden_states, attention_layer_norm_input


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, output_layer_norm_input, attention_mask):
        self_output = self.self(input_tensor, attention_mask)
        attention_output, layer_norm_input = self.output(self_output, output_layer_norm_input, input_tensor)
        return attention_output, layer_norm_input


class BertIntermediate(nn.Module):
    def __init__(self, config):
        super(BertIntermediate, self).__init__()
        self.dense_act = LinearActivation(config.hidden_size, config.intermediate_size, act=config.hidden_act)

    def forward(self, hidden_states):
        hidden_states = self.dense_act(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.hidden_size, eps=1e-12)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.layer_norm_large_model = config.layer_norm_large_model
        self.gate_grads_from_next_layer = False

    def forward(self, hidden_states,attention_layer_norm_input, input_tensor):
        hidden_states = self.dense(hidden_states)
        hidden_states = self.dropout(hidden_states)
        if not self.layer_norm_large_model:
            hidden_states = self.LayerNorm(hidden_states + input_tensor)
            output_layer_norm_input = 0
        else:
            output_layer_norm_input = hidden_states + attention_layer_norm_input
            hidden_states = output_layer_norm_input
            if self.gate_grads_from_next_layer:
                hidden_states = GateBackward(hidden_states)
            hidden_states = self.LayerNorm(hidden_states)
        return hidden_states, output_layer_norm_input


class BertLayer(nn.Module):
    def __init__(self, config):
        super(BertLayer, self).__init__()
        self.attention = BertAttention(config)
        self.intermediate = BertIntermediate(config)
        self.output = BertOutput(config)
        self.layer_norm_large_model = config.layer_norm_large_model

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


class BertEncoder(nn.Module):
    def __init__(self, config):
        super(BertEncoder, self).__init__()
        self._sanitize_gated_configuration(config)
        self.layer = nn.ModuleList([BertLayer(config) if idx not in config.gated_layers else BertGatedLayer(config)
                                    for idx in range(config.num_hidden_layers)])
        for idx in range(config.num_hidden_layers):
            self.layer[idx].output.gate_grads_from_next_layer = (idx + 1) in config.gated_layers
        self.output_all_encoded_layers = config.output_all_encoded_layers
        self._checkpoint_activations = False
        self._checkpoint_activations_interval = 1

    @staticmethod
    def _sanitize_gated_configuration(config):
        if not config.gated_layers:
            return

        gated_layers = list(set(config.gated_layers))
        if min(gated_layers) < 0:
            raise RuntimeError(f"Invalid gated_layers={gated_layers} - Negative layer detected")
        if max(gated_layers) >= config.num_hidden_layers:
            raise RuntimeError(f"Invalid gated_layers={gated_layers} for num_hidden_layers={config.num_hidden_layers}")

        n_non_gated = config.num_hidden_layers - len(gated_layers)
        if n_non_gated == 0:
            raise RuntimeError("Not supported to have all layers gated")

        # layer 0 gating is not supported
        # this is due to pre-ln implementation where layer N performs the Pre-LN for the MHA of the next layer
        if 0 in gated_layers:
            raise RuntimeError("Layer 0 must not be gated")

        gated_layers.sort()
        config.gated_layers = gated_layers
        logger.info(f"Using BertEncoder with gated_layers={config.gated_layers} n_non_gated={n_non_gated}")

    @torch.jit.unused
    def checkpointed_forward(self, hidden_states, layer_norm_input, attention_mask, interval):
        def custom(start, end):
            def custom_forward(*inputs):
                layers = self.layer[start:end]
                x_, layer_norm_input_, attention_mask_ = inputs
                for layer in layers:
                    x_, layer_norm_input_ = layer(x_, layer_norm_input_, attention_mask_)
                return x_, layer_norm_input_
            return custom_forward

        l = 0
        num_layers = len(self.layer)
        while l < num_layers:
            hidden_states, layer_norm_input = checkpoint(custom(l, l+interval),
                                                         hidden_states, layer_norm_input, attention_mask*1)
            l += interval

        return hidden_states, layer_norm_input

    def forward(self, hidden_states, attention_mask):
        all_encoder_layers = []

        layer_norm_input = 0
        if self._checkpoint_activations:
            hidden_states, layer_norm_input = self.checkpointed_forward(
                hidden_states, layer_norm_input, attention_mask, self._checkpoint_activations_interval)
        else:
            for i, layer_module in enumerate(self.layer):
                hidden_states, layer_norm_input = layer_module(hidden_states, layer_norm_input, attention_mask)

                if self.output_all_encoded_layers:
                    all_encoder_layers.append(hidden_states)

        if not self.output_all_encoded_layers or self._checkpoint_activations:
            all_encoder_layers.append(hidden_states)
        return all_encoder_layers


class BertPooler(nn.Module):
    def __init__(self, config):
        super(BertPooler, self).__init__()
        self.dense_act = LinearActivation(config.hidden_size, config.hidden_size, act="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_act(first_token_tensor)
        return pooled_output


class BertPredictionHeadTransform(nn.Module):
    def __init__(self, config):
        super(BertPredictionHeadTransform, self).__init__()
        self.dense_act = LinearActivation(config.hidden_size, config.hidden_size, act=config.hidden_act)
        self.LayerNorm = BertLayerNorm(config.hidden_size, eps=1e-12)

    def forward(self, hidden_states):
        hidden_states = self.dense_act(hidden_states)
        hidden_states = self.LayerNorm(hidden_states)
        return hidden_states


class BertLMPredictionHead(nn.Module):
    def __init__(self, config, bert_model_embedding_weights):
        super(BertLMPredictionHead, self).__init__()
        self.transform = BertPredictionHeadTransform(config)

        # The output weights are the same as the input embeddings, but there is
        # an output-only bias for each token.
        self.decoder = nn.Linear(bert_model_embedding_weights.size(1),
                                 bert_model_embedding_weights.size(0),
                                 bias=False)
        self.decoder.weight = bert_model_embedding_weights
        self.bias = nn.Parameter(torch.zeros(bert_model_embedding_weights.size(0)))

    def forward(self, hidden_states):
        hidden_states = self.transform(hidden_states)
        hidden_states = self.decoder(hidden_states) + self.bias
        return hidden_states


class BertOnlyMLMHead(nn.Module):
    def __init__(self, config, bert_model_embedding_weights):
        super(BertOnlyMLMHead, self).__init__()
        self.predictions = BertLMPredictionHead(config, bert_model_embedding_weights)

    def forward(self, sequence_output):
        prediction_scores = self.predictions(sequence_output)
        return prediction_scores


class BertOnlyNSPHead(nn.Module):
    def __init__(self, config):
        super(BertOnlyNSPHead, self).__init__()
        self.seq_relationship = nn.Linear(config.hidden_size, 2)

    def forward(self, pooled_output):
        seq_relationship_score = self.seq_relationship(pooled_output)
        return seq_relationship_score


class BertPreTrainingHeads(nn.Module):
    def __init__(self, config, bert_model_embedding_weights):
        super(BertPreTrainingHeads, self).__init__()
        self.predictions = BertLMPredictionHead(config, bert_model_embedding_weights)
        self.seq_relationship = nn.Linear(config.hidden_size, 2)

    def forward(self, sequence_output, pooled_output):
        prediction_scores = self.predictions(sequence_output)
        seq_relationship_score = self.seq_relationship(pooled_output)
        return prediction_scores, seq_relationship_score


class BertPreTrainedModel(nn.Module):
    """ An abstract class to handle weights initialization and
        a simple interface for dowloading and loading pretrained models.
    """
    def __init__(self, config, *inputs, **kwargs):
        super(BertPreTrainedModel, self).__init__()
        if not isinstance(config, BertConfig):
            raise ValueError(
                "Parameter config in `{}(config)` should be an instance of class `BertConfig`. "
                "To create a model from a Google pretrained model use "
                "`model = {}.from_pretrained(PRETRAINED_MODEL_NAME)`".format(
                    self.__class__.__name__, self.__class__.__name__
                ))
        self.config = config

    def init_bert_weights(self, module):
        """ Initialize the weights.
        """
        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=self.config.initializer_range)
        elif isinstance(module, BertLayerNorm):
            module.bias.data.zero_()
            module.weight.data.fill_(1.0)
        if isinstance(module, nn.Linear) and module.bias is not None:
            module.bias.data.zero_()

    def checkpoint_activations(self, enabled, interval=1):
        def _apply_flag(module):
            if hasattr(module, "_checkpoint_activations"):
                module._checkpoint_activations = enabled
                module._checkpoint_activations_interval = interval
        self.apply(_apply_flag)

    def enable_apex(self, val):
        def _apply_flag(module):
            if hasattr(module, "apex_enabled"):
                module.apex_enabled=val
        self.apply(_apply_flag)

    @classmethod
    def from_pretrained(cls, pretrained_model_name_or_path, state_dict=None, cache_dir=None,
                        from_tf=False, *inputs, **kwargs):
        """
        Instantiate a BertPreTrainedModel from a pre-trained model file or a pytorch state dict.
        Download and cache the pre-trained model file if needed.

        Params:
            pretrained_model_name_or_path: either:
                - a str with the name of a pre-trained model to load selected in the list of:
                    . `bert-base-uncased`
                    . `bert-large-uncased`
                    . `bert-base-cased`
                    . `bert-large-cased`
                    . `bert-base-multilingual-uncased`
                    . `bert-base-multilingual-cased`
                    . `bert-base-chinese`
                - a path or url to a pretrained model archive containing:
                    . `bert_config.json` a configuration file for the model
                    . `pytorch_model.bin` a PyTorch dump of a BertForPreTraining instance
                - a path or url to a pretrained model archive containing:
                    . `bert_config.json` a configuration file for the model
                    . `model.chkpt` a TensorFlow checkpoint
            from_tf: should we load the weights from a locally saved TensorFlow checkpoint
            cache_dir: an optional path to a folder in which the pre-trained models will be cached.
            state_dict: an optional state dictionnary (collections.OrderedDict object) to use instead of Google pre-trained models
            *inputs, **kwargs: additional input for the specific Bert class
                (ex: num_labels for BertForSequenceClassification)
        """
        if pretrained_model_name_or_path in PRETRAINED_MODEL_ARCHIVE_MAP:
            archive_file = PRETRAINED_MODEL_ARCHIVE_MAP[pretrained_model_name_or_path]
        else:
            archive_file = pretrained_model_name_or_path
        # redirect to the cache, if necessary
        try:
            from file_utils import cached_path
            resolved_archive_file = cached_path(archive_file, cache_dir=cache_dir)
        except EnvironmentError:
            logger.error(
                "Model name '{}' was not found in model name list ({}). "
                "We assumed '{}' was a path or url but couldn't find any file "
                "associated to this path or url.".format(
                    pretrained_model_name_or_path,
                    ', '.join(PRETRAINED_MODEL_ARCHIVE_MAP.keys()),
                    archive_file))
            return None
        if resolved_archive_file == archive_file:
            logger.info("loading archive file {}".format(archive_file))
        else:
            logger.info("loading archive file {} from cache at {}".format(
                archive_file, resolved_archive_file))
        tempdir = None
        if os.path.isdir(resolved_archive_file) or from_tf:
            serialization_dir = resolved_archive_file
        else:
            # Extract archive to temp dir
            tempdir = tempfile.mkdtemp()
            logger.info("extracting archive file {} to temp dir {}".format(
                resolved_archive_file, tempdir))
            with tarfile.open(resolved_archive_file, 'r:gz') as archive:
                def is_within_directory(directory, target):
                    abs_directory = os.path.abspath(directory)
                    abs_target = os.path.abspath(target)

                    prefix = os.path.commonprefix([abs_directory, abs_target])

                    return prefix == abs_directory

                def safe_extract(tar, path=".", members=None, *, numeric_owner=False):

                    for member in tar.getmembers():
                        member_path = os.path.join(path, member.name)
                        if not is_within_directory(path, member_path):
                            raise Exception("Attempted Path Traversal in Tar File")

                    tar.extractall(path, members, numeric_owner=numeric_owner)


                safe_extract(archive, tempdir)
            serialization_dir = tempdir
        # Load config
        config_file = os.path.join(serialization_dir, CONFIG_NAME)
        config = BertConfig.from_json_file(config_file)
        logger.info("Model config {}".format(config))
        # Instantiate model.
        model = cls(config, *inputs, **kwargs)
        if state_dict is None and not from_tf:
            weights_path = os.path.join(serialization_dir, WEIGHTS_NAME)
            state_dict = torch.load(weights_path, map_location='cpu' if not torch.cuda.is_available() else None)
        if tempdir:
            # Clean up temp dir
            shutil.rmtree(tempdir)
        if from_tf:
            # Directly load from a TensorFlow checkpoint
            weights_path = os.path.join(serialization_dir, TF_WEIGHTS_NAME)
            return load_tf_weights_in_bert(model, weights_path)
        # Load from a PyTorch state_dict
        old_keys = []
        new_keys = []
        for key in state_dict.keys():
            new_key = None
            if 'gamma' in key:
                new_key = key.replace('gamma', 'weight')
            if 'beta' in key:
                new_key = key.replace('beta', 'bias')
            if new_key:
                old_keys.append(key)
                new_keys.append(new_key)
        for old_key, new_key in zip(old_keys, new_keys):
            state_dict[new_key] = state_dict.pop(old_key)

        missing_keys = []
        unexpected_keys = []
        error_msgs = []
        # copy state_dict so _load_from_state_dict can modify it
        metadata = getattr(state_dict, '_metadata', None)
        state_dict = state_dict.copy()
        if metadata is not None:
            state_dict._metadata = metadata

        def load(module, prefix=''):
            local_metadata = {} if metadata is None else metadata.get(prefix[:-1], {})
            module._load_from_state_dict(
                state_dict, prefix, local_metadata, True, missing_keys, unexpected_keys, error_msgs)
            for name, child in module._modules.items():
                if child is not None:
                    load(child, prefix + name + '.')
        start_prefix = ''
        if not hasattr(model, 'bert') and any(s.startswith('bert.') for s in state_dict.keys()):
            start_prefix = 'bert.'
        load(model, prefix=start_prefix)
        if len(missing_keys) > 0:
            logger.info("Weights of {} not initialized from pretrained model: {}".format(
                model.__class__.__name__, missing_keys))
        if len(unexpected_keys) > 0:
            logger.info("Weights from pretrained model not used in {}: {}".format(
                model.__class__.__name__, unexpected_keys))
        if len(error_msgs) > 0:
            raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
                               model.__class__.__name__, "\n\t".join(error_msgs)))
        return model


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

    Params:
        config: a BertConfig class instance with the configuration to build a new model

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, sequence_length]
            with the word token indices in the vocabulary(see the tokens preprocessing logic in the scripts
            `extract_features.py`, `run_classifier.py` and `run_squad.py`)
        `token_type_ids`: an optional torch.LongTensor of shape [batch_size, sequence_length] with the token
            types indices selected in [0, 1]. Type 0 corresponds to a `sentence A` and type 1 corresponds to
            a `sentence B` token (see BERT paper for more details).
        `attention_mask`: an optional torch.LongTensor of shape [batch_size, sequence_length] with indices
            selected in [0, 1]. It's a mask to be used if the input sequence length is smaller than the max
            input sequence length in the current batch. It's the mask that we typically use for attention when
            a batch has varying length sentences.

    Outputs: Tuple of (encoded_layers, pooled_output)
        `encoded_layers`: controled by `output_all_encoded_layers` argument:
            - `output_all_encoded_layers=True`: outputs a list of the full sequences of encoded-hidden-states at the end
                of each attention block (i.e. 12 full sequences for BERT-base, 24 for BERT-large), each
                encoded-hidden-state is a torch.FloatTensor of size [batch_size, sequence_length, hidden_size],
            - `output_all_encoded_layers=False`: outputs only the full sequence of hidden-states corresponding
                to the last attention block of shape [batch_size, sequence_length, hidden_size],
        `pooled_output`: a torch.FloatTensor of size [batch_size, hidden_size] which is the output of a
            classifier pretrained on top of the hidden state associated to the first character of the
            input (`CLS`) to train on the Next-Sentence task (see BERT's paper).

    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, 1, 0]])

    config = modeling.BertConfig(vocab_size_or_config_json_file=32000, hidden_size=768,
        num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072)

    model = modeling.BertModel(config=config)
    all_encoder_layers, pooled_output = model(input_ids, token_type_ids, input_mask)
    ```
    """
    def __init__(self, config):
        super(BertModel, self).__init__(config)
        self.embeddings = BertEmbeddings(config)
        self.encoder = BertEncoder(config)
        self.pooler = BertPooler(config)
        self.apply(self.init_bert_weights)
        self.output_all_encoded_layers = config.output_all_encoded_layers

    def forward(self, input_ids, token_type_ids, attention_mask):
        # 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.to(dtype=self.embeddings.word_embeddings.weight.dtype) # fp16 compatibility
        extended_attention_mask = (1.0 - extended_attention_mask) * -10000.0

        embedding_output = self.embeddings(input_ids, token_type_ids)
        encoded_layers = self.encoder(embedding_output, extended_attention_mask)
        sequence_output = encoded_layers[-1]
        pooled_output = self.pooler(sequence_output)
        if not self.output_all_encoded_layers:
            encoded_layers = encoded_layers[-1:]
        return encoded_layers, pooled_output


class BertForPreTraining(BertPreTrainedModel):
    """BERT model with pre-training heads.
    This module comprises the BERT model followed by the two pre-training heads:
        - the masked language modeling head, and
        - the next sentence classification head.

    Params:
        config: a BertConfig class instance with the configuration to build a new model.

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, sequence_length]
            with the word token indices in the vocabulary(see the tokens preprocessing logic in the scripts
            `extract_features.py`, `run_classifier.py` and `run_squad.py`)
        `token_type_ids`: an optional torch.LongTensor of shape [batch_size, sequence_length] with the token
            types indices selected in [0, 1]. Type 0 corresponds to a `sentence A` and type 1 corresponds to
            a `sentence B` token (see BERT paper for more details).
        `attention_mask`: an optional torch.LongTensor of shape [batch_size, sequence_length] with indices
            selected in [0, 1]. It's a mask to be used if the input sequence length is smaller than the max
            input sequence length in the current batch. It's the mask that we typically use for attention when
            a batch has varying length sentences.
        `masked_lm_labels`: optional masked language modeling labels: torch.LongTensor of shape [batch_size, sequence_length]
            with indices selected in [-1, 0, ..., vocab_size]. All labels set to -1 are ignored (masked), the loss
            is only computed for the labels set in [0, ..., vocab_size]
        `next_sentence_label`: optional next sentence classification loss: torch.LongTensor of shape [batch_size]
            with indices selected in [0, 1].
            0 => next sentence is the continuation, 1 => next sentence is a random sentence.

    Outputs:
        if `masked_lm_labels` and `next_sentence_label` are not `None`:
            Outputs the total_loss which is the sum of the masked language modeling loss and the next
            sentence classification loss.
        if `masked_lm_labels` or `next_sentence_label` is `None`:
            Outputs a tuple comprising
            - the masked language modeling logits of shape [batch_size, sequence_length, vocab_size], and
            - the next sentence classification logits of shape [batch_size, 2].

    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, 1, 0]])

    config = BertConfig(vocab_size_or_config_json_file=32000, hidden_size=768,
        num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072)

    model = BertForPreTraining(config)
    masked_lm_logits_scores, seq_relationship_logits = model(input_ids, token_type_ids, input_mask)
    ```
    """
    def __init__(self, config):
        super(BertForPreTraining, self).__init__(config)
        self.bert = BertModel(config)
        self.cls = BertPreTrainingHeads(config, self.bert.embeddings.word_embeddings.weight)
        self.loss_fn = torch.nn.CrossEntropyLoss(ignore_index=-1)
        self.apply(self.init_bert_weights)

    def forward(self, input_ids, token_type_ids, attention_mask, masked_lm_labels=None, next_sentence_labels=None):
        encoded_layers, pooled_output = self.bert(input_ids, token_type_ids, attention_mask)
        sequence_output = encoded_layers[-1]
        prediction_scores, seq_relationship_score = self.cls(sequence_output, pooled_output)
        return prediction_scores, seq_relationship_score


class BertForMaskedLM(BertPreTrainedModel):
    """BERT model with the masked language modeling head.
    This module comprises the BERT model followed by the masked language modeling head.

    Params:
        config: a BertConfig class instance with the configuration to build a new model.

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, sequence_length]
            with the word token indices in the vocabulary(see the tokens preprocessing logic in the scripts
            `extract_features.py`, `run_classifier.py` and `run_squad.py`)
        `token_type_ids`: an optional torch.LongTensor of shape [batch_size, sequence_length] with the token
            types indices selected in [0, 1]. Type 0 corresponds to a `sentence A` and type 1 corresponds to
            a `sentence B` token (see BERT paper for more details).
        `attention_mask`: an optional torch.LongTensor of shape [batch_size, sequence_length] with indices
            selected in [0, 1]. It's a mask to be used if the input sequence length is smaller than the max
            input sequence length in the current batch. It's the mask that we typically use for attention when
            a batch has varying length sentences.
        `masked_lm_labels`: masked language modeling labels: torch.LongTensor of shape [batch_size, sequence_length]
            with indices selected in [-1, 0, ..., vocab_size]. All labels set to -1 are ignored (masked), the loss
            is only computed for the labels set in [0, ..., vocab_size]

    Outputs:
        if `masked_lm_labels` is  not `None`:
            Outputs the masked language modeling loss.
        if `masked_lm_labels` is `None`:
            Outputs the masked language modeling logits of shape [batch_size, sequence_length, vocab_size].

    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, 1, 0]])

    config = BertConfig(vocab_size_or_config_json_file=32000, hidden_size=768,
        num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072)

    model = BertForMaskedLM(config)
    masked_lm_logits_scores = model(input_ids, token_type_ids, input_mask)
    ```
    """
    def __init__(self, config):
        super(BertForMaskedLM, self).__init__(config)
        self.bert = BertModel(config)
        self.cls = BertOnlyMLMHead(config, self.bert.embeddings.word_embeddings.weight)
        self.apply(self.init_bert_weights)

    def forward(self, input_ids, token_type_ids=None, attention_mask=None, masked_lm_labels=None):
        encoded_layers, _ = self.bert(input_ids, token_type_ids, attention_mask)
        sequence_output = encoded_layers[-1]
        prediction_scores = self.cls(sequence_output)

        if masked_lm_labels is not None:
            loss_fct = CrossEntropyLoss(ignore_index=-1)
            masked_lm_loss = loss_fct(prediction_scores.view(-1, self.config.vocab_size), masked_lm_labels.view(-1))
            return masked_lm_loss
        else:
            return prediction_scores


class BertForNextSentencePrediction(BertPreTrainedModel):
    """BERT model with next sentence prediction head.
    This module comprises the BERT model followed by the next sentence classification head.

    Params:
        config: a BertConfig class instance with the configuration to build a new model.

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, sequence_length]
            with the word token indices in the vocabulary(see the tokens preprocessing logic in the scripts
            `extract_features.py`, `run_classifier.py` and `run_squad.py`)
        `token_type_ids`: an optional torch.LongTensor of shape [batch_size, sequence_length] with the token
            types indices selected in [0, 1]. Type 0 corresponds to a `sentence A` and type 1 corresponds to
            a `sentence B` token (see BERT paper for more details).
        `attention_mask`: an optional torch.LongTensor of shape [batch_size, sequence_length] with indices
            selected in [0, 1]. It's a mask to be used if the input sequence length is smaller than the max
            input sequence length in the current batch. It's the mask that we typically use for attention when
            a batch has varying length sentences.
        `next_sentence_label`: next sentence classification loss: torch.LongTensor of shape [batch_size]
            with indices selected in [0, 1].
            0 => next sentence is the continuation, 1 => next sentence is a random sentence.

    Outputs:
        if `next_sentence_label` is not `None`:
            Outputs the total_loss which is the sum of the masked language modeling loss and the next
            sentence classification loss.
        if `next_sentence_label` is `None`:
            Outputs the next sentence classification logits of shape [batch_size, 2].

    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, 1, 0]])

    config = BertConfig(vocab_size_or_config_json_file=32000, hidden_size=768,
        num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072)

    model = BertForNextSentencePrediction(config)
    seq_relationship_logits = model(input_ids, token_type_ids, input_mask)
    ```
    """
    def __init__(self, config):
        super(BertForNextSentencePrediction, self).__init__(config)
        self.bert = BertModel(config)
        self.cls = BertOnlyNSPHead(config)
        self.apply(self.init_bert_weights)

    def forward(self, input_ids, token_type_ids=None, attention_mask=None, next_sentence_label=None):
        _, pooled_output = self.bert(input_ids, token_type_ids, attention_mask)
        seq_relationship_score = self.cls( pooled_output)

        if next_sentence_label is not None:
            loss_fct = CrossEntropyLoss(ignore_index=-1)
            next_sentence_loss = loss_fct(seq_relationship_score.view(-1, 2), next_sentence_label.view(-1))
            return next_sentence_loss
        else:
            return seq_relationship_score


class BertForSequenceClassification(BertPreTrainedModel):
    """BERT model for classification.
    This module is composed of the BERT model with a linear layer on top of
    the pooled output.

    Params:
        `config`: a BertConfig class instance with the configuration to build a new model.
        `num_labels`: the number of classes for the classifier. Default = 2.

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, sequence_length]
            with the word token indices in the vocabulary(see the tokens preprocessing logic in the scripts
            `extract_features.py`, `run_classifier.py` and `run_squad.py`)
        `token_type_ids`: an optional torch.LongTensor of shape [batch_size, sequence_length] with the token
            types indices selected in [0, 1]. Type 0 corresponds to a `sentence A` and type 1 corresponds to
            a `sentence B` token (see BERT paper for more details).
        `attention_mask`: an optional torch.LongTensor of shape [batch_size, sequence_length] with indices
            selected in [0, 1]. It's a mask to be used if the input sequence length is smaller than the max
            input sequence length in the current batch. It's the mask that we typically use for attention when
            a batch has varying length sentences.
        `labels`: labels for the classification output: torch.LongTensor of shape [batch_size]
            with indices selected in [0, ..., num_labels].

    Outputs:
        if `labels` is not `None`:
            Outputs the CrossEntropy classification loss of the output with the labels.
        if `labels` is `None`:
            Outputs the classification logits of shape [batch_size, num_labels].

    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, 1, 0]])

    config = BertConfig(vocab_size_or_config_json_file=32000, hidden_size=768,
        num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072)

    num_labels = 2

    model = BertForSequenceClassification(config, num_labels)
    logits = model(input_ids, token_type_ids, input_mask)
    ```
    """
    def __init__(self, config, num_labels):
        super(BertForSequenceClassification, self).__init__(config)
        self.num_labels = num_labels
        self.bert = BertModel(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(config.hidden_size, num_labels)
        self.apply(self.init_bert_weights)

    def forward(self, input_ids, token_type_ids=None, attention_mask=None):
        _, pooled_output = self.bert(input_ids, token_type_ids, attention_mask)
        pooled_output = self.dropout(pooled_output)
        return self.classifier(pooled_output)


class BertForMultipleChoice(BertPreTrainedModel):
    """BERT model for multiple choice tasks.
    This module is composed of the BERT model with a linear layer on top of
    the pooled output.

    Params:
        `config`: a BertConfig class instance with the configuration to build a new model.
        `num_choices`: the number of classes for the classifier. Default = 2.

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, num_choices, sequence_length]
            with the word token indices in the vocabulary(see the tokens preprocessing logic in the scripts
            `extract_features.py`, `run_classifier.py` and `run_squad.py`)
        `token_type_ids`: an optional torch.LongTensor of shape [batch_size, num_choices, sequence_length]
            with the token types indices selected in [0, 1]. Type 0 corresponds to a `sentence A`
            and type 1 corresponds to a `sentence B` token (see BERT paper for more details).
        `attention_mask`: an optional torch.LongTensor of shape [batch_size, num_choices, sequence_length] with indices
            selected in [0, 1]. It's a mask to be used if the input sequence length is smaller than the max
            input sequence length in the current batch. It's the mask that we typically use for attention when
            a batch has varying length sentences.
        `labels`: labels for the classification output: torch.LongTensor of shape [batch_size]
            with indices selected in [0, ..., num_choices].

    Outputs:
        if `labels` is not `None`:
            Outputs the CrossEntropy classification loss of the output with the labels.
        if `labels` is `None`:
            Outputs the classification logits of shape [batch_size, num_labels].

    Example usage:
    ```python
    # Already been converted into WordPiece token ids
    input_ids = torch.LongTensor([[[31, 51, 99], [15, 5, 0]], [[12, 16, 42], [14, 28, 57]]])
    input_mask = torch.LongTensor([[[1, 1, 1], [1, 1, 0]],[[1,1,0], [1, 0, 0]]])
    token_type_ids = torch.LongTensor([[[0, 0, 1], [0, 1, 0]],[[0, 1, 1], [0, 0, 1]]])
    config = BertConfig(vocab_size_or_config_json_file=32000, hidden_size=768,
        num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072)

    num_choices = 2

    model = BertForMultipleChoice(config, num_choices)
    logits = model(input_ids, token_type_ids, input_mask)
    ```
    """
    def __init__(self, config, num_choices):
        super(BertForMultipleChoice, self).__init__(config)
        self.num_choices = num_choices
        self.bert = BertModel(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(config.hidden_size, 1)
        self.apply(self.init_bert_weights)

    def forward(self, input_ids, token_type_ids=None, attention_mask=None, labels=None):
        flat_input_ids = input_ids.view(-1, input_ids.size(-1))
        flat_token_type_ids = token_type_ids.view(-1, token_type_ids.size(-1))
        flat_attention_mask = attention_mask.view(-1, attention_mask.size(-1))
        _, pooled_output = self.bert(flat_input_ids, flat_token_type_ids, flat_attention_mask)
        pooled_output = self.dropout(pooled_output)
        logits = self.classifier(pooled_output)
        reshaped_logits = logits.view(-1, self.num_choices)

        if labels is not None:
            loss_fct = CrossEntropyLoss()
            loss = loss_fct(reshaped_logits, labels)
            return loss
        else:
            return reshaped_logits


class BertForTokenClassification(BertPreTrainedModel):
    """BERT model for token-level classification.
    This module is composed of the BERT model with a linear layer on top of
    the full hidden state of the last layer.

    Params:
        `config`: a BertConfig class instance with the configuration to build a new model.
        `num_labels`: the number of classes for the classifier. Default = 2.

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, sequence_length]
            with the word token indices in the vocabulary(see the tokens preprocessing logic in the scripts
            `extract_features.py`, `run_classifier.py` and `run_squad.py`)
        `token_type_ids`: an optional torch.LongTensor of shape [batch_size, sequence_length] with the token
            types indices selected in [0, 1]. Type 0 corresponds to a `sentence A` and type 1 corresponds to
            a `sentence B` token (see BERT paper for more details).
        `attention_mask`: an optional torch.LongTensor of shape [batch_size, sequence_length] with indices
            selected in [0, 1]. It's a mask to be used if the input sequence length is smaller than the max
            input sequence length in the current batch. It's the mask that we typically use for attention when
            a batch has varying length sentences.
        `labels`: labels for the classification output: torch.LongTensor of shape [batch_size, sequence_length]
            with indices selected in [0, ..., num_labels].

    Outputs:
        if `labels` is not `None`:
            Outputs the CrossEntropy classification loss of the output with the labels.
        if `labels` is `None`:
            Outputs the classification logits of shape [batch_size, sequence_length, num_labels].

    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, 1, 0]])

    config = BertConfig(vocab_size_or_config_json_file=32000, hidden_size=768,
        num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072)

    num_labels = 2

    model = BertForTokenClassification(config, num_labels)
    logits = model(input_ids, token_type_ids, input_mask)
    ```
    """
    def __init__(self, config, num_labels):
        super(BertForTokenClassification, self).__init__(config)
        self.num_labels = num_labels
        self.bert = BertModel(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(config.hidden_size, num_labels)
        self.apply(self.init_bert_weights)

    def forward(self, input_ids, token_type_ids=None, attention_mask=None, labels=None):
        encoded_layers, _ = self.bert(input_ids, token_type_ids, attention_mask)
        sequence_output = encoded_layers[-1]
        sequence_output = self.dropout(sequence_output)
        logits = self.classifier(sequence_output)

        if labels is not None:
            loss_fct = CrossEntropyLoss()
            # Only keep active parts of the loss
            if attention_mask is not None:
                active_loss = attention_mask.view(-1) == 1
                active_logits = logits.view(-1, self.num_labels)[active_loss]
                active_labels = labels.view(-1)[active_loss]
                loss = loss_fct(active_logits, active_labels)
            else:
                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
            return loss
        else:
            return logits


class BertForQuestionAnswering(BertPreTrainedModel):
    """BERT model for Question Answering (span extraction).
    This module is composed of the BERT model with a linear layer on top of
    the sequence output that computes start_logits and end_logits

    Params:
        `config`: a BertConfig class instance with the configuration to build a new model.

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, sequence_length]
            with the word token indices in the vocabulary(see the tokens preprocessing logic in the scripts
            `extract_features.py`, `run_classifier.py` and `run_squad.py`)
        `token_type_ids`: an optional torch.LongTensor of shape [batch_size, sequence_length] with the token
            types indices selected in [0, 1]. Type 0 corresponds to a `sentence A` and type 1 corresponds to
            a `sentence B` token (see BERT paper for more details).
        `attention_mask`: an optional torch.LongTensor of shape [batch_size, sequence_length] with indices
            selected in [0, 1]. It's a mask to be used if the input sequence length is smaller than the max
            input sequence length in the current batch. It's the mask that we typically use for attention when
            a batch has varying length sentences.

    Outputs:
         Outputs a tuple of start_logits, end_logits which are the logits respectively for the start and end
         position tokens of shape [batch_size, sequence_length].

    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, 1, 0]])

    config = BertConfig(vocab_size_or_config_json_file=32000, hidden_size=768,
        num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072)

    model = BertForQuestionAnswering(config)
    start_logits, end_logits = model(input_ids, token_type_ids, input_mask)
    ```
    """
    def __init__(self, config):
        super(BertForQuestionAnswering, 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.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.qa_outputs = nn.Linear(config.hidden_size, 2)
        self.apply(self.init_bert_weights)

    def forward(self, input_ids, token_type_ids, attention_mask):
        encoded_layers, _ = self.bert(input_ids, token_type_ids, attention_mask)
        sequence_output = encoded_layers[-1]
        logits = self.qa_outputs(sequence_output)
        start_logits, end_logits = logits.split(1, dim=-1)
        start_logits = start_logits.squeeze(-1)
        end_logits = end_logits.squeeze(-1)
        return start_logits, end_logits


# =============================================================================================================
# Gated BERT
# =============================================================================================================
class GateForwardFunction(torch.autograd.Function):
    @staticmethod
    def forward(ctx, x):
        ctx.mark_dirty(x)
        output = x
        output.mul_(0.)
        return output

    @staticmethod
    def backward(ctx, grad_output):
        grad_input = grad_output
        return grad_input

    @staticmethod
    def jvp(ctx, *grad_inputs):
        raise NotImplementedError("Not supported for GateForwardFunction")


class GateBackwardFunction(torch.autograd.Function):
    @staticmethod
    def forward(ctx, x):
        return x

    @staticmethod
    def backward(ctx, grad_output):
        grad_input = torch.zeros_like(grad_output)
        return grad_input

    @staticmethod
    def jvp(ctx, *grad_inputs):
        raise NotImplementedError("Not supported for GateBackwardFunction")


GateForward = GateForwardFunction.apply
GateBackward = GateBackwardFunction.apply


class BertGatedSelfOutput(nn.Module):
    def __init__(self, config):
        super(BertGatedSelfOutput, self).__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.LayerNorm = BertLayerNorm(config.hidden_size, eps=1e-12)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)

    def forward(self, hidden_states, output_layer_norm_input):
        hidden_states = self.dense(hidden_states)
        hidden_states = self.dropout(hidden_states)
        hidden_states = GateForward(hidden_states)
        attention_layer_norm_input = hidden_states + output_layer_norm_input
        hidden_states = GateBackward(attention_layer_norm_input)
        hidden_states = self.LayerNorm(hidden_states)
        return hidden_states, attention_layer_norm_input


class BertGatedAttention(nn.Module):
    def __init__(self, config):
        super(BertGatedAttention, self).__init__()
        self.self = BertSelfAttention(config)
        self.output = BertGatedSelfOutput(config)

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


class BertGatedOutput(nn.Module):
    def __init__(self, config):
        super(BertGatedOutput, self).__init__()
        self.dense = nn.Linear(config.intermediate_size, config.hidden_size)
        self.LayerNorm = BertLayerNorm(config.hidden_size, eps=1e-12)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.gate_grads_from_next_layer = False

    def forward(self, hidden_states, attention_layer_norm_input, input_tensor):
        hidden_states = self.dense(hidden_states)
        hidden_states = self.dropout(hidden_states)
        hidden_states = GateForward(hidden_states)
        output_layer_norm_input = hidden_states + attention_layer_norm_input
        hidden_states = output_layer_norm_input
        if self.gate_grads_from_next_layer:
            hidden_states = GateBackward(hidden_states)
        hidden_states = self.LayerNorm(hidden_states)
        return hidden_states, output_layer_norm_input


class BertGatedLayer(nn.Module):
    def __init__(self, config):
        super(BertGatedLayer, self).__init__()
        if not config.layer_norm_large_model:
            raise NotImplementedError("BertGatedLayer is only supported for PreLN BERT")
        self.attention = BertGatedAttention(config)
        self.intermediate = BertIntermediate(config)
        self.output = BertGatedOutput(config)

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