modeling_transfo_xl.py

# coding=utf-8
# Copyright 2018 Google AI, Google Brain and Carnegie Mellon University Authors and the HugginFace Inc. team.
# Copyright (c) 2018, NVIDIA CORPORATION.  All rights reserved.
#
# 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 Transformer XL model.
    Directly adapted from https://github.com/kimiyoung/transformer-xl.
    In particular https://github.com/kimiyoung/transformer-xl/blob/master/pytorch/mem_transformer.py
"""

import os
import copy
import json
import math
import logging
import tarfile
import tempfile
import shutil
import collections

import torch
import torch.nn as nn
from torch.nn import CrossEntropyLoss
from torch.nn.parameter import Parameter

from .modeling import BertLayerNorm as LayerNorm
from .file_utils import cached_path

logger = logging.getLogger(__name__)

PRETRAINED_MODEL_ARCHIVE_MAP = {
    'transfo-xl': "https://s3.amazonaws.com/models.huggingface.co/bert/transfo-xl.tar.gz",
}
CONFIG_NAME = 'transfo_xl_config.json'
WEIGHTS_NAME = 'pytorch_model.bin'

class TransfoXLConfig(object):
    """Configuration class to store the configuration of a `TransfoXLModel`.
    """
    def __init__(self,
                 vocab_size_or_config_json_file=267735,
                 cutoffs=[20000, 40000, 200000],
                 d_model=410,
                 d_embed=410,
                 d_head=41,
                 d_inner=2100,
                 div_val=1.0,
                 pre_lnorm=False,
                 n_layer=16,
                 n_head=10,
                 tgt_len=150,
                 ext_len=0,
                 mem_len=150,
                 same_length=False,
                 attn_type=0,
                 clamp_len=-1,
                 sample_softmax=-1,
                 adaptive=True,
                 tied=True,
                 dropout=0.1,
                 dropatt=0.0,
                 init="normal",
                 init_range=0.01,
                 proj_init_std=0.01,
                 init_std=0.02):
        """Constructs TransfoXLConfig.

        Args:
            vocab_size_or_config_json_file: Vocabulary size of `inputs_ids` in `TransfoXLModel` or a configuration json file.
            cutoffs: cutoffs for the adaptive softmax
            d_model: Dimensionality of the model's hidden states.
            d_embed: Dimensionality of the embeddings
            d_head: Dimensionality of the model's heads.
            div_val: divident value for adapative input and softmax
            pre_lnorm: apply LayerNorm to the input instead of the output
            d_inner: Inner dimension in FF
            n_layer: Number of hidden layers in the Transformer encoder.
            n_head: Number of attention heads for each attention layer in
                the Transformer encoder.
            tgt_len: number of tokens to predict
            ext_len: length of the extended context
            mem_len: length of the retained previous heads
            same_length: use the same attn length for all tokens
            attn_type: attention type. 0 for Transformer-XL, 1 for Shaw et al, 2 for Vaswani et al, 3 for Al Rfou et al.
            clamp_len: use the same pos embeddings after clamp_len
            sample_softmax: number of samples in sampled softmax
            adaptive: use adaptive softmax
            tied: tie the word embedding and softmax weights
            dropout: The dropout probabilitiy for all fully connected
                layers in the embeddings, encoder, and pooler.
            dropatt: The dropout ratio for the attention probabilities.
            embd_pdrop: The dropout ratio for the embeddings.
            init: parameter initializer to use
            init_range: parameters initialized by U(-init_range, init_range).
            proj_init_std: parameters initialized by N(0, init_std)
            init_std: parameters initialized by N(0, init_std)
        """
        if isinstance(vocab_size_or_config_json_file, str):
            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.cutoffs = []
            self.cutoffs.extend(cutoffs)
            self.tie_projs = [False] + [True] * len(self.cutoffs)
            self.d_model = d_model
            self.d_embed = d_embed
            self.d_head = d_head
            self.d_inner = d_inner
            self.div_val = div_val
            self.pre_lnorm = pre_lnorm
            self.n_layer = n_layer
            self.n_head = n_head
            self.tgt_len = tgt_len
            self.ext_len = ext_len
            self.mem_len = mem_len
            self.same_length = same_length
            self.attn_type = attn_type
            self.clamp_len = clamp_len
            self.sample_softmax = sample_softmax
            self.adaptive = adaptive
            self.tied = tied
            self.dropout = dropout
            self.dropatt = dropatt
            self.init = init
            self.init_range = init_range
            self.proj_init_std = proj_init_std
            self.init_std = init_std
        else:
            raise ValueError("First argument must be either a vocabulary size (int)"
                             "or the path to a pretrained model config file (str)")

    @property
    def total_num_embeddings(self):
        return self.vocab_size + self.n_special + self.n_ctx

    @classmethod
    def from_dict(cls, json_object):
        """Constructs a `TransfoXLConfig` from a Python dictionary of parameters."""
        config = TransfoXLConfig(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 `TransfoXLConfig` 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 PositionalEmbedding(nn.Module):
    def __init__(self, demb):
        super(PositionalEmbedding, self).__init__()

        self.demb = demb

        inv_freq = 1 / (10000 ** (torch.arange(0.0, demb, 2.0) / demb))
        self.register_buffer('inv_freq', inv_freq)

    def forward(self, pos_seq, bsz=None):
        sinusoid_inp = torch.ger(pos_seq, self.inv_freq)
        pos_emb = torch.cat([sinusoid_inp.sin(), sinusoid_inp.cos()], dim=-1)

        if bsz is not None:
            return pos_emb[:,None,:].expand(-1, bsz, -1)
        else:
            return pos_emb[:,None,:]


class PositionwiseFF(nn.Module):
    def __init__(self, d_model, d_inner, dropout, pre_lnorm=False):
        super(PositionwiseFF, self).__init__()

        self.d_model = d_model
        self.d_inner = d_inner
        self.dropout = dropout

        self.CoreNet = nn.Sequential(
            nn.Linear(d_model, d_inner), nn.ReLU(inplace=True),
            nn.Dropout(dropout),
            nn.Linear(d_inner, d_model),
            nn.Dropout(dropout),
        )

        self.layer_norm = nn.LayerNorm(d_model)

        self.pre_lnorm = pre_lnorm

    def forward(self, inp):
        if self.pre_lnorm:
            ##### layer normalization + positionwise feed-forward
            core_out = self.CoreNet(self.layer_norm(inp))

            ##### residual connection
            output = core_out + inp
        else:
            ##### positionwise feed-forward
            core_out = self.CoreNet(inp)

            ##### residual connection + layer normalization
            output = self.layer_norm(inp + core_out)

        return output

class MultiHeadAttn(nn.Module):
    def __init__(self, n_head, d_model, d_head, dropout, dropatt=0, 
                 pre_lnorm=False):
        super(MultiHeadAttn, self).__init__()

        self.n_head = n_head
        self.d_model = d_model
        self.d_head = d_head
        self.dropout = dropout

        self.q_net = nn.Linear(d_model, n_head * d_head, bias=False)
        self.kv_net = nn.Linear(d_model, 2 * n_head * d_head, bias=False)

        self.drop = nn.Dropout(dropout)
        self.dropatt = nn.Dropout(dropatt)
        self.o_net = nn.Linear(n_head * d_head, d_model, bias=False)

        self.layer_norm = nn.LayerNorm(d_model)

        self.scale = 1 / (d_head ** 0.5)

        self.pre_lnorm = pre_lnorm

    def forward(self, h, attn_mask=None, mems=None):
        ##### multihead attention
        # [hlen x bsz x n_head x d_head]

        if mems is not None:
            c = torch.cat([mems, h], 0)
        else:
            c = h

        if self.pre_lnorm:
            ##### layer normalization
            c = self.layer_norm(c)

        head_q = self.q_net(h)
        head_k, head_v = torch.chunk(self.kv_net(c), 2, -1)

        head_q = head_q.view(h.size(0), h.size(1), self.n_head, self.d_head)
        head_k = head_k.view(c.size(0), c.size(1), self.n_head, self.d_head)
        head_v = head_v.view(c.size(0), c.size(1), self.n_head, self.d_head)

        # [qlen x klen x bsz x n_head]
        attn_score = torch.einsum('ibnd,jbnd->ijbn', (head_q, head_k))
        attn_score.mul_(self.scale)
        if attn_mask is not None and attn_mask.any().item():
            if attn_mask.dim() == 2:
                attn_score.masked_fill_(attn_mask[None,:,:,None], -float('inf'))
            elif attn_mask.dim() == 3:
                attn_score.masked_fill_(attn_mask[:,:,:,None], -float('inf'))

        # [qlen x klen x bsz x n_head]
        attn_prob = F.softmax(attn_score, dim=1)
        attn_prob = self.dropatt(attn_prob)

        # [qlen x klen x bsz x n_head] + [klen x bsz x n_head x d_head] -> [qlen x bsz x n_head x d_head]
        attn_vec = torch.einsum('ijbn,jbnd->ibnd', (attn_prob, head_v))
        attn_vec = attn_vec.contiguous().view(
            attn_vec.size(0), attn_vec.size(1), self.n_head * self.d_head)

        ##### linear projection
        attn_out = self.o_net(attn_vec)
        attn_out = self.drop(attn_out)

        if self.pre_lnorm:
            ##### residual connection
            output = h + attn_out
        else:
            ##### residual connection + layer normalization
            output = self.layer_norm(h + attn_out)

        return output

class RelMultiHeadAttn(nn.Module):
    def __init__(self, n_head, d_model, d_head, dropout, dropatt=0,
                 tgt_len=None, ext_len=None, mem_len=None, pre_lnorm=False):
        super(RelMultiHeadAttn, self).__init__()

        self.n_head = n_head
        self.d_model = d_model
        self.d_head = d_head
        self.dropout = dropout

        self.qkv_net = nn.Linear(d_model, 3 * n_head * d_head, bias=False)

        self.drop = nn.Dropout(dropout)
        self.dropatt = nn.Dropout(dropatt)
        self.o_net = nn.Linear(n_head * d_head, d_model, bias=False)

        self.layer_norm = nn.LayerNorm(d_model)

        self.scale = 1 / (d_head ** 0.5)

        self.pre_lnorm = pre_lnorm

    def _parallelogram_mask(self, h, w, left=False):
        mask = torch.ones((h, w)).byte()
        m = min(h, w)
        mask[:m,:m] = torch.triu(mask[:m,:m])
        mask[-m:,-m:] = torch.tril(mask[-m:,-m:])

        if left:
            return mask
        else:
            return mask.flip(0)

    def _shift(self, x, qlen, klen, mask, left=False):
        if qlen > 1:
            zero_pad = torch.zeros((x.size(0), qlen-1, x.size(2), x.size(3)),
                                    device=x.device, dtype=x.dtype)
        else:
            zero_pad = torch.zeros(0, device=x.device, dtype=x.dtype)

        if left:
            mask = mask.flip(1)
            x_padded = torch.cat([zero_pad, x], dim=1).expand(qlen, -1, -1, -1)
        else:
            x_padded = torch.cat([x, zero_pad], dim=1).expand(qlen, -1, -1, -1)

        x = x_padded.masked_select(mask[:,:,None,None]) \
                    .view(qlen, klen, x.size(2), x.size(3))

        return x

    def _rel_shift(self, x, zero_triu=False):
        zero_pad = torch.zeros((x.size(0), 1, *x.size()[2:]),
                               device=x.device, dtype=x.dtype)
        x_padded = torch.cat([zero_pad, x], dim=1)

        x_padded = x_padded.view(x.size(1) + 1, x.size(0), *x.size()[2:])

        x = x_padded[1:].view_as(x)

        if zero_triu:
            ones = torch.ones((x.size(0), x.size(1)))
            x = x * torch.tril(ones, x.size(1) - x.size(0))[:,:,None,None]

        return x

    def forward(self, w, r, attn_mask=None, mems=None):
        raise NotImplementedError

class RelPartialLearnableMultiHeadAttn(RelMultiHeadAttn):
    def __init__(self, *args, **kwargs):
        super(RelPartialLearnableMultiHeadAttn, self).__init__(*args, **kwargs)

        self.r_net = nn.Linear(self.d_model, self.n_head * self.d_head, bias=False)

    def forward(self, w, r, r_w_bias, r_r_bias, attn_mask=None, mems=None):
        qlen, rlen, bsz = w.size(0), r.size(0), w.size(1)

        if mems is not None:
            cat = torch.cat([mems, w], 0)
            if self.pre_lnorm:
                w_heads = self.qkv_net(self.layer_norm(cat))
            else:
                w_heads = self.qkv_net(cat)
            r_head_k = self.r_net(r)

            w_head_q, w_head_k, w_head_v = torch.chunk(w_heads, 3, dim=-1)
            w_head_q = w_head_q[-qlen:]
        else:
            if self.pre_lnorm:
                w_heads = self.qkv_net(self.layer_norm(w))
            else:
                w_heads = self.qkv_net(w)
            r_head_k = self.r_net(r)

            w_head_q, w_head_k, w_head_v = torch.chunk(w_heads, 3, dim=-1)

        klen = w_head_k.size(0)

        w_head_q = w_head_q.view(qlen, bsz, self.n_head, self.d_head)           # qlen x bsz x n_head x d_head
        w_head_k = w_head_k.view(klen, bsz, self.n_head, self.d_head)           # qlen x bsz x n_head x d_head
        w_head_v = w_head_v.view(klen, bsz, self.n_head, self.d_head)           # qlen x bsz x n_head x d_head

        r_head_k = r_head_k.view(rlen, self.n_head, self.d_head)                # qlen x n_head x d_head

        #### compute attention score
        rw_head_q = w_head_q + r_w_bias                                         # qlen x bsz x n_head x d_head
        AC = torch.einsum('ibnd,jbnd->ijbn', (rw_head_q, w_head_k))             # qlen x klen x bsz x n_head

        rr_head_q = w_head_q + r_r_bias
        BD = torch.einsum('ibnd,jnd->ijbn', (rr_head_q, r_head_k))              # qlen x klen x bsz x n_head
        BD = self._rel_shift(BD)

        # [qlen x klen x bsz x n_head]
        attn_score = AC + BD
        attn_score.mul_(self.scale)

        #### compute attention probability
        if attn_mask is not None and attn_mask.any().item():
            if attn_mask.dim() == 2:
                attn_score = attn_score.float().masked_fill(
                    attn_mask[None,:,:,None], -float('inf')).type_as(attn_score)
            elif attn_mask.dim() == 3:
                attn_score = attn_score.float().masked_fill(
                    attn_mask[:,:,:,None], -float('inf')).type_as(attn_score)

        # [qlen x klen x bsz x n_head]
        attn_prob = F.softmax(attn_score, dim=1)
        attn_prob = self.dropatt(attn_prob)

        #### compute attention vector
        attn_vec = torch.einsum('ijbn,jbnd->ibnd', (attn_prob, w_head_v))

        # [qlen x bsz x n_head x d_head]
        attn_vec = attn_vec.contiguous().view(
            attn_vec.size(0), attn_vec.size(1), self.n_head * self.d_head)

        ##### linear projection
        attn_out = self.o_net(attn_vec)
        attn_out = self.drop(attn_out)

        if self.pre_lnorm:
            ##### residual connection
            output = w + attn_out
        else:
            ##### residual connection + layer normalization
            output = self.layer_norm(w + attn_out)

        return output

class RelLearnableMultiHeadAttn(RelMultiHeadAttn):
    def __init__(self, *args, **kwargs):
        super(RelLearnableMultiHeadAttn, self).__init__(*args, **kwargs)

    def forward(self, w, r_emb, r_w_bias, r_bias, attn_mask=None, mems=None):
        # r_emb: [klen, n_head, d_head], used for term B
        # r_w_bias: [n_head, d_head], used for term C
        # r_bias: [klen, n_head], used for term D

        qlen, bsz = w.size(0), w.size(1)

        if mems is not None:
            cat = torch.cat([mems, w], 0)
            if self.pre_lnorm:
                w_heads = self.qkv_net(self.layer_norm(cat))
            else:
                w_heads = self.qkv_net(cat)
            w_head_q, w_head_k, w_head_v = torch.chunk(w_heads, 3, dim=-1)

            w_head_q = w_head_q[-qlen:]
        else:
            if self.pre_lnorm:
                w_heads = self.qkv_net(self.layer_norm(w))
            else:
                w_heads = self.qkv_net(w)
            w_head_q, w_head_k, w_head_v = torch.chunk(w_heads, 3, dim=-1)

        klen = w_head_k.size(0)

        w_head_q = w_head_q.view(qlen, bsz, self.n_head, self.d_head)
        w_head_k = w_head_k.view(klen, bsz, self.n_head, self.d_head)
        w_head_v = w_head_v.view(klen, bsz, self.n_head, self.d_head)

        if klen > r_emb.size(0):
            r_emb_pad = r_emb[0:1].expand(klen-r_emb.size(0), -1, -1)
            r_emb = torch.cat([r_emb_pad, r_emb], 0)
            r_bias_pad = r_bias[0:1].expand(klen-r_bias.size(0), -1)
            r_bias = torch.cat([r_bias_pad, r_bias], 0)
        else:
            r_emb = r_emb[-klen:]
            r_bias = r_bias[-klen:]

        #### compute attention score
        rw_head_q = w_head_q + r_w_bias[None]                                   # qlen x bsz x n_head x d_head

        AC = torch.einsum('ibnd,jbnd->ijbn', (rw_head_q, w_head_k))             # qlen x klen x bsz x n_head
        B_ = torch.einsum('ibnd,jnd->ijbn', (w_head_q, r_emb))                  # qlen x klen x bsz x n_head
        D_ = r_bias[None, :, None]                                              # 1    x klen x 1   x n_head
        BD = self._rel_shift(B_ + D_)

        # [qlen x klen x bsz x n_head]
        attn_score = AC + BD
        attn_score.mul_(self.scale)

        #### compute attention probability
        if attn_mask is not None and attn_mask.any().item():
            if attn_mask.dim() == 2:
                attn_score.masked_fill_(attn_mask[None,:,:,None], -float('inf'))
            elif attn_mask.dim() == 3:
                attn_score.masked_fill_(attn_mask[:,:,:,None], -float('inf'))

        # [qlen x klen x bsz x n_head]
        attn_prob = F.softmax(attn_score, dim=1)
        attn_prob = self.dropatt(attn_prob)

        #### compute attention vector
        attn_vec = torch.einsum('ijbn,jbnd->ibnd', (attn_prob, w_head_v))

        # [qlen x bsz x n_head x d_head]
        attn_vec = attn_vec.contiguous().view(
            attn_vec.size(0), attn_vec.size(1), self.n_head * self.d_head)

        ##### linear projection
        attn_out = self.o_net(attn_vec)
        attn_out = self.drop(attn_out)

        if self.pre_lnorm:
            ##### residual connection
            output = w + attn_out
        else:
            ##### residual connection + layer normalization
            output = self.layer_norm(w + attn_out)

        return output

class DecoderLayer(nn.Module):
    def __init__(self, n_head, d_model, d_head, d_inner, dropout, **kwargs):
        super(DecoderLayer, self).__init__()

        self.dec_attn = MultiHeadAttn(n_head, d_model, d_head, dropout, **kwargs)
        self.pos_ff = PositionwiseFF(d_model, d_inner, dropout, 
                                     pre_lnorm=kwargs.get('pre_lnorm'))

    def forward(self, dec_inp, dec_attn_mask=None, mems=None):

        output = self.dec_attn(dec_inp, attn_mask=dec_attn_mask,
                               mems=mems)
        output = self.pos_ff(output)

        return output

class RelLearnableDecoderLayer(nn.Module):
    def __init__(self, n_head, d_model, d_head, d_inner, dropout,
                 **kwargs):
        super(RelLearnableDecoderLayer, self).__init__()

        self.dec_attn = RelLearnableMultiHeadAttn(n_head, d_model, d_head, dropout,
                                         **kwargs)
        self.pos_ff = PositionwiseFF(d_model, d_inner, dropout, 
                                     pre_lnorm=kwargs.get('pre_lnorm'))

    def forward(self, dec_inp, r_emb, r_w_bias, r_bias, dec_attn_mask=None, mems=None):

        output = self.dec_attn(dec_inp, r_emb, r_w_bias, r_bias,
                               attn_mask=dec_attn_mask,
                               mems=mems)
        output = self.pos_ff(output)

        return output

class RelPartialLearnableDecoderLayer(nn.Module):
    def __init__(self, n_head, d_model, d_head, d_inner, dropout,
                 **kwargs):
        super(RelPartialLearnableDecoderLayer, self).__init__()

        self.dec_attn = RelPartialLearnableMultiHeadAttn(n_head, d_model,
                            d_head, dropout, **kwargs)
        self.pos_ff = PositionwiseFF(d_model, d_inner, dropout, 
                                     pre_lnorm=kwargs.get('pre_lnorm'))

    def forward(self, dec_inp, r, r_w_bias, r_r_bias, dec_attn_mask=None, mems=None):

        output = self.dec_attn(dec_inp, r, r_w_bias, r_r_bias,
                               attn_mask=dec_attn_mask,
                               mems=mems)
        output = self.pos_ff(output)

        return output


class AdaptiveEmbedding(nn.Module):
    def __init__(self, n_token, d_embed, d_proj, cutoffs, div_val=1, 
                 sample_softmax=False):
        super(AdaptiveEmbedding, self).__init__()

        self.n_token = n_token
        self.d_embed = d_embed

        self.cutoffs = cutoffs + [n_token]
        self.div_val = div_val
        self.d_proj = d_proj

        self.emb_scale = d_proj ** 0.5

        self.cutoff_ends = [0] + self.cutoffs

        self.emb_layers = nn.ModuleList()
        self.emb_projs = nn.ParameterList()
        if div_val == 1:
            self.emb_layers.append(
                nn.Embedding(n_token, d_embed, sparse=sample_softmax>0)
            )
            if d_proj != d_embed:
                self.emb_projs.append(nn.Parameter(torch.Tensor(d_proj, d_embed)))
        else:
            for i in range(len(self.cutoffs)):
                l_idx, r_idx = self.cutoff_ends[i], self.cutoff_ends[i+1]
                d_emb_i = d_embed // (div_val ** i)
                self.emb_layers.append(nn.Embedding(r_idx-l_idx, d_emb_i))
                self.emb_projs.append(nn.Parameter(torch.Tensor(d_proj, d_emb_i)))

    def forward(self, inp):
        if self.div_val == 1:
            embed = self.emb_layers[0](inp)
            if self.d_proj != self.d_embed:
                embed  = F.linear(embed, self.emb_projs[0])
        else:
            param = next(self.parameters())
            inp_flat = inp.view(-1)
            emb_flat = torch.zeros([inp_flat.size(0), self.d_proj], 
                dtype=param.dtype, device=param.device)
            for i in range(len(self.cutoffs)):
                l_idx, r_idx = self.cutoff_ends[i], self.cutoff_ends[i + 1]

                mask_i = (inp_flat >= l_idx) & (inp_flat < r_idx)
                indices_i = mask_i.nonzero().squeeze()

                if indices_i.numel() == 0:
                    continue

                inp_i = inp_flat.index_select(0, indices_i) - l_idx
                emb_i = self.emb_layers[i](inp_i)
                emb_i = F.linear(emb_i, self.emb_projs[i])

                emb_flat.index_copy_(0, indices_i, emb_i)

            embed = emb_flat.view(*inp.size(), self.d_proj)

        embed.mul_(self.emb_scale)

        return embed

class MemTransformerLM(nn.Module):
    def __init__(self, n_token, n_layer, n_head, d_model, d_head, d_inner,
                 dropout, dropatt, tie_weight=True, d_embed=None, 
                 div_val=1, tie_projs=[False], pre_lnorm=False,
                 tgt_len=None, ext_len=None, mem_len=None, 
                 cutoffs=[], adapt_inp=False,
                 same_length=False, attn_type=0, clamp_len=-1, 
                 sample_softmax=-1):
        super(MemTransformerLM, self).__init__()
        self.n_token = n_token

        d_embed = d_model if d_embed is None else d_embed
        self.d_embed = d_embed
        self.d_model = d_model
        self.n_head = n_head
        self.d_head = d_head

        self.word_emb = AdaptiveEmbedding(n_token, d_embed, d_model, cutoffs, 
                                          div_val=div_val)

        self.drop = nn.Dropout(dropout)

        self.n_layer = n_layer

        self.tgt_len = tgt_len
        self.mem_len = mem_len
        self.ext_len = ext_len
        self.max_klen = tgt_len + ext_len + mem_len

        self.attn_type = attn_type

        self.layers = nn.ModuleList()
        if attn_type == 0: # the default attention
            for i in range(n_layer):
                self.layers.append(
                    RelPartialLearnableDecoderLayer(
                        n_head, d_model, d_head, d_inner, dropout,
                        tgt_len=tgt_len, ext_len=ext_len, mem_len=mem_len,
                        dropatt=dropatt, pre_lnorm=pre_lnorm)
                )
        elif attn_type == 1: # learnable embeddings
            for i in range(n_layer):
                self.layers.append(
                    RelLearnableDecoderLayer(
                        n_head, d_model, d_head, d_inner, dropout,
                        tgt_len=tgt_len, ext_len=ext_len, mem_len=mem_len,
                        dropatt=dropatt, pre_lnorm=pre_lnorm)
                )
        elif attn_type in [2, 3]: # absolute embeddings
            for i in range(n_layer):
                self.layers.append(
                    DecoderLayer(
                        n_head, d_model, d_head, d_inner, dropout,
                        dropatt=dropatt, pre_lnorm=pre_lnorm)
                )

        self.sample_softmax = sample_softmax
        # use sampled softmax
        if sample_softmax > 0:
            self.out_layer = nn.Linear(d_model, n_token)
            if tie_weight:
                self.out_layer.weight = self.word_emb.weight
            self.tie_weight = tie_weight
            self.sampler = LogUniformSampler(n_token, sample_softmax)

        # use adaptive softmax (including standard softmax)
        else:
            self.crit = ProjectedAdaptiveLogSoftmax(n_token, d_embed, d_model, 
                                                    cutoffs, div_val=div_val)

            if tie_weight:
                for i in range(len(self.crit.out_layers)):
                    self.crit.out_layers[i].weight = self.word_emb.emb_layers[i].weight

            if tie_projs:
                for i, tie_proj in enumerate(tie_projs):
                    if tie_proj and div_val == 1 and d_model != d_embed:
                        self.crit.out_projs[i] = self.word_emb.emb_projs[0]
                    elif tie_proj and div_val != 1:
                        self.crit.out_projs[i] = self.word_emb.emb_projs[i]

        self.same_length = same_length
        self.clamp_len = clamp_len

        self._create_params()

    def backward_compatible(self):
        self.sample_softmax = -1

    def _create_params(self):
        if self.attn_type == 0: # default attention
            self.pos_emb = PositionalEmbedding(self.d_model)
            self.r_w_bias = nn.Parameter(torch.Tensor(self.n_head, self.d_head))
            self.r_r_bias = nn.Parameter(torch.Tensor(self.n_head, self.d_head))
        elif self.attn_type == 1: # learnable
            self.r_emb = nn.Parameter(torch.Tensor(
                    self.n_layer, self.max_klen, self.n_head, self.d_head))
            self.r_w_bias = nn.Parameter(torch.Tensor(
                    self.n_layer, self.n_head, self.d_head))
            self.r_bias = nn.Parameter(torch.Tensor(
                    self.n_layer, self.max_klen, self.n_head))
        elif self.attn_type == 2: # absolute standard
            self.pos_emb = PositionalEmbedding(self.d_model)
        elif self.attn_type == 3: # absolute deeper SA
            self.r_emb = nn.Parameter(torch.Tensor(
                    self.n_layer, self.max_klen, self.n_head, self.d_head))

    def reset_length(self, tgt_len, ext_len, mem_len):
        self.tgt_len = tgt_len
        self.mem_len = mem_len
        self.ext_len = ext_len

    def init_mems(self):
        if self.mem_len > 0:
            mems = []
            param = next(self.parameters())
            for i in range(self.n_layer+1):
                empty = torch.empty(0, dtype=param.dtype, device=param.device)
                mems.append(empty)

            return mems
        else:
            return None

    def _update_mems(self, hids, mems, qlen, mlen):
        # does not deal with None
        if mems is None: return None

        # mems is not None
        assert len(hids) == len(mems), 'len(hids) != len(mems)'

        # There are `mlen + qlen` steps that can be cached into mems
        # For the next step, the last `ext_len` of the `qlen` tokens
        # will be used as the extended context. Hence, we only cache
        # the tokens from `mlen + qlen - self.ext_len - self.mem_len`
        # to `mlen + qlen - self.ext_len`.
        with torch.no_grad():
            new_mems = []
            end_idx = mlen + max(0, qlen - 0 - self.ext_len)
            beg_idx = max(0, end_idx - self.mem_len)
            for i in range(len(hids)):

                cat = torch.cat([mems[i], hids[i]], dim=0)
                new_mems.append(cat[beg_idx:end_idx].detach())

        return new_mems

    def _forward(self, dec_inp, mems=None):
        qlen, bsz = dec_inp.size()

        word_emb = self.word_emb(dec_inp)

        mlen = mems[0].size(0) if mems is not None else 0
        klen = mlen + qlen
        if self.same_length:
            all_ones = word_emb.new_ones(qlen, klen)
            mask_len = klen - self.mem_len
            if mask_len > 0:
                mask_shift_len = qlen - mask_len
            else:
                mask_shift_len = qlen
            dec_attn_mask = (torch.triu(all_ones, 1+mlen)
                    + torch.tril(all_ones, -mask_shift_len)).byte()[:, :, None] # -1
        else:
            dec_attn_mask = torch.triu(
                word_emb.new_ones(qlen, klen), diagonal=1+mlen).byte()[:,:,None]

        hids = []
        if self.attn_type == 0: # default
            pos_seq = torch.arange(klen-1, -1, -1.0, device=word_emb.device, 
                                   dtype=word_emb.dtype)
            if self.clamp_len > 0:
                pos_seq.clamp_(max=self.clamp_len)
            pos_emb = self.pos_emb(pos_seq)

            core_out = self.drop(word_emb)
            pos_emb = self.drop(pos_emb)

            hids.append(core_out)
            for i, layer in enumerate(self.layers):
                mems_i = None if mems is None else mems[i]
                core_out = layer(core_out, pos_emb, self.r_w_bias,
                        self.r_r_bias, dec_attn_mask=dec_attn_mask, mems=mems_i)
                hids.append(core_out)
        elif self.attn_type == 1: # learnable
            core_out = self.drop(word_emb)
            hids.append(core_out)
            for i, layer in enumerate(self.layers):
                if self.clamp_len > 0:
                    r_emb = self.r_emb[i][-self.clamp_len :]
                    r_bias = self.r_bias[i][-self.clamp_len :]
                else:
                    r_emb, r_bias = self.r_emb[i], self.r_bias[i]

                mems_i = None if mems is None else mems[i]
                core_out = layer(core_out, r_emb, self.r_w_bias[i],
                        r_bias, dec_attn_mask=dec_attn_mask, mems=mems_i)
                hids.append(core_out)
        elif self.attn_type == 2: # absolute
            pos_seq = torch.arange(klen - 1, -1, -1.0, device=word_emb.device,
                                   dtype=word_emb.dtype)
            if self.clamp_len > 0:
                pos_seq.clamp_(max=self.clamp_len)
            pos_emb = self.pos_emb(pos_seq)

            core_out = self.drop(word_emb + pos_emb[-qlen:])

            hids.append(core_out)
            for i, layer in enumerate(self.layers):
                mems_i = None if mems is None else mems[i]
                if mems_i is not None and i == 0:
                    mems_i += pos_emb[:mlen]
                core_out = layer(core_out, dec_attn_mask=dec_attn_mask,
                                 mems=mems_i)
                hids.append(core_out)
        elif self.attn_type == 3:
            core_out = self.drop(word_emb)

            hids.append(core_out)
            for i, layer in enumerate(self.layers):
                mems_i = None if mems is None else mems[i]
                if mems_i is not None and mlen > 0:
                    cur_emb = self.r_emb[i][:-qlen]
                    cur_size = cur_emb.size(0)
                    if cur_size < mlen:
                        cur_emb_pad = cur_emb[0:1].expand(mlen-cur_size, -1, -1)
                        cur_emb = torch.cat([cur_emb_pad, cur_emb], 0)
                    else:
                        cur_emb = cur_emb[-mlen:]
                    mems_i += cur_emb.view(mlen, 1, -1)
                core_out += self.r_emb[i][-qlen:].view(qlen, 1, -1)

                core_out = layer(core_out, dec_attn_mask=dec_attn_mask,
                                 mems=mems_i)
                hids.append(core_out)

        core_out = self.drop(core_out)

        new_mems = self._update_mems(hids, mems, mlen, qlen)

        return core_out, new_mems

    def forward(self, data, target, *mems):
        # nn.DataParallel does not allow size(0) tensors to be broadcasted.
        # So, have to initialize size(0) mems inside the model forward.
        # Moreover, have to return new_mems to allow nn.DataParallel to piece
        # them together.
        if not mems: mems = self.init_mems()

        tgt_len = target.size(0)
        hidden, new_mems = self._forward(data, mems=mems)

        pred_hid = hidden[-tgt_len:]
        if self.sample_softmax > 0 and self.training:
            assert self.tie_weight
            logit = sample_logits(self.word_emb,
                self.out_layer.bias, target, pred_hid, self.sampler)
            loss = -F.log_softmax(logit, -1)[:, :, 0]
        else:
            loss = self.crit(pred_hid.view(-1, pred_hid.size(-1)), target.view(-1))
            loss = loss.view(tgt_len, -1)

        if new_mems is None:
            return [loss]
        else:
            return [loss] + new_mems


class TransfoXLPreTrainedModel(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(TransfoXLPreTrainedModel, self).__init__()
        if not isinstance(config, TransfoXLConfig):
            raise ValueError(
                "Parameter config in `{}(config)` should be an instance of class `TransfoXLConfig`. "
                "To create a model from a pretrained model use "
                "`model = {}.from_pretrained(PRETRAINED_MODEL_NAME)`".format(
                    self.__class__.__name__, self.__class__.__name__
                ))
        self.config = config

    def init_weight(weight):
        if self.config.init == 'uniform':
            nn.init.uniform_(weight, -self.config.init_range, self.config.init_range)
        elif self.config.init == 'normal':
            nn.init.normal_(weight, 0.0, self.config.init_std)

    def init_bias(bias):
        nn.init.constant_(bias, 0.0)

    def init_weights(self, m):
        """ Initialize the weights.
        """
        classname = m.__class__.__name__
        if classname.find('Linear') != -1:
            if hasattr(m, 'weight') and m.weight is not None:
                self.init_weight(m.weight)
            if hasattr(m, 'bias') and m.bias is not None:
                self.init_bias(m.bias)
        elif classname.find('AdaptiveEmbedding') != -1:
            if hasattr(m, 'emb_projs'):
                for i in range(len(m.emb_projs)):
                    if m.emb_projs[i] is not None:
                        nn.init.normal_(m.emb_projs[i], 0.0, self.config.proj_init_std)
        elif classname.find('Embedding') != -1:
            if hasattr(m, 'weight'):
                self.init_weight(m.weight)
        elif classname.find('ProjectedAdaptiveLogSoftmax') != -1:
            if hasattr(m, 'cluster_weight') and m.cluster_weight is not None:
                self.init_weight(m.cluster_weight)
            if hasattr(m, 'cluster_bias') and m.cluster_bias is not None:
                self.init_bias(m.cluster_bias)
            if hasattr(m, 'out_projs'):
                for i in range(len(m.out_projs)):
                    if m.out_projs[i] is not None:
                        nn.init.normal_(m.out_projs[i], 0.0, self.config.proj_init_std)
        elif classname.find('LayerNorm') != -1:
            if hasattr(m, 'weight'):
                nn.init.normal_(m.weight, 1.0, self.config.init_std)
            if hasattr(m, 'bias') and m.bias is not None:
                self.init_bias(m.bias)
        elif classname.find('TransformerLM') != -1:
            if hasattr(m, 'r_emb'):
                self.init_weight(m.r_emb)
            if hasattr(m, 'r_w_bias'):
                self.init_weight(m.r_w_bias)
            if hasattr(m, 'r_r_bias'):
                self.init_weight(m.r_r_bias)
            if hasattr(m, 'r_bias'):
                self.init_bias(m.r_bias)

    def set_num_special_tokens(self, num_special_tokens):
        pass

    @classmethod
    def from_pretrained(cls, pretrained_model_name, num_special_tokens=0, state_dict=None, cache_dir=None,
                        *inputs, **kwargs):
        """
        Instantiate a TransfoXLPreTrainedModel 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: either:
                - a str with the name of a pre-trained model to load selected in the list of:
                    . `transfo-xl`
                - a path or url to a pretrained model archive containing:
                    . `transfo_xl_config.json` a configuration file for the model
                    . `pytorch_model.bin` a PyTorch dump of a TransfoXLModel instance
            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 pre-trained models
            *inputs, **kwargs: additional input for the specific Bert class
                (ex: num_labels for BertForSequenceClassification)
        """
        if pretrained_model_name in PRETRAINED_MODEL_ARCHIVE_MAP:
            archive_file = PRETRAINED_MODEL_ARCHIVE_MAP[pretrained_model_name]
        else:
            archive_file = pretrained_model_name
        # redirect to the cache, if necessary
        try:
            resolved_archive_file = cached_path(archive_file, cache_dir=cache_dir)
        except FileNotFoundError:
            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,
                    ', '.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):
            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:
                archive.extractall(tempdir)
            serialization_dir = tempdir
        # Load config
        config_file = os.path.join(serialization_dir, CONFIG_NAME)
        config = TransfoXLConfig.from_json_file(config_file)
        logger.info("Model config {}".format(config))
        # Instantiate model.
        model = cls(config, *inputs, **kwargs)
        if state_dict is None:
            weights_path = os.path.join(serialization_dir, WEIGHTS_NAME)
            state_dict = torch.load(weights_path)

        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 + '.')
        load(model.transformer if hasattr(model, 'transformer') else model, 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)))
        # Add additional embeddings for special tokens if needed
        if num_special_tokens != config.n_special:
            model.set_num_special_tokens(num_special_tokens)
        if tempdir:
            # Clean up temp dir
            shutil.rmtree(tempdir)
        return model






###################




class TransfoXLLMHead(nn.Module):
    """ Language Model Head for the transformer """

    def __init__(self, model_embeddings_weights, config):
        super(TransfoXLLMHead, self).__init__()
        self.n_embd = config.n_embd
        self.set_embeddings_weights(model_embeddings_weights)

    def set_embeddings_weights(self, model_embeddings_weights):
        embed_shape = model_embeddings_weights.shape
        self.decoder = nn.Linear(embed_shape[1], embed_shape[0], bias=False)
        self.decoder.weight = model_embeddings_weights # Tied weights

    def forward(self, hidden_state):
        # Truncated Language modeling logits (we remove the last token)
        # h_trunc = h[:, :-1].contiguous().view(-1, self.n_embd)
        lm_logits = self.decoder(hidden_state)
        return lm_logits


class TransfoXLMultipleChoiceHead(nn.Module):
    """ Classifier Head for the transformer """

    def __init__(self, config):
        super(TransfoXLMultipleChoiceHead, self).__init__()
        self.n_embd = config.n_embd
        # self.multiple_choice_token = multiple_choice_token
        self.dropout = nn.Dropout2d(config.resid_pdrop)  # To reproduce the noise_shape parameter of TF implementation
        self.linear = nn.Linear(config.n_embd, 1)

        nn.init.normal_(self.linear.weight, std = 0.02)
        nn.init.normal_(self.linear.bias, 0)

    def forward(self, hidden_states, multiple_choice_token_mask):
        # Classification logits
        # hidden_states = hidden_states.view(-1, self.n_embd)
        # multiple_choice_token_mask = multiple_choice_token_mask.view(-1, 1).expand_as(hidden_states)
        multiple_choice_h = hidden_states * multiple_choice_token_mask.unsqueeze(-1)
        multiple_choice_h = multiple_choice_h.sum(dim=-2)
        # flat = x[..., 0].contiguous().view(-1)
        # multiple_choice_h = multiple_choice_h[flat == self.multiple_choice_token, :]
        # multiple_choice_h = multiple_choice_h.view(-1, x.size(1), self.n_embd, 1)
        # # This double transposition is there to replicate the behavior
        # # of the noise_shape argument in the tensorflow
        # # implementation.  For more details, see
        # # https://github.com/huggingface/pytorch-openai-transformer-lm/issues/11
        # multiple_choice_h = self.dropout(multiple_choice_h.transpose(1, 2)).transpose(1, 2)
        # multiple_choice_h = multiple_choice_h.contiguous().view(-1, self.n_embd)
        multiple_choice_logits = self.linear(multiple_choice_h).squeeze(-1)
        return multiple_choice_logits


class TransfoXLModel(TransfoXLPreTrainedModel):
    """OpenAI GPT model ("Improving Language Understanding by Generative Pre-Training").

    The main implementation difference between BERT and the OpenAI is the use, in OpenAI GPT, of a single embedding matrix
    to store the word, special ([SEP], [CLS]...) and position embeddings.
    The embeddings are ordered as follow in the word embeddings matrice:
        [0,                                                         ----------------------
         ...                                                        -> word embeddings
         config.vocab_size - 1,                                     ______________________
         config.vocab_size,
         ...                                                        -> special embeddings
         config.vocab_size + config.n_special - 1,                  ______________________
         config.vocab_size + config.n_special,
         ...                                                        -> position embeddings
         total_num_embeddings - 1]                                  ______________________

    where total_num_embeddings can be obtained as config.total_num_embeddings and is:
        total_num_embeddings = config.vocab_size + config.n_special + config.n_ctx
    You should use the associate indices to index the embeddings.

    The special embeddings ([SEP], [CLS]...) are not pre-trained and need to be trained during the fine-tuning if you use them.
    The number of special embeddings can be controled using the `set_num_special_tokens(num_special_tokens)` function.

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

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, sequence_length] (or more generally [d_1, ..., d_n, sequence_length]
            were d_1 ... d_n are arbitrary dimensions) with the word BPE token indices selected in the range [0, config.vocab_size[
        `position_ids`: an optional torch.LongTensor with the same shape as input_ids
            with the position indices (selected in the range [config.vocab_size + config.n_special, config.vocab_size + config.n_special + config.n_ctx - 1[.
        `token_type_ids`: an optional torch.LongTensor with the same shape as input_ids
            You can use it to add a third embedding (the previous two being the word and position embeddings)
            to each token in the sentence.

    Outputs:
        `hidden_states`: the encoded-hidden-states at the top of the model
            as a torch.FloatTensor of size [batch_size, sequence_length, hidden_size]
            (or more generally [d_1, ..., d_n, hidden_size] were d_1 ... d_n are the dimension of input_ids)

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

    config = modeling_transfo_xl.TransfoXLConfig()

    model = modeling_transfo_xl.TransfoXLModel(config)
    hidden_states = model(input_ids)
    ```
    """
    def __init__(self, config):
        super(TransfoXLModel, self).__init__(config)
        total_embeddings_size = config.vocab_size + config.n_special + config.n_ctx
        self.embed = nn.Embedding(total_embeddings_size, config.n_embd)
        self.drop = nn.Dropout(config.embd_pdrop)
        block = Block(config.n_ctx, config, scale=True)
        self.h = nn.ModuleList([copy.deepcopy(block) for _ in range(config.n_layer)])

        self.apply(self.init_weights)
        # nn.init.normal_(self.embed.weight, std=0.02)

    def set_num_special_tokens(self, num_special_tokens):
        " Update input embeddings with new embedding matrice "
        # Update config
        self.config.n_special = num_special_tokens
        # # Build new embeddings and initialize
        old_embed = self.embed
        self.embed = nn.Embedding(self.config.total_num_embeddings, self.config.n_embd)
        # Initialize all new embeddings (in particular the special tokens)
        self.init_weights(self.embed)
        # Copy word and positional embeddings from the previous weights
        self.embed.weight.data[:self.config.vocab_size, :] = old_embed.weight.data[:self.config.vocab_size, :]
        self.embed.weight.data[-self.config.n_ctx:, :] = old_embed.weight.data[-self.config.n_ctx:, :]

    def forward(self, input_ids, position_ids=None, token_type_ids=None):
        if position_ids is None:
            start = self.config.vocab_size + self.config.n_special
            end = start + input_ids.size(-1)
            position_ids = torch.arange(start, end, dtype=torch.long, device=input_ids.device)
            position_ids = position_ids.unsqueeze(0).expand_as(input_ids)

        input_shape = input_ids.size()
        input_ids = input_ids.view(-1, input_ids.size(-1))
        position_ids = position_ids.view(-1, position_ids.size(-1))

        inputs_embeds = self.embed(input_ids)
        position_embeds = self.embed(position_ids)
        if token_type_ids is not None:
            token_type_ids = token_type_ids.view(-1, token_type_ids.size(-1))
            token_type_embeds = self.embed(token_type_ids)
        else:
            token_type_embeds = 0
        # Add the position information to the input embeddings
        # h = e.sum(dim=2)
        hidden_states = inputs_embeds + position_embeds + token_type_embeds
        for block in self.h:
            hidden_states = block(hidden_states)
        return hidden_states.view(*input_shape, hidden_states.size(-1))

class TransfoXLLMHeadModel(TransfoXLPreTrainedModel):
    """OpenAI GPT model with a Language Modeling head ("Improving Language Understanding by Generative Pre-Training").

    There are two main implementation differences between BERT and the OpenAI GPT:
        - the use of an LM loss in OpenAI GPT which means the Transformer is trained to predict the NEXT token for each input token
            vs. predict the SAME token for BERT (i.e. you need to shift your labels to the right)
        - the use, in OpenAI GPT, of a single embedding matrix to store the word, special ([SEP], [CLS]...) and position embeddings.
    The embeddings are ordered as follow in the word embeddings matrice:
        [0,                                                         ----------------------
         ...                                                        -> word embeddings
         config.vocab_size - 1,                                     ______________________
         config.vocab_size,
         ...                                                        -> special embeddings
         config.vocab_size + config.n_special - 1,                  ______________________
         config.vocab_size + config.n_special,
         ...                                                        -> position embeddings
         total_num_embeddings - 1]                                  ______________________

    where total_num_embeddings can be obtained as config.total_num_embeddings and is:
        total_num_embeddings = config.vocab_size + config.n_special + config.n_ctx
    You should use these indices to index the word, special and position embeddings.

    The special embeddings ([SEP], [CLS]...) are not pre-trained and need to be trained during the fine-tuning if you use them.
    The number of special embeddings can be controled using the `set_num_special_tokens(num_special_tokens)` function.

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

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, sequence_length] (or more generally [d_1, ..., d_n, sequence_length]
            were d_1 ... d_n are arbitrary dimensions) with the word BPE token indices selected in the range [0, config.vocab_size[
        `position_ids`: an optional torch.LongTensor with the same shape as input_ids
            with the position indices (selected in the range [config.vocab_size + config.n_special, config.vocab_size + config.n_special + config.n_ctx - 1[.
        `token_type_ids`: an optional torch.LongTensor with the same shape as input_ids
            You can use it to add a third embedding (the previous two being the word and position embeddings)
            to each token in the sentence.
        `lm_labels`: optional 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 `lm_labels` is not `None`:
            Outputs the language modeling loss.
        else:
            `lm_logits`: the language modeling logits as a torch.FloatTensor of size [batch_size, sequence_length, total_num_embeddings]
                (or more generally [d_1, ..., d_n, total_num_embeddings] were d_1 ... d_n are the dimension of input_ids)

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

    config = modeling_transfo_xl.TransfoXLConfig()

    model = modeling_transfo_xl.TransfoXLLMHeadModel(config)
    lm_logits = model(input_ids)
    ```
    """
    def __init__(self, config):
        super(TransfoXLLMHeadModel, self).__init__(config)
        self.transformer = TransfoXLModel(config)
        self.lm_head = TransfoXLLMHead(self.transformer.embed.weight, config)
        self.apply(self.init_weights)

    def set_num_special_tokens(self, num_special_tokens):
        " Update input and output embeddings with new embedding matrice "
        self.transformer.set_num_special_tokens(num_special_tokens)
        self.lm_head.set_embeddings_weights(self.transformer.embed.weight)

    def forward(self, input_ids, position_ids=None, token_type_ids=None, lm_labels=None):
        hidden_states = self.transformer(input_ids, position_ids, token_type_ids)
        lm_logits = self.lm_head(hidden_states)
        if lm_labels is not None:
            loss_fct = CrossEntropyLoss(ignore_index=-1)
            loss = loss_fct(lm_logits.view(-1, lm_logits.size(-1)), lm_labels.view(-1))
            return loss
        return lm_logits

class TransfoXLDoubleHeadsModel(TransfoXLPreTrainedModel):
    """OpenAI GPT model with a Language Modeling and a Multiple Choice heads ("Improving Language Understanding by Generative Pre-Training").

    There are two main implementation differences between BERT and the OpenAI GPT:
        - the use of an LM loss in OpenAI GPT which means the Transformer is trained to predict the NEXT token for each input token
            vs. predict the SAME token for BERT (i.e. you need to shift your labels to the right)
        - the use, in OpenAI GPT, of a single embedding matrix to store the word, special ([SEP], [CLS]...) and position embeddings.
    The embeddings are ordered as follow in the word embeddings matrice:
        [0,                                                         ----------------------
         ...                                                        -> word embeddings
         config.vocab_size - 1,                                     ______________________
         config.vocab_size,
         ...                                                        -> special embeddings
         config.vocab_size + config.n_special - 1,                  ______________________
         config.vocab_size + config.n_special,
         ...                                                        -> position embeddings
         total_num_embeddings - 1]                                  ______________________

    where total_num_embeddings can be obtained as config.total_num_embeddings and is:
        total_num_embeddings = config.vocab_size + config.n_special + config.n_ctx
    You should use these indices to index the word, special and position embeddings.

    The special embeddings ([SEP], [CLS]...) are not pre-trained and need to be trained during the fine-tuning if you use them.
    The number of special embeddings can be controled using the `set_num_special_tokens(num_special_tokens)` function.

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

    Inputs:
        `input_ids`: a torch.LongTensor of shape [batch_size, num_choices, sequence_length]
            with the word BPE token indices selected in the range [0, config.vocab_size[
        `multiple_choice_token_mask`: a torch.LongTensor of shape [batch_size, num_choices, sequence_length]
            with a value of 1 were the last hidden state is (usually the [CLS] token) and 0 otherwise.
        `position_ids`: an optional torch.LongTensor with the same shape as input_ids
            with the position indices (selected in the range [config.vocab_size + config.n_special,
            config.vocab_size + config.n_special + config.n_ctx - 1[.
        `token_type_ids`: an optional torch.LongTensor with the same shape as input_ids
            You can use it to add a third embedding (the previous two being the word and position embeddings)
            to each token in the sentence.
        `lm_labels`: optional language modeling labels: torch.LongTensor of shape [batch_size, num_choices, sequence_length]
            with indices selected in [-1, 0, ..., total_num_embeddings]. All labels set to -1 are ignored (masked), the loss
            is only computed for the labels set in [0, ..., total_num_embeddings]
        `multiple_choice_labels`: optional multiple choice labels: torch.LongTensor of shape [batch_size]
            with indices selected in [0, ..., num_choices].

    Outputs:
        if `lm_labels` and `multiple_choice_labels` are not `None`:
            Outputs a tuple of losses with the language modeling loss and the multiple choice loss.
        else: a tuple with
            `lm_logits`: the language modeling logits as a torch.FloatTensor of size [batch_size, num_choices, sequence_length, total_num_embeddings]
            `multiple_choice_logits`: the multiple choice logits as a torch.FloatTensor of size [batch_size, num_choices]

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

    config = modeling_transfo_xl.TransfoXLConfig()

    model = modeling_transfo_xl.TransfoXLLMHeadModel(config)
    lm_logits, multiple_choice_logits = model(input_ids, multiple_choice_token_mask)
    ```
    """
    def __init__(self, config):
        super(TransfoXLDoubleHeadsModel, self).__init__(config)
        self.transformer = TransfoXLModel(config)
        self.lm_head = TransfoXLLMHead(self.transformer.embed.weight, config)
        self.multiple_choice_head = TransfoXLMultipleChoiceHead(config)
        self.apply(self.init_weights)

    def set_num_special_tokens(self, num_special_tokens):
        " Update input and output embeddings with new embedding matrice "
        self.transformer.set_num_special_tokens(num_special_tokens)
        self.lm_head.set_embeddings_weights(self.transformer.embed.weight)

    def forward(self, input_ids, multiple_choice_token_mask, position_ids=None, token_type_ids=None,
                lm_labels=None, multiple_choice_labels=None):
        hidden_states = self.transformer(input_ids, position_ids, token_type_ids)
        lm_logits = self.lm_head(hidden_states)
        multiple_choice_logits = self.multiple_choice_head(hidden_states, multiple_choice_token_mask)
        losses = []
        if lm_labels is not None:
            loss_fct = CrossEntropyLoss(ignore_index=-1)
            losses.append(loss_fct(lm_logits.view(-1, lm_logits.size(-1)), lm_labels.view(-1)))
        if multiple_choice_labels is not None:
            loss_fct = CrossEntropyLoss()
            losses.append(loss_fct(multiple_choice_logits, multiple_choice_labels.view(-1)))
        if losses:
            return losses
        return lm_logits, multiple_choice_logits