modeling_utils.py

# coding=utf-8
# Copyright 2018 The Google AI Language Team Authors, Facebook AI Research authors and The HuggingFace 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.

import inspect
import os
import re
import warnings
from contextlib import contextmanager
from dataclasses import dataclass
from typing import Any, Callable, Dict, List, Optional, Set, Tuple, Union

import torch
from torch import Tensor, device, nn
from torch.nn import CrossEntropyLoss

from .activations import get_activation
from .configuration_utils import PretrainedConfig
from .deepspeed import deepspeed_config, is_deepspeed_zero3_enabled
from .file_utils import (
    DUMMY_INPUTS,
    FLAX_WEIGHTS_NAME,
    TF2_WEIGHTS_NAME,
    TF_WEIGHTS_NAME,
    WEIGHTS_NAME,
    ModelOutput,
    PushToHubMixin,
    cached_path,
    hf_bucket_url,
    is_offline_mode,
    is_remote_url,
    replace_return_docstrings,
)
from .generation_utils import GenerationMixin
from .utils import logging


logger = logging.get_logger(__name__)


_init_weights = True


@contextmanager
def no_init_weights(_enable=True):
    """
    Context manager to globally disable weight initialization to speed up loading large models.

    TODO(Patrick): Delete safety argument `_enable=True` at next major version. .
    """
    global _init_weights
    if _enable:
        _init_weights = False
    try:
        yield
    finally:
        _init_weights = True


try:
    from torch.nn import Identity
except ImportError:
    # Older PyTorch compatibility
    class Identity(nn.Module):
        r"""A placeholder identity operator that is argument-insensitive."""

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

        def forward(self, input):
            return input


def find_pruneable_heads_and_indices(
    heads: List[int], n_heads: int, head_size: int, already_pruned_heads: Set[int]
) -> Tuple[Set[int], torch.LongTensor]:
    """
    Finds the heads and their indices taking :obj:`already_pruned_heads` into account.

    Args:
        heads (:obj:`List[int]`): List of the indices of heads to prune.
        n_heads (:obj:`int`): The number of heads in the model.
        head_size (:obj:`int`): The size of each head.
        already_pruned_heads (:obj:`Set[int]`): A set of already pruned heads.

    Returns:
        :obj:`Tuple[Set[int], torch.LongTensor]`: A tuple with the remaining heads and their corresponding indices.
    """
    mask = torch.ones(n_heads, head_size)
    heads = set(heads) - already_pruned_heads  # Convert to set and remove already pruned heads
    for head in heads:
        # Compute how many pruned heads are before the head and move the index accordingly
        head = head - sum(1 if h < head else 0 for h in already_pruned_heads)
        mask[head] = 0
    mask = mask.view(-1).contiguous().eq(1)
    index: torch.LongTensor = torch.arange(len(mask))[mask].long()
    return heads, index


def get_parameter_device(parameter: Union[nn.Module, GenerationMixin, "ModuleUtilsMixin"]):
    try:
        return next(parameter.parameters()).device
    except StopIteration:
        # For nn.DataParallel compatibility in PyTorch 1.5

        def find_tensor_attributes(module: nn.Module) -> List[Tuple[str, Tensor]]:
            tuples = [(k, v) for k, v in module.__dict__.items() if torch.is_tensor(v)]
            return tuples

        gen = parameter._named_members(get_members_fn=find_tensor_attributes)
        first_tuple = next(gen)
        return first_tuple[1].device


def get_parameter_dtype(parameter: Union[nn.Module, GenerationMixin, "ModuleUtilsMixin"]):
    try:
        return next(parameter.parameters()).dtype
    except StopIteration:
        # For nn.DataParallel compatibility in PyTorch 1.5

        def find_tensor_attributes(module: nn.Module) -> List[Tuple[str, Tensor]]:
            tuples = [(k, v) for k, v in module.__dict__.items() if torch.is_tensor(v)]
            return tuples

        gen = parameter._named_members(get_members_fn=find_tensor_attributes)
        first_tuple = next(gen)
        return first_tuple[1].dtype


class ModuleUtilsMixin:
    """
    A few utilities for :obj:`torch.nn.Modules`, to be used as a mixin.
    """

    @staticmethod
    def _hook_rss_memory_pre_forward(module, *args, **kwargs):
        try:
            import psutil
        except (ImportError):
            raise ImportError("You need to install psutil (pip install psutil) to use memory tracing.")

        process = psutil.Process(os.getpid())
        mem = process.memory_info()
        module.mem_rss_pre_forward = mem.rss
        return None

    @staticmethod
    def _hook_rss_memory_post_forward(module, *args, **kwargs):
        try:
            import psutil
        except (ImportError):
            raise ImportError("You need to install psutil (pip install psutil) to use memory tracing.")

        process = psutil.Process(os.getpid())
        mem = process.memory_info()
        module.mem_rss_post_forward = mem.rss
        mem_rss_diff = module.mem_rss_post_forward - module.mem_rss_pre_forward
        module.mem_rss_diff = mem_rss_diff + (module.mem_rss_diff if hasattr(module, "mem_rss_diff") else 0)
        return None

    def add_memory_hooks(self):
        """
        Add a memory hook before and after each sub-module forward pass to record increase in memory consumption.

        Increase in memory consumption is stored in a :obj:`mem_rss_diff` attribute for each module and can be reset to
        zero with :obj:`model.reset_memory_hooks_state()`.
        """
        for module in self.modules():
            module.register_forward_pre_hook(self._hook_rss_memory_pre_forward)
            module.register_forward_hook(self._hook_rss_memory_post_forward)
        self.reset_memory_hooks_state()

    def reset_memory_hooks_state(self):
        """
        Reset the :obj:`mem_rss_diff` attribute of each module (see
        :func:`~transformers.modeling_utils.ModuleUtilsMixin.add_memory_hooks`).
        """
        for module in self.modules():
            module.mem_rss_diff = 0
            module.mem_rss_post_forward = 0
            module.mem_rss_pre_forward = 0

    @property
    def device(self) -> device:
        """
        :obj:`torch.device`: The device on which the module is (assuming that all the module parameters are on the same
        device).
        """
        return get_parameter_device(self)

    @property
    def dtype(self) -> torch.dtype:
        """
        :obj:`torch.dtype`: The dtype of the module (assuming that all the module parameters have the same dtype).
        """
        return get_parameter_dtype(self)

    def invert_attention_mask(self, encoder_attention_mask: Tensor) -> Tensor:
        """
        Invert an attention mask (e.g., switches 0. and 1.).

        Args:
            encoder_attention_mask (:obj:`torch.Tensor`): An attention mask.

        Returns:
            :obj:`torch.Tensor`: The inverted attention mask.
        """
        if encoder_attention_mask.dim() == 3:
            encoder_extended_attention_mask = encoder_attention_mask[:, None, :, :]
        if encoder_attention_mask.dim() == 2:
            encoder_extended_attention_mask = encoder_attention_mask[:, None, None, :]
        # T5 has a mask that can compare sequence ids, we can simulate this here with this transposition
        # Cf. https://github.com/tensorflow/mesh/blob/8d2465e9bc93129b913b5ccc6a59aa97abd96ec6/mesh_tensorflow
        # /transformer/transformer_layers.py#L270
        # encoder_extended_attention_mask = (encoder_extended_attention_mask ==
        # encoder_extended_attention_mask.transpose(-1, -2))
        encoder_extended_attention_mask = encoder_extended_attention_mask.to(dtype=self.dtype)  # fp16 compatibility

        if self.dtype == torch.float16:
            encoder_extended_attention_mask = (1.0 - encoder_extended_attention_mask) * -1e4
        elif self.dtype == torch.float32:
            encoder_extended_attention_mask = (1.0 - encoder_extended_attention_mask) * -1e9
        else:
            raise ValueError(
                f"{self.dtype} not recognized. `dtype` should be set to either `torch.float32` or `torch.float16`"
            )

        return encoder_extended_attention_mask

    def get_extended_attention_mask(self, attention_mask: Tensor, input_shape: Tuple[int], device: device) -> Tensor:
        """
        Makes broadcastable attention and causal masks so that future and masked tokens are ignored.

        Arguments:
            attention_mask (:obj:`torch.Tensor`):
                Mask with ones indicating tokens to attend to, zeros for tokens to ignore.
            input_shape (:obj:`Tuple[int]`):
                The shape of the input to the model.
            device: (:obj:`torch.device`):
                The device of the input to the model.

        Returns:
            :obj:`torch.Tensor` The extended attention mask, with a the same dtype as :obj:`attention_mask.dtype`.
        """
        # We can provide a self-attention mask of dimensions [batch_size, from_seq_length, to_seq_length]
        # ourselves in which case we just need to make it broadcastable to all heads.
        if attention_mask.dim() == 3:
            extended_attention_mask = attention_mask[:, None, :, :]
        elif attention_mask.dim() == 2:
            # Provided a padding mask of dimensions [batch_size, seq_length]
            # - if the model is a decoder, apply a causal mask in addition to the padding mask
            # - if the model is an encoder, make the mask broadcastable to [batch_size, num_heads, seq_length, seq_length]
            if self.config.is_decoder:
                batch_size, seq_length = input_shape
                seq_ids = torch.arange(seq_length, device=device)
                causal_mask = seq_ids[None, None, :].repeat(batch_size, seq_length, 1) <= seq_ids[None, :, None]
                # in case past_key_values are used we need to add a prefix ones mask to the causal mask
                # causal and attention masks must have same type with pytorch version < 1.3
                causal_mask = causal_mask.to(attention_mask.dtype)

                if causal_mask.shape[1] < attention_mask.shape[1]:
                    prefix_seq_len = attention_mask.shape[1] - causal_mask.shape[1]
                    causal_mask = torch.cat(
                        [
                            torch.ones(
                                (batch_size, seq_length, prefix_seq_len), device=device, dtype=causal_mask.dtype
                            ),
                            causal_mask,
                        ],
                        axis=-1,
                    )

                extended_attention_mask = causal_mask[:, None, :, :] * attention_mask[:, None, None, :]
            else:
                extended_attention_mask = attention_mask[:, None, None, :]
        else:
            raise ValueError(
                f"Wrong shape for input_ids (shape {input_shape}) or attention_mask (shape {attention_mask.shape})"
            )

        # 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.dtype)  # fp16 compatibility
        extended_attention_mask = (1.0 - extended_attention_mask) * -10000.0
        return extended_attention_mask

    def get_head_mask(
        self, head_mask: Optional[Tensor], num_hidden_layers: int, is_attention_chunked: bool = False
    ) -> Tensor:
        """
        Prepare the head mask if needed.

        Args:
            head_mask (:obj:`torch.Tensor` with shape :obj:`[num_heads]` or :obj:`[num_hidden_layers x num_heads]`, `optional`):
                The mask indicating if we should keep the heads or not (1.0 for keep, 0.0 for discard).
            num_hidden_layers (:obj:`int`):
                The number of hidden layers in the model.
            is_attention_chunked: (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether or not the attentions scores are computed by chunks or not.

        Returns:
            :obj:`torch.Tensor` with shape :obj:`[num_hidden_layers x batch x num_heads x seq_length x seq_length]` or
            list with :obj:`[None]` for each layer.
        """
        if head_mask is not None:
            head_mask = self._convert_head_mask_to_5d(head_mask, num_hidden_layers)
            if is_attention_chunked is True:
                head_mask = head_mask.unsqueeze(-1)
        else:
            head_mask = [None] * num_hidden_layers

        return head_mask

    def _convert_head_mask_to_5d(self, head_mask, num_hidden_layers):
        """-> [num_hidden_layers x batch x num_heads x seq_length x seq_length]"""
        if head_mask.dim() == 1:
            head_mask = head_mask.unsqueeze(0).unsqueeze(0).unsqueeze(-1).unsqueeze(-1)
            head_mask = head_mask.expand(num_hidden_layers, -1, -1, -1, -1)
        elif head_mask.dim() == 2:
            head_mask = head_mask.unsqueeze(1).unsqueeze(-1).unsqueeze(-1)  # We can specify head_mask for each layer
        assert head_mask.dim() == 5, f"head_mask.dim != 5, instead {head_mask.dim()}"
        head_mask = head_mask.to(dtype=self.dtype)  # switch to float if need + fp16 compatibility
        return head_mask

    def num_parameters(self, only_trainable: bool = False, exclude_embeddings: bool = False) -> int:
        """
        Get number of (optionally, trainable or non-embeddings) parameters in the module.

        Args:
            only_trainable (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether or not to return only the number of trainable parameters

            exclude_embeddings (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether or not to return only the number of non-embeddings parameters

        Returns:
            :obj:`int`: The number of parameters.
        """

        if exclude_embeddings:
            embedding_param_names = [
                f"{name}.weight" for name, module_type in self.named_modules() if isinstance(module_type, nn.Embedding)
            ]
            non_embedding_parameters = [
                parameter for name, parameter in self.named_parameters() if name not in embedding_param_names
            ]
            return sum(p.numel() for p in non_embedding_parameters if p.requires_grad or not only_trainable)
        else:
            return sum(p.numel() for p in self.parameters() if p.requires_grad or not only_trainable)

    def estimate_tokens(self, input_dict: Dict[str, Union[torch.Tensor, Any]]) -> int:
        """
        Helper function to estimate the total number of tokens from the model inputs.

        Args:
            inputs (:obj:`dict`): The model inputs.

        Returns:
            :obj:`int`: The total number of tokens.
        """
        token_inputs = [tensor for key, tensor in input_dict.items() if "input" in key]
        if token_inputs:
            return sum([token_input.numel() for token_input in token_inputs])
        else:
            warnings.warn(
                "Could not estimate the number of tokens of the input, floating-point operations will not be computed"
            )
            return 0

    def floating_point_ops(
        self, input_dict: Dict[str, Union[torch.Tensor, Any]], exclude_embeddings: bool = True
    ) -> int:
        """
        Get number of (optionally, non-embeddings) floating-point operations for the forward and backward passes of a
        batch with this transformer model. Default approximation neglects the quadratic dependency on the number of
        tokens (valid if :obj:`12 * d_model << sequence_length`) as laid out in `this paper
        <https://arxiv.org/pdf/2001.08361.pdf>`__ section 2.1. Should be overridden for transformers with parameter
        re-use e.g. Albert or Universal Transformers, or if doing long-range modeling with very high sequence lengths.

        Args:
            batch_size (:obj:`int`):
                The batch size for the forward pass.

            sequence_length (:obj:`int`):
                The number of tokens in each line of the batch.

            exclude_embeddings (:obj:`bool`, `optional`, defaults to :obj:`True`):
                Whether or not to count embedding and softmax operations.

        Returns:
            :obj:`int`: The number of floating-point operations.
        """

        return 6 * self.estimate_tokens(input_dict) * self.num_parameters(exclude_embeddings=exclude_embeddings)


class PreTrainedModel(nn.Module, ModuleUtilsMixin, GenerationMixin, PushToHubMixin):
    r"""
    Base class for all models.

    :class:`~transformers.PreTrainedModel` takes care of storing the configuration of the models and handles methods
    for loading, downloading and saving models as well as a few methods common to all models to:

        * resize the input embeddings,
        * prune heads in the self-attention heads.

    Class attributes (overridden by derived classes):

        - **config_class** (:class:`~transformers.PretrainedConfig`) -- A subclass of
          :class:`~transformers.PretrainedConfig` to use as configuration class for this model architecture.
        - **load_tf_weights** (:obj:`Callable`) -- A python `method` for loading a TensorFlow checkpoint in a PyTorch
          model, taking as arguments:

            - **model** (:class:`~transformers.PreTrainedModel`) -- An instance of the model on which to load the
              TensorFlow checkpoint.
            - **config** (:class:`~transformers.PreTrainedConfig`) -- An instance of the configuration associated to
              the model.
            - **path** (:obj:`str`) -- A path to the TensorFlow checkpoint.

        - **base_model_prefix** (:obj:`str`) -- A string indicating the attribute associated to the base model in
          derived classes of the same architecture adding modules on top of the base model.
        - **is_parallelizable** (:obj:`bool`) -- A flag indicating whether this model supports model parallelization.
    """
    config_class = None
    base_model_prefix = ""
    # a list of re pattern of tensor names to ignore from the model when loading the model weights
    # (and avoid unnecessary warnings).
    _keys_to_ignore_on_load_missing = None
    # a list of re pattern of tensor names to ignore from the weights when loading the model weights
    # (and avoid unnecessary warnings).
    _keys_to_ignore_on_load_unexpected = None
    # a list of of tensor names to ignore when saving the model (useful for keys that aren't
    # trained, but which are deterministic)
    _keys_to_ignore_on_save = None

    is_parallelizable = False

    @property
    def dummy_inputs(self) -> Dict[str, torch.Tensor]:
        """
        :obj:`Dict[str, torch.Tensor]`: Dummy inputs to do a forward pass in the network.
        """
        return {"input_ids": torch.tensor(DUMMY_INPUTS)}

    def __init__(self, config: PretrainedConfig, *inputs, **kwargs):
        super().__init__()
        if not isinstance(config, PretrainedConfig):
            raise ValueError(
                f"Parameter config in `{self.__class__.__name__}(config)` should be an instance of class "
                "`PretrainedConfig`. To create a model from a pretrained model use "
                f"`model = {self.__class__.__name__}.from_pretrained(PRETRAINED_MODEL_NAME)`"
            )
        # Save config and origin of the pretrained weights if given in model
        self.config = config
        self.name_or_path = config.name_or_path

    @classmethod
    def _from_config(cls, config, **kwargs):
        """
        All context managers that the model should be initialized under go here.

        Args:
            torch_dtype (:obj:`torch.dtype`, `optional`):
                Override the default ``torch.dtype`` and load the model under this dtype.
        """
        torch_dtype = kwargs.pop("torch_dtype", None)

        # override default dtype if needed
        dtype_orig = None
        if torch_dtype is not None:
            dtype_orig = cls._set_default_torch_dtype(torch_dtype)

        if is_deepspeed_zero3_enabled():
            import deepspeed

            logger.info("Detected DeepSpeed ZeRO-3: activating zero.init() for this model")
            # this immediately partitions the model across all gpus, to avoid the overhead in time
            # and memory copying it on CPU or each GPU first
            with deepspeed.zero.Init(config=deepspeed_config()):
                model = cls(config, **kwargs)
        else:
            model = cls(config, **kwargs)

        # restore default dtype if it was modified
        if dtype_orig is not None:
            torch.set_default_dtype(dtype_orig)

        return model

    @classmethod
    def _set_default_torch_dtype(cls, dtype: torch.dtype) -> torch.dtype:
        """
        Change the default dtype and return the previous one. This is needed when wanting to instantiate the model
        under specific dtype.

        Args:
            dtype (:obj:`torch.dtype`):
                a floating dtype to set to.

        Returns:
            :obj:`torch.dtype`: the original ``dtype`` that can be used to restore ``torch.set_default_dtype(dtype)``
            if it was modified. If it wasn't, returns :obj:`None`.

        Note ``set_default_dtype`` currently only works with floating-point types and asserts if for example,
        ``torch.int64`` is passed. So if a non-float ``dtype`` is passed this functions will throw an exception.
        """
        if not dtype.is_floating_point:
            raise ValueError(
                f"Can't instantiate {cls.__name__} model under dtype={dtype} since it is not a floating point dtype"
            )

        logger.info(f"Instantiating {cls.__name__} model under default dtype {dtype}.")
        dtype_orig = torch.get_default_dtype()
        torch.set_default_dtype(dtype)
        return dtype_orig

    @property
    def base_model(self) -> nn.Module:
        """
        :obj:`torch.nn.Module`: The main body of the model.
        """
        return getattr(self, self.base_model_prefix, self)

    def get_input_embeddings(self) -> nn.Module:
        """
        Returns the model's input embeddings.

        Returns:
            :obj:`nn.Module`: A torch module mapping vocabulary to hidden states.
        """
        base_model = getattr(self, self.base_model_prefix, self)
        if base_model is not self:
            return base_model.get_input_embeddings()
        else:
            raise NotImplementedError

    def set_input_embeddings(self, value: nn.Module):
        """
        Set model's input embeddings.

        Args:
            value (:obj:`nn.Module`): A module mapping vocabulary to hidden states.
        """
        base_model = getattr(self, self.base_model_prefix, self)
        if base_model is not self:
            base_model.set_input_embeddings(value)
        else:
            raise NotImplementedError

    def get_output_embeddings(self) -> nn.Module:
        """
        Returns the model's output embeddings.

        Returns:
            :obj:`nn.Module`: A torch module mapping hidden states to vocabulary.
        """
        return None  # Overwrite for models with output embeddings

    def _init_weights(self, module):
        """
        Initialize the weights. This method should be overridden by derived class.
        """
        raise NotImplementedError(f"Make sure `_init_weigths` is implemented for {self.__class__}")

    def tie_weights(self):
        """
        Tie the weights between the input embeddings and the output embeddings.

        If the :obj:`torchscript` flag is set in the configuration, can't handle parameter sharing so we are cloning
        the weights instead.
        """
        output_embeddings = self.get_output_embeddings()
        if output_embeddings is not None and self.config.tie_word_embeddings:
            self._tie_or_clone_weights(output_embeddings, self.get_input_embeddings())

        if self.config.is_encoder_decoder and self.config.tie_encoder_decoder:
            if hasattr(self, self.base_model_prefix):
                self = getattr(self, self.base_model_prefix)
            self._tie_encoder_decoder_weights(self.encoder, self.decoder, self.base_model_prefix)

    @staticmethod
    def _tie_encoder_decoder_weights(encoder: nn.Module, decoder: nn.Module, base_model_prefix: str):
        uninitialized_encoder_weights: List[str] = []
        if decoder.__class__ != encoder.__class__:
            logger.info(
                f"{decoder.__class__} and {encoder.__class__} are not equal. In this case make sure that all encoder weights are correctly initialized."
            )

        def tie_encoder_to_decoder_recursively(
            decoder_pointer: nn.Module,
            encoder_pointer: nn.Module,
            module_name: str,
            uninitialized_encoder_weights: List[str],
            depth=0,
        ):
            assert isinstance(decoder_pointer, nn.Module) and isinstance(
                encoder_pointer, nn.Module
            ), f"{decoder_pointer} and {encoder_pointer} have to be of type nn.Module"
            if hasattr(decoder_pointer, "weight"):
                assert hasattr(encoder_pointer, "weight")
                encoder_pointer.weight = decoder_pointer.weight
                if hasattr(decoder_pointer, "bias"):
                    assert hasattr(encoder_pointer, "bias")
                    encoder_pointer.bias = decoder_pointer.bias
                return

            encoder_modules = encoder_pointer._modules
            decoder_modules = decoder_pointer._modules
            if len(decoder_modules) > 0:
                assert (
                    len(encoder_modules) > 0
                ), f"Encoder module {encoder_pointer} does not match decoder module {decoder_pointer}"

                all_encoder_weights = set([module_name + "/" + sub_name for sub_name in encoder_modules.keys()])
                encoder_layer_pos = 0
                for name, module in decoder_modules.items():
                    if name.isdigit():
                        encoder_name = str(int(name) + encoder_layer_pos)
                        decoder_name = name
                        if not isinstance(decoder_modules[decoder_name], type(encoder_modules[encoder_name])) and len(
                            encoder_modules
                        ) != len(decoder_modules):
                            # this can happen if the name corresponds to the position in a list module list of layers
                            # in this case the decoder has added a cross-attention that the encoder does not have
                            # thus skip this step and subtract one layer pos from encoder
                            encoder_layer_pos -= 1
                            continue
                    elif name not in encoder_modules:
                        continue
                    elif depth > 500:
                        raise ValueError(
                            "Max depth of recursive function `tie_encoder_to_decoder` reached. It seems that there is a circular dependency between two or more `nn.Modules` of your model."
                        )
                    else:
                        decoder_name = encoder_name = name
                    tie_encoder_to_decoder_recursively(
                        decoder_modules[decoder_name],
                        encoder_modules[encoder_name],
                        module_name + "/" + name,
                        uninitialized_encoder_weights,
                        depth=depth + 1,
                    )
                    all_encoder_weights.remove(module_name + "/" + encoder_name)

                uninitialized_encoder_weights += list(all_encoder_weights)

        # tie weights recursively
        tie_encoder_to_decoder_recursively(decoder, encoder, base_model_prefix, uninitialized_encoder_weights)
        if len(uninitialized_encoder_weights) > 0:
            logger.warning(
                f"The following encoder weights were not tied to the decoder {uninitialized_encoder_weights}"
            )

    def _tie_or_clone_weights(self, output_embeddings, input_embeddings):
        """Tie or clone module weights depending of whether we are using TorchScript or not"""
        if self.config.torchscript:
            output_embeddings.weight = nn.Parameter(input_embeddings.weight.clone())
        else:
            output_embeddings.weight = input_embeddings.weight

        if getattr(output_embeddings, "bias", None) is not None:
            output_embeddings.bias.data = nn.functional.pad(
                output_embeddings.bias.data,
                (
                    0,
                    output_embeddings.weight.shape[0] - output_embeddings.bias.shape[0],
                ),
                "constant",
                0,
            )
        if hasattr(output_embeddings, "out_features") and hasattr(input_embeddings, "num_embeddings"):
            output_embeddings.out_features = input_embeddings.num_embeddings

    def resize_token_embeddings(self, new_num_tokens: Optional[int] = None) -> nn.Embedding:
        """
        Resizes input token embeddings matrix of the model if :obj:`new_num_tokens != config.vocab_size`.

        Takes care of tying weights embeddings afterwards if the model class has a :obj:`tie_weights()` method.

        Arguments:
            new_num_tokens (:obj:`int`, `optional`):
                The number of new tokens in the embedding matrix. Increasing the size will add newly initialized
                vectors at the end. Reducing the size will remove vectors from the end. If not provided or :obj:`None`,
                just returns a pointer to the input tokens :obj:`torch.nn.Embedding` module of the model without doing
                anything.

        Return:
            :obj:`torch.nn.Embedding`: Pointer to the input tokens Embeddings Module of the model.
        """
        model_embeds = self._resize_token_embeddings(new_num_tokens)
        if new_num_tokens is None:
            return model_embeds

        # Update base model and current model config
        self.config.vocab_size = new_num_tokens
        self.vocab_size = new_num_tokens

        # Tie weights again if needed
        self.tie_weights()

        return model_embeds

    def _resize_token_embeddings(self, new_num_tokens):
        old_embeddings = self.get_input_embeddings()
        new_embeddings = self._get_resized_embeddings(old_embeddings, new_num_tokens)
        self.set_input_embeddings(new_embeddings)

        # if word embeddings are not tied, make sure that lm head is resized as well
        if self.get_output_embeddings() is not None and not self.config.tie_word_embeddings:
            old_lm_head = self.get_output_embeddings()
            new_lm_head = self._get_resized_lm_head(old_lm_head, new_num_tokens)
            self.set_output_embeddings(new_lm_head)

        return self.get_input_embeddings()

    def _get_resized_embeddings(
        self, old_embeddings: nn.Embedding, new_num_tokens: Optional[int] = None
    ) -> nn.Embedding:
        """
        Build a resized Embedding Module from a provided token Embedding Module. Increasing the size will add newly
        initialized vectors at the end. Reducing the size will remove vectors from the end

        Args:
            old_embeddings (:obj:`torch.nn.Embedding`):
                Old embeddings to be resized.
            new_num_tokens (:obj:`int`, `optional`):
                New number of tokens in the embedding matrix.

                Increasing the size will add newly initialized vectors at the end. Reducing the size will remove
                vectors from the end. If not provided or :obj:`None`, just returns a pointer to the input tokens
                :obj:`torch.nn.Embedding`` module of the model without doing anything.

        Return:
            :obj:`torch.nn.Embedding`: Pointer to the resized Embedding Module or the old Embedding Module if
            :obj:`new_num_tokens` is :obj:`None`
        """
        if new_num_tokens is None:
            return old_embeddings

        if is_deepspeed_zero3_enabled():
            import deepspeed

            with deepspeed.zero.GatheredParameters(old_embeddings.weight, modifier_rank=None):
                old_num_tokens, old_embedding_dim = old_embeddings.weight.size()
        else:
            old_num_tokens, old_embedding_dim = old_embeddings.weight.size()

        if old_num_tokens == new_num_tokens:
            return old_embeddings

        if not isinstance(old_embeddings, nn.Embedding):
            raise TypeError(
                f"Old embeddings are of type {type(old_embeddings)}, which is not an instance of {nn.Embedding}."
                f"You should either use a different resize function or make sure that `old_embeddings` are an instance of {nn.Embedding}."
            )

        # Build new embeddings
        new_embeddings = nn.Embedding(new_num_tokens, old_embedding_dim).to(
            self.device, dtype=old_embeddings.weight.dtype
        )

        # initialize all new embeddings (in particular added tokens)
        self._init_weights(new_embeddings)

        # Copy token embeddings from the previous weights

        # numbers of tokens to copy
        n = min(old_num_tokens, new_num_tokens)
        if is_deepspeed_zero3_enabled():
            import deepspeed

            with deepspeed.zero.GatheredParameters(old_embeddings.weight, modifier_rank=0):
                if torch.distributed.get_rank() == 0:
                    new_embeddings.weight.data[:n, :] = old_embeddings.weight.data[:n, :]
        else:
            new_embeddings.weight.data[:n, :] = old_embeddings.weight.data[:n, :]

        return new_embeddings

    def _get_resized_lm_head(
        self, old_lm_head: nn.Linear, new_num_tokens: Optional[int] = None, transposed: Optional[bool] = False
    ) -> nn.Linear:
        """
        Build a resized Linear Module from a provided old Linear Module. Increasing the size will add newly initialized
        vectors at the end. Reducing the size will remove vectors from the end

        Args:
            old_lm_head (:obj:`torch.nn.Linear`):
                Old lm head liner layer to be resized.
            new_num_tokens (:obj:`int`, `optional`):
                New number of tokens in the linear matrix.

                Increasing the size will add newly initialized vectors at the end. Reducing the size will remove
                vectors from the end. If not provided or :obj:`None`, just returns a pointer to the input tokens
                :obj:`torch.nn.Linear`` module of the model without doing anything.
            transposed (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether ``old_lm_head`` is transposed or not. If True ``old_lm_head.size()`` is ``lm_head_dim,
                vocab_size`` else ``vocab_size, lm_head_dim``.

        Return:
            :obj:`torch.nn.Linear`: Pointer to the resized Linear Module or the old Linear Module if
            :obj:`new_num_tokens` is :obj:`None`
        """
        if new_num_tokens is None:
            return old_lm_head

        old_num_tokens, old_lm_head_dim = (
            old_lm_head.weight.size() if not transposed else old_lm_head.weight.t().size()
        )

        if old_num_tokens == new_num_tokens:
            return old_lm_head

        if not isinstance(old_lm_head, nn.Linear):
            raise TypeError(
                f"Old language model head is of type {type(old_lm_head)}, which is not an instance of {nn.Linear}."
                f"You should either use a different resize function or make sure that `old_embeddings` are an instance of {nn.Linear}."
            )

        # Build new lm head
        new_lm_head_shape = (old_lm_head_dim, new_num_tokens) if not transposed else (new_num_tokens, old_lm_head_dim)
        has_new_lm_head_bias = old_lm_head.bias is not None
        new_lm_head = nn.Linear(*new_lm_head_shape, bias=has_new_lm_head_bias).to(self.device)

        # initialize new lm head (in particular added tokens)
        self._init_weights(new_lm_head)

        num_tokens_to_copy = min(old_num_tokens, new_num_tokens)

        # Copy old lm head weights to new lm head
        if not transposed:
            new_lm_head.weight.data[:num_tokens_to_copy, :] = old_lm_head.weight.data[:num_tokens_to_copy, :]
        else:
            new_lm_head.weight.data[:, :num_tokens_to_copy] = old_lm_head.weight.data[:, :num_tokens_to_copy]

        # Copy bias weights to new lm head
        if has_new_lm_head_bias:
            new_lm_head.bias.data[:num_tokens_to_copy] = old_lm_head.bias.data[:num_tokens_to_copy]

        return new_lm_head

    def init_weights(self):
        """
        If needed prunes and maybe initializes weights.
        """
        # Prune heads if needed
        if self.config.pruned_heads:
            self.prune_heads(self.config.pruned_heads)

        if _init_weights:
            # Initialize weights
            self.apply(self._init_weights)

            # Tie weights should be skipped when not initializing all weights
            # since from_pretrained(...) calls tie weights anyways
            self.tie_weights()

    def prune_heads(self, heads_to_prune: Dict[int, List[int]]):
        """
        Prunes heads of the base model.

        Arguments:
            heads_to_prune (:obj:`Dict[int, List[int]]`):
                Dictionary with keys being selected layer indices (:obj:`int`) and associated values being the list of
                heads to prune in said layer (list of :obj:`int`). For instance {1: [0, 2], 2: [2, 3]} will prune heads
                0 and 2 on layer 1 and heads 2 and 3 on layer 2.
        """
        # save new sets of pruned heads as union of previously stored pruned heads and newly pruned heads
        for layer, heads in heads_to_prune.items():
            union_heads = set(self.config.pruned_heads.get(layer, [])) | set(heads)
            self.config.pruned_heads[layer] = list(union_heads)  # Unfortunately we have to store it as list for JSON

        self.base_model._prune_heads(heads_to_prune)

    def save_pretrained(
        self,
        save_directory: Union[str, os.PathLike],
        save_config: bool = True,
        state_dict: Optional[dict] = None,
        save_function: Callable = torch.save,
        push_to_hub: bool = False,
        **kwargs,
    ):
        """
        Save a model and its configuration file to a directory, so that it can be re-loaded using the
        `:func:`~transformers.PreTrainedModel.from_pretrained`` class method.

        Arguments:
            save_directory (:obj:`str` or :obj:`os.PathLike`):
                Directory to which to save. Will be created if it doesn't exist.
            save_config (:obj:`bool`, `optional`, defaults to :obj:`True`):
                Whether or not to save the config of the model. Useful when in distributed training like TPUs and need
                to call this function on all processes. In this case, set :obj:`save_config=True` only on the main
                process to avoid race conditions.
            state_dict (nested dictionary of :obj:`torch.Tensor`):
                The state dictionary of the model to save. Will default to :obj:`self.state_dict()`, but can be used to
                only save parts of the model or if special precautions need to be taken when recovering the state
                dictionary of a model (like when using model parallelism).
            save_function (:obj:`Callable`):
                The function to use to save the state dictionary. Useful on distributed training like TPUs when one
                need to replace :obj:`torch.save` by another method.
            push_to_hub (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether or not to push your model to the Hugging Face model hub after saving it.

                .. warning::

                    Using :obj:`push_to_hub=True` will synchronize the repository you are pushing to with
                    :obj:`save_directory`, which requires :obj:`save_directory` to be a local clone of the repo you are
                    pushing to if it's an existing folder. Pass along :obj:`temp_dir=True` to use a temporary directory
                    instead.

            kwargs:
                Additional key word arguments passed along to the
                :meth:`~transformers.file_utils.PushToHubMixin.push_to_hub` method.
        """
        if os.path.isfile(save_directory):
            logger.error(f"Provided path ({save_directory}) should be a directory, not a file")
            return

        if push_to_hub:
            commit_message = kwargs.pop("commit_message", None)
            repo = self._create_or_get_repo(save_directory, **kwargs)

        os.makedirs(save_directory, exist_ok=True)

        # Only save the model itself if we are using distributed training
        model_to_save = unwrap_model(self)

        # save the string version of dtype to the config, e.g. convert torch.float32 => "float32"
        # we currently don't use this setting automatically, but may start to use with v5
        dtype = get_parameter_dtype(model_to_save)
        model_to_save.config.torch_dtype = str(dtype).split(".")[1]

        # Attach architecture to the config
        model_to_save.config.architectures = [model_to_save.__class__.__name__]

        # Save the config
        if save_config:
            model_to_save.config.save_pretrained(save_directory)

        # Save the model
        if state_dict is None:
            state_dict = model_to_save.state_dict()

        # Handle the case where some state_dict keys shouldn't be saved
        if self._keys_to_ignore_on_save is not None:
            state_dict = {k: v for k, v in state_dict.items() if k not in self._keys_to_ignore_on_save}

        # If we save using the predefined names, we can load using `from_pretrained`
        output_model_file = os.path.join(save_directory, WEIGHTS_NAME)
        save_function(state_dict, output_model_file)

        logger.info(f"Model weights saved in {output_model_file}")

        if push_to_hub:
            url = self._push_to_hub(repo, commit_message=commit_message)
            logger.info(f"Model pushed to the hub in this commit: {url}")

    @classmethod
    def from_pretrained(cls, pretrained_model_name_or_path: Optional[Union[str, os.PathLike]], *model_args, **kwargs):
        r"""
        Instantiate a pretrained pytorch model from a pre-trained model configuration.

        The model is set in evaluation mode by default using ``model.eval()`` (Dropout modules are deactivated). To
        train the model, you should first set it back in training mode with ``model.train()``.

        The warning `Weights from XXX not initialized from pretrained model` means that the weights of XXX do not come
        pretrained with the rest of the model. It is up to you to train those weights with a downstream fine-tuning
        task.

        The warning `Weights from XXX not used in YYY` means that the layer XXX is not used by YYY, therefore those
        weights are discarded.

        Parameters:
            pretrained_model_name_or_path (:obj:`str` or :obj:`os.PathLike`, `optional`):
                Can be either:

                    - A string, the `model id` of a pretrained model hosted inside a model repo on huggingface.co.
                      Valid model ids can be located at the root-level, like ``bert-base-uncased``, or namespaced under
                      a user or organization name, like ``dbmdz/bert-base-german-cased``.
                    - A path to a `directory` containing model weights saved using
                      :func:`~transformers.PreTrainedModel.save_pretrained`, e.g., ``./my_model_directory/``.
                    - A path or url to a `tensorflow index checkpoint file` (e.g, ``./tf_model/model.ckpt.index``). In
                      this case, ``from_tf`` should be set to :obj:`True` and a configuration object should be provided
                      as ``config`` argument. This loading path is slower than converting the TensorFlow checkpoint in
                      a PyTorch model using the provided conversion scripts and loading the PyTorch model afterwards.
                    - A path or url to a model folder containing a `flax checkpoint file` in `.msgpack` format (e.g,
                      ``./flax_model/`` containing ``flax_model.msgpack``). In this case, ``from_flax`` should be set
                      to :obj:`True`.
                    - :obj:`None` if you are both providing the configuration and state dictionary (resp. with keyword
                      arguments ``config`` and ``state_dict``).
            model_args (sequence of positional arguments, `optional`):
                All remaning positional arguments will be passed to the underlying model's ``__init__`` method.
            config (:obj:`Union[PretrainedConfig, str, os.PathLike]`, `optional`):
                Can be either:

                    - an instance of a class derived from :class:`~transformers.PretrainedConfig`,
                    - a string or path valid as input to :func:`~transformers.PretrainedConfig.from_pretrained`.

                Configuration for the model to use instead of an automatically loaded configuation. Configuration can
                be automatically loaded when:

                    - The model is a model provided by the library (loaded with the `model id` string of a pretrained
                      model).
                    - The model was saved using :func:`~transformers.PreTrainedModel.save_pretrained` and is reloaded
                      by supplying the save directory.
                    - The model is loaded by supplying a local directory as ``pretrained_model_name_or_path`` and a
                      configuration JSON file named `config.json` is found in the directory.
            state_dict (:obj:`Dict[str, torch.Tensor]`, `optional`):
                A state dictionary to use instead of a state dictionary loaded from saved weights file.

                This option can be used if you want to create a model from a pretrained configuration but load your own
                weights. In this case though, you should check if using
                :func:`~transformers.PreTrainedModel.save_pretrained` and
                :func:`~transformers.PreTrainedModel.from_pretrained` is not a simpler option.
            cache_dir (:obj:`Union[str, os.PathLike]`, `optional`):
                Path to a directory in which a downloaded pretrained model configuration should be cached if the
                standard cache should not be used.
            from_tf (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Load the model weights from a TensorFlow checkpoint save file (see docstring of
                ``pretrained_model_name_or_path`` argument).
            from_flax (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Load the model weights from a Flax checkpoint save file (see docstring of
                ``pretrained_model_name_or_path`` argument).
            force_download (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether or not to force the (re-)download of the model weights and configuration files, overriding the
                cached versions if they exist.
            resume_download (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether or not to delete incompletely received files. Will attempt to resume the download if such a
                file exists.
            proxies (:obj:`Dict[str, str], `optional`):
                A dictionary of proxy servers to use by protocol or endpoint, e.g., :obj:`{'http': 'foo.bar:3128',
                'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request.
            output_loading_info(:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether ot not to also return a dictionary containing missing keys, unexpected keys and error messages.
            local_files_only(:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether or not to only look at local files (i.e., do not try to download the model).
            use_auth_token (:obj:`str` or `bool`, `optional`):
                The token to use as HTTP bearer authorization for remote files. If :obj:`True`, will use the token
                generated when running :obj:`transformers-cli login` (stored in :obj:`~/.huggingface`).
            revision(:obj:`str`, `optional`, defaults to :obj:`"main"`):
                The specific model version to use. It can be a branch name, a tag name, or a commit id, since we use a
                git-based system for storing models and other artifacts on huggingface.co, so ``revision`` can be any
                identifier allowed by git.
            mirror(:obj:`str`, `optional`):
                Mirror source to accelerate downloads in China. If you are from China and have an accessibility
                problem, you can set this option to resolve it. Note that we do not guarantee the timeliness or safety.
                Please refer to the mirror site for more information.
            _fast_init(:obj:`bool`, `optional`, defaults to `:obj:`True`):
                Whether or not to disable fast initialization.
            torch_dtype (:obj:`str` or :obj:`torch.dtype`, `optional`):
                Override the default ``torch.dtype`` and load the model under this dtype. If ``"auto"`` is passed the
                dtype will be automatically derived from the model's weights.

                .. warning::

                    One should only disable `_fast_init` to ensure backwards compatibility with
                    ``transformers.__version__ < 4.6.0`` for seeded model initialization. This argument will be removed
                    at the next major version. See `pull request 11471
                    <https://github.com/huggingface/transformers/pull/11471>`__ for more information.

            kwargs (remaining dictionary of keyword arguments, `optional`):
                Can be used to update the configuration object (after it being loaded) and initiate the model (e.g.,
                :obj:`output_attentions=True`). Behaves differently depending on whether a ``config`` is provided or
                automatically loaded:

                    - If a configuration is provided with ``config``, ``**kwargs`` will be directly passed to the
                      underlying model's ``__init__`` method (we assume all relevant updates to the configuration have
                      already been done)
                    - If a configuration is not provided, ``kwargs`` will be first passed to the configuration class
                      initialization function (:func:`~transformers.PretrainedConfig.from_pretrained`). Each key of
                      ``kwargs`` that corresponds to a configuration attribute will be used to override said attribute
                      with the supplied ``kwargs`` value. Remaining keys that do not correspond to any configuration
                      attribute will be passed to the underlying model's ``__init__`` function.

        .. note::

            Passing :obj:`use_auth_token=True` is required when you want to use a private model.

        .. note::

            Activate the special `"offline-mode"
            <https://huggingface.co/transformers/installation.html#offline-mode>`__ to use this method in a firewalled
            environment.

        Examples::

            >>> from transformers import BertConfig, BertModel
            >>> # Download model and configuration from huggingface.co and cache.
            >>> model = BertModel.from_pretrained('bert-base-uncased')
            >>> # Model was saved using `save_pretrained('./test/saved_model/')` (for example purposes, not runnable).
            >>> model = BertModel.from_pretrained('./test/saved_model/')
            >>> # Update configuration during loading.
            >>> model = BertModel.from_pretrained('bert-base-uncased', output_attentions=True)
            >>> assert model.config.output_attentions == True
            >>> # Loading from a TF checkpoint file instead of a PyTorch model (slower, for example purposes, not runnable).
            >>> config = BertConfig.from_json_file('./tf_model/my_tf_model_config.json')
            >>> model = BertModel.from_pretrained('./tf_model/my_tf_checkpoint.ckpt.index', from_tf=True, config=config)
            >>> # Loading from a Flax checkpoint file instead of a PyTorch model (slower)
            >>> model = BertModel.from_pretrained('bert-base-uncased', from_flax=True)

        """
        config = kwargs.pop("config", None)
        state_dict = kwargs.pop("state_dict", None)
        cache_dir = kwargs.pop("cache_dir", None)
        from_tf = kwargs.pop("from_tf", False)
        from_flax = kwargs.pop("from_flax", False)
        force_download = kwargs.pop("force_download", False)
        resume_download = kwargs.pop("resume_download", False)
        proxies = kwargs.pop("proxies", None)
        output_loading_info = kwargs.pop("output_loading_info", False)
        local_files_only = kwargs.pop("local_files_only", False)
        use_auth_token = kwargs.pop("use_auth_token", None)
        revision = kwargs.pop("revision", None)
        mirror = kwargs.pop("mirror", None)
        from_pipeline = kwargs.pop("_from_pipeline", None)
        from_auto_class = kwargs.pop("_from_auto", False)
        _fast_init = kwargs.pop("_fast_init", True)
        torch_dtype = kwargs.pop("torch_dtype", None)

        from_pt = not (from_tf | from_flax)

        user_agent = {"file_type": "model", "framework": "pytorch", "from_auto_class": from_auto_class}
        if from_pipeline is not None:
            user_agent["using_pipeline"] = from_pipeline

        if is_offline_mode() and not local_files_only:
            logger.info("Offline mode: forcing local_files_only=True")
            local_files_only = True

        # Load config if we don't provide a configuration
        if not isinstance(config, PretrainedConfig):
            config_path = config if config is not None else pretrained_model_name_or_path
            config, model_kwargs = cls.config_class.from_pretrained(
                config_path,
                *model_args,
                cache_dir=cache_dir,
                return_unused_kwargs=True,
                force_download=force_download,
                resume_download=resume_download,
                proxies=proxies,
                local_files_only=local_files_only,
                use_auth_token=use_auth_token,
                revision=revision,
                _from_auto=from_auto_class,
                _from_pipeline=from_pipeline,
                **kwargs,
            )
        else:
            model_kwargs = kwargs

        # Load model
        if pretrained_model_name_or_path is not None:
            pretrained_model_name_or_path = str(pretrained_model_name_or_path)
            if os.path.isdir(pretrained_model_name_or_path):
                if from_tf and os.path.isfile(os.path.join(pretrained_model_name_or_path, TF_WEIGHTS_NAME + ".index")):
                    # Load from a TF 1.0 checkpoint in priority if from_tf
                    archive_file = os.path.join(pretrained_model_name_or_path, TF_WEIGHTS_NAME + ".index")
                elif from_tf and os.path.isfile(os.path.join(pretrained_model_name_or_path, TF2_WEIGHTS_NAME)):
                    # Load from a TF 2.0 checkpoint in priority if from_tf
                    archive_file = os.path.join(pretrained_model_name_or_path, TF2_WEIGHTS_NAME)
                elif from_flax and os.path.isfile(os.path.join(pretrained_model_name_or_path, FLAX_WEIGHTS_NAME)):
                    # Load from a Flax checkpoint in priority if from_flax
                    archive_file = os.path.join(pretrained_model_name_or_path, FLAX_WEIGHTS_NAME)
                elif os.path.isfile(os.path.join(pretrained_model_name_or_path, WEIGHTS_NAME)):
                    # Load from a PyTorch checkpoint
                    archive_file = os.path.join(pretrained_model_name_or_path, WEIGHTS_NAME)
                else:
                    raise EnvironmentError(
                        f"Error no file named {[WEIGHTS_NAME, TF2_WEIGHTS_NAME, TF_WEIGHTS_NAME + '.index', FLAX_WEIGHTS_NAME]} found in "
                        f"directory {pretrained_model_name_or_path} or `from_tf` and `from_flax` set to False."
                    )
            elif os.path.isfile(pretrained_model_name_or_path) or is_remote_url(pretrained_model_name_or_path):
                archive_file = pretrained_model_name_or_path
            elif os.path.isfile(pretrained_model_name_or_path + ".index"):
                if not from_tf:
                    raise ValueError(
                        f"We found a TensorFlow checkpoint at {pretrained_model_name_or_path + '.index'}, please set "
                        "from_tf to True to load from this checkpoint."
                    )
                archive_file = pretrained_model_name_or_path + ".index"
            else:
                # set correct filename
                if from_tf:
                    filename = TF2_WEIGHTS_NAME
                elif from_flax:
                    filename = FLAX_WEIGHTS_NAME
                else:
                    filename = WEIGHTS_NAME

                archive_file = hf_bucket_url(
                    pretrained_model_name_or_path,
                    filename=filename,
                    revision=revision,
                    mirror=mirror,
                )

            try:
                # Load from URL or cache if already cached
                resolved_archive_file = cached_path(
                    archive_file,
                    cache_dir=cache_dir,
                    force_download=force_download,
                    proxies=proxies,
                    resume_download=resume_download,
                    local_files_only=local_files_only,
                    use_auth_token=use_auth_token,
                    user_agent=user_agent,
                )
            except EnvironmentError as err:
                logger.error(err)
                msg = (
                    f"Can't load weights for '{pretrained_model_name_or_path}'. Make sure that:\n\n"
                    f"- '{pretrained_model_name_or_path}' is a correct model identifier listed on 'https://huggingface.co/models'\n\n"
                    f"- or '{pretrained_model_name_or_path}' is the correct path to a directory containing a file named one of {WEIGHTS_NAME}, {TF2_WEIGHTS_NAME}, {TF_WEIGHTS_NAME}.\n\n"
                )
                raise EnvironmentError(msg)

            if resolved_archive_file == archive_file:
                logger.info(f"loading weights file {archive_file}")
            else:
                logger.info(f"loading weights file {archive_file} from cache at {resolved_archive_file}")
        else:
            resolved_archive_file = None

        # load pt weights early so that we know which dtype to init the model under
        if from_pt:
            if state_dict is None:
                try:
                    state_dict = torch.load(resolved_archive_file, map_location="cpu")
                except Exception:
                    raise OSError(
                        f"Unable to load weights from pytorch checkpoint file for '{pretrained_model_name_or_path}' "
                        f"at '{resolved_archive_file}'"
                        "If you tried to load a PyTorch model from a TF 2.0 checkpoint, please set from_tf=True. "
                    )

            # set dtype to instantiate the model under:
            # 1. If torch_dtype is not None, we use that dtype
            # 2. If torch_dtype is "auto", we auto-detect dtype from the loaded state_dict, by checking its first
            #    weights entry - we assume all weights are of the same dtype
            # we also may have config.torch_dtype available, but we won't rely on it till v5
            dtype_orig = None
            if torch_dtype is not None:
                if isinstance(torch_dtype, str):
                    if torch_dtype == "auto":
                        torch_dtype = next(iter(state_dict.values())).dtype
                    else:
                        raise ValueError(
                            f"`torch_dtype` can be either a `torch.dtype` or `auto`, but received {torch_dtype}"
                        )
                dtype_orig = cls._set_default_torch_dtype(torch_dtype)

        config.name_or_path = pretrained_model_name_or_path

        # Instantiate model.
        if is_deepspeed_zero3_enabled():
            import deepspeed

            logger.info("Detected DeepSpeed ZeRO-3: activating zero.init() for this model")
            # this immediately partitions the model across all gpus, to avoid the overhead in time
            # and memory copying it on CPU or each GPU first
            with deepspeed.zero.Init(config=deepspeed_config()):
                with no_init_weights(_enable=_fast_init):
                    model = cls(config, *model_args, **model_kwargs)
        else:
            with no_init_weights(_enable=_fast_init):
                model = cls(config, *model_args, **model_kwargs)

        if from_pt:
            # restore default dtype
            if dtype_orig is not None:
                torch.set_default_dtype(dtype_orig)

        if from_tf:
            if resolved_archive_file.endswith(".index"):
                # Load from a TensorFlow 1.X checkpoint - provided by original authors
                model = cls.load_tf_weights(model, config, resolved_archive_file[:-6])  # Remove the '.index'
            else:
                # Load from our TensorFlow 2.0 checkpoints
                try:
                    from .modeling_tf_pytorch_utils import load_tf2_checkpoint_in_pytorch_model

                    model = load_tf2_checkpoint_in_pytorch_model(model, resolved_archive_file, allow_missing_keys=True)
                except ImportError:
                    logger.error(
                        "Loading a TensorFlow model in PyTorch, requires both PyTorch and TensorFlow to be installed. Please see "
                        "https://pytorch.org/ and https://www.tensorflow.org/install/ for installation instructions."
                    )
                    raise
        elif from_flax:
            try:
                from .modeling_flax_pytorch_utils import load_flax_checkpoint_in_pytorch_model

                model = load_flax_checkpoint_in_pytorch_model(model, resolved_archive_file)
            except ImportError:
                logger.error(
                    "Loading a Flax model in PyTorch, requires both PyTorch and Flax to be installed. Please see "
                    "https://pytorch.org/ and https://flax.readthedocs.io/en/latest/installation.html for installation instructions."
                )
                raise
        elif from_pt:
            model, missing_keys, unexpected_keys, error_msgs = cls._load_state_dict_into_model(
                model, state_dict, pretrained_model_name_or_path, _fast_init=_fast_init
            )

        # make sure token embedding weights are still tied if needed
        model.tie_weights()

        # Set model in evaluation mode to deactivate DropOut modules by default
        model.eval()

        if output_loading_info:
            loading_info = {
                "missing_keys": missing_keys,
                "unexpected_keys": unexpected_keys,
                "error_msgs": error_msgs,
            }
            return model, loading_info

        return model

    @classmethod
    def _load_state_dict_into_model(cls, model, state_dict, pretrained_model_name_or_path, _fast_init=True):

        # Convert old format to new format if needed 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)

        # Retrieve missing & unexpected_keys
        expected_keys = list(model.state_dict().keys())
        loaded_keys = list(state_dict.keys())
        prefix = model.base_model_prefix

        has_prefix_module = any(s.startswith(prefix) for s in loaded_keys)
        expects_prefix_module = any(s.startswith(prefix) for s in expected_keys)

        # key re-naming operations are never done on the keys
        # that are loaded, but always on the keys of the newly initialized model
        remove_prefix = not has_prefix_module and expects_prefix_module
        add_prefix = has_prefix_module and not expects_prefix_module

        if remove_prefix:
            expected_keys = [".".join(s.split(".")[1:]) if s.startswith(prefix) else s for s in expected_keys]
        elif add_prefix:
            expected_keys = [".".join([prefix, s]) for s in expected_keys]

        missing_keys = list(set(expected_keys) - set(loaded_keys))
        unexpected_keys = list(set(loaded_keys) - set(expected_keys))

        # Some models may have keys that are not in the state by design, removing them before needlessly warning
        # the user.
        if cls._keys_to_ignore_on_load_missing is not None:
            for pat in cls._keys_to_ignore_on_load_missing:
                missing_keys = [k for k in missing_keys if re.search(pat, k) is None]

        if cls._keys_to_ignore_on_load_unexpected is not None:
            for pat in cls._keys_to_ignore_on_load_unexpected:
                unexpected_keys = [k for k in unexpected_keys if re.search(pat, k) is None]

        if _fast_init:
            # retrieve unintialized modules and initialize
            unintialized_modules = model.retrieve_modules_from_names(
                missing_keys, add_prefix=add_prefix, remove_prefix=remove_prefix
            )
            for module in unintialized_modules:
                model._init_weights(module)

        # 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

        error_msgs = []

        # PyTorch's `_load_from_state_dict` does not copy parameters in a module's descendants
        # so we need to apply the function recursively.
        def load(module: nn.Module, prefix=""):
            local_metadata = {} if metadata is None else metadata.get(prefix[:-1], {})
            args = (state_dict, prefix, local_metadata, True, [], [], error_msgs)
            if is_deepspeed_zero3_enabled():
                import deepspeed

                # because zero3 puts placeholders in model params, this context
                # manager gathers (unpartitions) the params of the current layer, then loads from
                # the state dict and then re-partitions them again
                with deepspeed.zero.GatheredParameters(list(module.parameters(recurse=False)), modifier_rank=0):
                    if torch.distributed.get_rank() == 0:
                        module._load_from_state_dict(*args)
            else:
                module._load_from_state_dict(*args)

            for name, child in module._modules.items():
                if child is not None:
                    load(child, prefix + name + ".")

        # Make sure we are able to load base models as well as derived models (with heads)
        start_prefix = ""
        model_to_load = model
        if not hasattr(model, cls.base_model_prefix) and has_prefix_module:
            start_prefix = cls.base_model_prefix + "."
        if hasattr(model, cls.base_model_prefix) and not has_prefix_module:
            model_to_load = getattr(model, cls.base_model_prefix)

        load(model_to_load, prefix=start_prefix)

        if len(unexpected_keys) > 0:
            logger.warning(
                f"Some weights of the model checkpoint at {pretrained_model_name_or_path} were not used when "
                f"initializing {model.__class__.__name__}: {unexpected_keys}\n"
                f"- This IS expected if you are initializing {model.__class__.__name__} from the checkpoint of a model trained on another task "
                f"or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n"
                f"- This IS NOT expected if you are initializing {model.__class__.__name__} from the checkpoint of a model that you expect "
                f"to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model)."
            )
        else:
            logger.info(f"All model checkpoint weights were used when initializing {model.__class__.__name__}.\n")
        if len(missing_keys) > 0:
            logger.warning(
                f"Some weights of {model.__class__.__name__} were not initialized from the model checkpoint at {pretrained_model_name_or_path} "
                f"and are newly initialized: {missing_keys}\n"
                f"You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference."
            )
        else:
            logger.info(
                f"All the weights of {model.__class__.__name__} were initialized from the model checkpoint at {pretrained_model_name_or_path}.\n"
                f"If your task is similar to the task the model of the checkpoint was trained on, "
                f"you can already use {model.__class__.__name__} for predictions without further training."
            )
        if len(error_msgs) > 0:
            error_msg = "\n\t".join(error_msgs)
            raise RuntimeError(f"Error(s) in loading state_dict for {model.__class__.__name__}:\n\t{error_msg}")

        return model, missing_keys, unexpected_keys, error_msgs

    def retrieve_modules_from_names(self, names, add_prefix=False, remove_prefix=False):
        module_keys = set([".".join(key.split(".")[:-1]) for key in names])

        # torch.nn.ParameterList is a special case where two parameter keywords
        # are appended to the module name, *e.g.* bert.special_embeddings.0
        module_keys = module_keys.union(set([".".join(key.split(".")[:-2]) for key in names if key[-1].isdigit()]))

        retrieved_modules = []
        # retrieve all modules that has at least one missing weight name
        for name, module in self.named_modules():
            if remove_prefix:
                name = ".".join(name.split(".")[1:]) if name.startswith(self.base_model_prefix) else name
            elif add_prefix:
                name = ".".join([self.base_model_prefix, name]) if len(name) > 0 else self.base_model_prefix

            if name in module_keys:
                retrieved_modules.append(module)

        return retrieved_modules


class Conv1D(nn.Module):
    """
    1D-convolutional layer as defined by Radford et al. for OpenAI GPT (and also used in GPT-2).

    Basically works like a linear layer but the weights are transposed.

    Args:
        nf (:obj:`int`): The number of output features.
        nx (:obj:`int`): The number of input features.
    """

    def __init__(self, nf, nx):
        super().__init__()
        self.nf = nf
        w = torch.empty(nx, nf)
        nn.init.normal_(w, std=0.02)
        self.weight = nn.Parameter(w)
        self.bias = nn.Parameter(torch.zeros(nf))

    def forward(self, x):
        size_out = x.size()[:-1] + (self.nf,)
        x = torch.addmm(self.bias, x.view(-1, x.size(-1)), self.weight)
        x = x.view(*size_out)
        return x


class PoolerStartLogits(nn.Module):
    """
    Compute SQuAD start logits from sequence hidden states.

    Args:
        config (:class:`~transformers.PretrainedConfig`):
            The config used by the model, will be used to grab the :obj:`hidden_size` of the model.
    """

    def __init__(self, config: PretrainedConfig):
        super().__init__()
        self.dense = nn.Linear(config.hidden_size, 1)

    def forward(
        self, hidden_states: torch.FloatTensor, p_mask: Optional[torch.FloatTensor] = None
    ) -> torch.FloatTensor:
        """
        Args:
            hidden_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, seq_len, hidden_size)`):
                The final hidden states of the model.
            p_mask (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, seq_len)`, `optional`):
                Mask for tokens at invalid position, such as query and special symbols (PAD, SEP, CLS). 1.0 means token
                should be masked.

        Returns:
            :obj:`torch.FloatTensor`: The start logits for SQuAD.
        """
        x = self.dense(hidden_states).squeeze(-1)

        if p_mask is not None:
            if get_parameter_dtype(self) == torch.float16:
                x = x * (1 - p_mask) - 65500 * p_mask
            else:
                x = x * (1 - p_mask) - 1e30 * p_mask

        return x


class PoolerEndLogits(nn.Module):
    """
    Compute SQuAD end logits from sequence hidden states.

    Args:
        config (:class:`~transformers.PretrainedConfig`):
            The config used by the model, will be used to grab the :obj:`hidden_size` of the model and the
            :obj:`layer_norm_eps` to use.
    """

    def __init__(self, config: PretrainedConfig):
        super().__init__()
        self.dense_0 = nn.Linear(config.hidden_size * 2, config.hidden_size)
        self.activation = nn.Tanh()
        self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
        self.dense_1 = nn.Linear(config.hidden_size, 1)

    def forward(
        self,
        hidden_states: torch.FloatTensor,
        start_states: Optional[torch.FloatTensor] = None,
        start_positions: Optional[torch.LongTensor] = None,
        p_mask: Optional[torch.FloatTensor] = None,
    ) -> torch.FloatTensor:
        """
        Args:
            hidden_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, seq_len, hidden_size)`):
                The final hidden states of the model.
            start_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, seq_len, hidden_size)`, `optional`):
                The hidden states of the first tokens for the labeled span.
            start_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
                The position of the first token for the labeled span.
            p_mask (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, seq_len)`, `optional`):
                Mask for tokens at invalid position, such as query and special symbols (PAD, SEP, CLS). 1.0 means token
                should be masked.

        .. note::

            One of ``start_states`` or ``start_positions`` should be not obj:`None`. If both are set,
            ``start_positions`` overrides ``start_states``.

        Returns:
            :obj:`torch.FloatTensor`: The end logits for SQuAD.
        """
        assert (
            start_states is not None or start_positions is not None
        ), "One of start_states, start_positions should be not None"
        if start_positions is not None:
            slen, hsz = hidden_states.shape[-2:]
            start_positions = start_positions[:, None, None].expand(-1, -1, hsz)  # shape (bsz, 1, hsz)
            start_states = hidden_states.gather(-2, start_positions)  # shape (bsz, 1, hsz)
            start_states = start_states.expand(-1, slen, -1)  # shape (bsz, slen, hsz)

        x = self.dense_0(torch.cat([hidden_states, start_states], dim=-1))
        x = self.activation(x)
        x = self.LayerNorm(x)
        x = self.dense_1(x).squeeze(-1)

        if p_mask is not None:
            if get_parameter_dtype(self) == torch.float16:
                x = x * (1 - p_mask) - 65500 * p_mask
            else:
                x = x * (1 - p_mask) - 1e30 * p_mask

        return x


class PoolerAnswerClass(nn.Module):
    """
    Compute SQuAD 2.0 answer class from classification and start tokens hidden states.

    Args:
        config (:class:`~transformers.PretrainedConfig`):
            The config used by the model, will be used to grab the :obj:`hidden_size` of the model.
    """

    def __init__(self, config):
        super().__init__()
        self.dense_0 = nn.Linear(config.hidden_size * 2, config.hidden_size)
        self.activation = nn.Tanh()
        self.dense_1 = nn.Linear(config.hidden_size, 1, bias=False)

    def forward(
        self,
        hidden_states: torch.FloatTensor,
        start_states: Optional[torch.FloatTensor] = None,
        start_positions: Optional[torch.LongTensor] = None,
        cls_index: Optional[torch.LongTensor] = None,
    ) -> torch.FloatTensor:
        """
        Args:
            hidden_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, seq_len, hidden_size)`):
                The final hidden states of the model.
            start_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, seq_len, hidden_size)`, `optional`):
                The hidden states of the first tokens for the labeled span.
            start_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
                The position of the first token for the labeled span.
            cls_index (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
                Position of the CLS token for each sentence in the batch. If :obj:`None`, takes the last token.

        .. note::

            One of ``start_states`` or ``start_positions`` should be not obj:`None`. If both are set,
            ``start_positions`` overrides ``start_states``.

        Returns:
            :obj:`torch.FloatTensor`: The SQuAD 2.0 answer class.
        """
        # No dependency on end_feature so that we can obtain one single `cls_logits` for each sample.
        hsz = hidden_states.shape[-1]
        assert (
            start_states is not None or start_positions is not None
        ), "One of start_states, start_positions should be not None"
        if start_positions is not None:
            start_positions = start_positions[:, None, None].expand(-1, -1, hsz)  # shape (bsz, 1, hsz)
            start_states = hidden_states.gather(-2, start_positions).squeeze(-2)  # shape (bsz, hsz)

        if cls_index is not None:
            cls_index = cls_index[:, None, None].expand(-1, -1, hsz)  # shape (bsz, 1, hsz)
            cls_token_state = hidden_states.gather(-2, cls_index).squeeze(-2)  # shape (bsz, hsz)
        else:
            cls_token_state = hidden_states[:, -1, :]  # shape (bsz, hsz)

        x = self.dense_0(torch.cat([start_states, cls_token_state], dim=-1))
        x = self.activation(x)
        x = self.dense_1(x).squeeze(-1)

        return x


@dataclass
class SquadHeadOutput(ModelOutput):
    """
    Base class for outputs of question answering models using a :class:`~transformers.modeling_utils.SQuADHead`.

    Args:
        loss (:obj:`torch.FloatTensor` of shape :obj:`(1,)`, `optional`, returned if both :obj:`start_positions` and :obj:`end_positions` are provided):
            Classification loss as the sum of start token, end token (and is_impossible if provided) classification
            losses.
        start_top_log_probs (``torch.FloatTensor`` of shape ``(batch_size, config.start_n_top)``, `optional`, returned if ``start_positions`` or ``end_positions`` is not provided):
            Log probabilities for the top config.start_n_top start token possibilities (beam-search).
        start_top_index (``torch.LongTensor`` of shape ``(batch_size, config.start_n_top)``, `optional`, returned if ``start_positions`` or ``end_positions`` is not provided):
            Indices for the top config.start_n_top start token possibilities (beam-search).
        end_top_log_probs (``torch.FloatTensor`` of shape ``(batch_size, config.start_n_top * config.end_n_top)``, `optional`, returned if ``start_positions`` or ``end_positions`` is not provided):
            Log probabilities for the top ``config.start_n_top * config.end_n_top`` end token possibilities
            (beam-search).
        end_top_index (``torch.LongTensor`` of shape ``(batch_size, config.start_n_top * config.end_n_top)``, `optional`, returned if ``start_positions`` or ``end_positions`` is not provided):
            Indices for the top ``config.start_n_top * config.end_n_top`` end token possibilities (beam-search).
        cls_logits (``torch.FloatTensor`` of shape ``(batch_size,)``, `optional`, returned if ``start_positions`` or ``end_positions`` is not provided):
            Log probabilities for the ``is_impossible`` label of the answers.

    """

    loss: Optional[torch.FloatTensor] = None
    start_top_log_probs: Optional[torch.FloatTensor] = None
    start_top_index: Optional[torch.LongTensor] = None
    end_top_log_probs: Optional[torch.FloatTensor] = None
    end_top_index: Optional[torch.LongTensor] = None
    cls_logits: Optional[torch.FloatTensor] = None


class SQuADHead(nn.Module):
    r"""
    A SQuAD head inspired by XLNet.

    Args:
        config (:class:`~transformers.PretrainedConfig`):
            The config used by the model, will be used to grab the :obj:`hidden_size` of the model and the
            :obj:`layer_norm_eps` to use.
    """

    def __init__(self, config):
        super().__init__()
        self.start_n_top = config.start_n_top
        self.end_n_top = config.end_n_top

        self.start_logits = PoolerStartLogits(config)
        self.end_logits = PoolerEndLogits(config)
        self.answer_class = PoolerAnswerClass(config)

    @replace_return_docstrings(output_type=SquadHeadOutput, config_class=PretrainedConfig)
    def forward(
        self,
        hidden_states: torch.FloatTensor,
        start_positions: Optional[torch.LongTensor] = None,
        end_positions: Optional[torch.LongTensor] = None,
        cls_index: Optional[torch.LongTensor] = None,
        is_impossible: Optional[torch.LongTensor] = None,
        p_mask: Optional[torch.FloatTensor] = None,
        return_dict: bool = False,
    ) -> Union[SquadHeadOutput, Tuple[torch.FloatTensor]]:
        """
        Args:
            hidden_states (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, seq_len, hidden_size)`):
                Final hidden states of the model on the sequence tokens.
            start_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
                Positions of the first token for the labeled span.
            end_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
                Positions of the last token for the labeled span.
            cls_index (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
                Position of the CLS token for each sentence in the batch. If :obj:`None`, takes the last token.
            is_impossible (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`):
                Whether the question has a possible answer in the paragraph or not.
            p_mask (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, seq_len)`, `optional`):
                Mask for tokens at invalid position, such as query and special symbols (PAD, SEP, CLS). 1.0 means token
                should be masked.
            return_dict (:obj:`bool`, `optional`, defaults to :obj:`False`):
                Whether or not to return a :class:`~transformers.file_utils.ModelOutput` instead of a plain tuple.

        Returns:
        """
        start_logits = self.start_logits(hidden_states, p_mask=p_mask)

        if start_positions is not None and end_positions is not None:
            # If we are on multi-GPU, let's remove the dimension added by batch splitting
            for x in (start_positions, end_positions, cls_index, is_impossible):
                if x is not None and x.dim() > 1:
                    x.squeeze_(-1)

            # during training, compute the end logits based on the ground truth of the start position
            end_logits = self.end_logits(hidden_states, start_positions=start_positions, p_mask=p_mask)

            loss_fct = CrossEntropyLoss()
            start_loss = loss_fct(start_logits, start_positions)
            end_loss = loss_fct(end_logits, end_positions)
            total_loss = (start_loss + end_loss) / 2

            if cls_index is not None and is_impossible is not None:
                # Predict answerability from the representation of CLS and START
                cls_logits = self.answer_class(hidden_states, start_positions=start_positions, cls_index=cls_index)
                loss_fct_cls = nn.BCEWithLogitsLoss()
                cls_loss = loss_fct_cls(cls_logits, is_impossible)

                # note(zhiliny): by default multiply the loss by 0.5 so that the scale is comparable to start_loss and end_loss
                total_loss += cls_loss * 0.5

            return SquadHeadOutput(loss=total_loss) if return_dict else (total_loss,)

        else:
            # during inference, compute the end logits based on beam search
            bsz, slen, hsz = hidden_states.size()
            start_log_probs = nn.functional.softmax(start_logits, dim=-1)  # shape (bsz, slen)

            start_top_log_probs, start_top_index = torch.topk(
                start_log_probs, self.start_n_top, dim=-1
            )  # shape (bsz, start_n_top)
            start_top_index_exp = start_top_index.unsqueeze(-1).expand(-1, -1, hsz)  # shape (bsz, start_n_top, hsz)
            start_states = torch.gather(hidden_states, -2, start_top_index_exp)  # shape (bsz, start_n_top, hsz)
            start_states = start_states.unsqueeze(1).expand(-1, slen, -1, -1)  # shape (bsz, slen, start_n_top, hsz)

            hidden_states_expanded = hidden_states.unsqueeze(2).expand_as(
                start_states
            )  # shape (bsz, slen, start_n_top, hsz)
            p_mask = p_mask.unsqueeze(-1) if p_mask is not None else None
            end_logits = self.end_logits(hidden_states_expanded, start_states=start_states, p_mask=p_mask)
            end_log_probs = nn.functional.softmax(end_logits, dim=1)  # shape (bsz, slen, start_n_top)

            end_top_log_probs, end_top_index = torch.topk(
                end_log_probs, self.end_n_top, dim=1
            )  # shape (bsz, end_n_top, start_n_top)
            end_top_log_probs = end_top_log_probs.view(-1, self.start_n_top * self.end_n_top)
            end_top_index = end_top_index.view(-1, self.start_n_top * self.end_n_top)

            start_states = torch.einsum("blh,bl->bh", hidden_states, start_log_probs)
            cls_logits = self.answer_class(hidden_states, start_states=start_states, cls_index=cls_index)

            if not return_dict:
                return (start_top_log_probs, start_top_index, end_top_log_probs, end_top_index, cls_logits)
            else:
                return SquadHeadOutput(
                    start_top_log_probs=start_top_log_probs,
                    start_top_index=start_top_index,
                    end_top_log_probs=end_top_log_probs,
                    end_top_index=end_top_index,
                    cls_logits=cls_logits,
                )


class SequenceSummary(nn.Module):
    r"""
    Compute a single vector summary of a sequence hidden states.

    Args:
        config (:class:`~transformers.PretrainedConfig`):
            The config used by the model. Relevant arguments in the config class of the model are (refer to the actual
            config class of your model for the default values it uses):

            - **summary_type** (:obj:`str`) -- The method to use to make this summary. Accepted values are:

                - :obj:`"last"` -- Take the last token hidden state (like XLNet)
                - :obj:`"first"` -- Take the first token hidden state (like Bert)
                - :obj:`"mean"` -- Take the mean of all tokens hidden states
                - :obj:`"cls_index"` -- Supply a Tensor of classification token position (GPT/GPT-2)
                - :obj:`"attn"` -- Not implemented now, use multi-head attention

            - **summary_use_proj** (:obj:`bool`) -- Add a projection after the vector extraction.
            - **summary_proj_to_labels** (:obj:`bool`) -- If :obj:`True`, the projection outputs to
              :obj:`config.num_labels` classes (otherwise to :obj:`config.hidden_size`).
            - **summary_activation** (:obj:`Optional[str]`) -- Set to :obj:`"tanh"` to add a tanh activation to the
              output, another string or :obj:`None` will add no activation.
            - **summary_first_dropout** (:obj:`float`) -- Optional dropout probability before the projection and
              activation.
            - **summary_last_dropout** (:obj:`float`)-- Optional dropout probability after the projection and
              activation.
    """

    def __init__(self, config: PretrainedConfig):
        super().__init__()

        self.summary_type = getattr(config, "summary_type", "last")
        if self.summary_type == "attn":
            # We should use a standard multi-head attention module with absolute positional embedding for that.
            # Cf. https://github.com/zihangdai/xlnet/blob/master/modeling.py#L253-L276
            # We can probably just use the multi-head attention module of PyTorch >=1.1.0
            raise NotImplementedError

        self.summary = Identity()
        if hasattr(config, "summary_use_proj") and config.summary_use_proj:
            if hasattr(config, "summary_proj_to_labels") and config.summary_proj_to_labels and config.num_labels > 0:
                num_classes = config.num_labels
            else:
                num_classes = config.hidden_size
            self.summary = nn.Linear(config.hidden_size, num_classes)

        activation_string = getattr(config, "summary_activation", None)
        self.activation: Callable = get_activation(activation_string) if activation_string else Identity()

        self.first_dropout = Identity()
        if hasattr(config, "summary_first_dropout") and config.summary_first_dropout > 0:
            self.first_dropout = nn.Dropout(config.summary_first_dropout)

        self.last_dropout = Identity()
        if hasattr(config, "summary_last_dropout") and config.summary_last_dropout > 0:
            self.last_dropout = nn.Dropout(config.summary_last_dropout)

    def forward(
        self, hidden_states: torch.FloatTensor, cls_index: Optional[torch.LongTensor] = None
    ) -> torch.FloatTensor:
        """
        Compute a single vector summary of a sequence hidden states.

        Args:
            hidden_states (:obj:`torch.FloatTensor` of shape :obj:`[batch_size, seq_len, hidden_size]`):
                The hidden states of the last layer.
            cls_index (:obj:`torch.LongTensor` of shape :obj:`[batch_size]` or :obj:`[batch_size, ...]` where ... are optional leading dimensions of :obj:`hidden_states`, `optional`):
                Used if :obj:`summary_type == "cls_index"` and takes the last token of the sequence as classification
                token.

        Returns:
            :obj:`torch.FloatTensor`: The summary of the sequence hidden states.
        """
        if self.summary_type == "last":
            output = hidden_states[:, -1]
        elif self.summary_type == "first":
            output = hidden_states[:, 0]
        elif self.summary_type == "mean":
            output = hidden_states.mean(dim=1)
        elif self.summary_type == "cls_index":
            if cls_index is None:
                cls_index = torch.full_like(
                    hidden_states[..., :1, :],
                    hidden_states.shape[-2] - 1,
                    dtype=torch.long,
                )
            else:
                cls_index = cls_index.unsqueeze(-1).unsqueeze(-1)
                cls_index = cls_index.expand((-1,) * (cls_index.dim() - 1) + (hidden_states.size(-1),))
            # shape of cls_index: (bsz, XX, 1, hidden_size) where XX are optional leading dim of hidden_states
            output = hidden_states.gather(-2, cls_index).squeeze(-2)  # shape (bsz, XX, hidden_size)
        elif self.summary_type == "attn":
            raise NotImplementedError

        output = self.first_dropout(output)
        output = self.summary(output)
        output = self.activation(output)
        output = self.last_dropout(output)

        return output


def unwrap_model(model: nn.Module) -> nn.Module:
    """
    Recursively unwraps a model from potential containers (as used in distributed training).

    Args:
        model (:obj:`torch.nn.Module`): The model to unwrap.
    """
    # since there could be multiple levels of wrapping, unwrap recursively
    if hasattr(model, "module"):
        return unwrap_model(model.module)
    else:
        return model


def prune_linear_layer(layer: nn.Linear, index: torch.LongTensor, dim: int = 0) -> nn.Linear:
    """
    Prune a linear layer to keep only entries in index.

    Used to remove heads.

    Args:
        layer (:obj:`torch.nn.Linear`): The layer to prune.
        index (:obj:`torch.LongTensor`): The indices to keep in the layer.
        dim (:obj:`int`, `optional`, defaults to 0): The dimension on which to keep the indices.

    Returns:
        :obj:`torch.nn.Linear`: The pruned layer as a new layer with :obj:`requires_grad=True`.
    """
    index = index.to(layer.weight.device)
    W = layer.weight.index_select(dim, index).clone().detach()
    if layer.bias is not None:
        if dim == 1:
            b = layer.bias.clone().detach()
        else:
            b = layer.bias[index].clone().detach()
    new_size = list(layer.weight.size())
    new_size[dim] = len(index)
    new_layer = nn.Linear(new_size[1], new_size[0], bias=layer.bias is not None).to(layer.weight.device)
    new_layer.weight.requires_grad = False
    new_layer.weight.copy_(W.contiguous())
    new_layer.weight.requires_grad = True
    if layer.bias is not None:
        new_layer.bias.requires_grad = False
        new_layer.bias.copy_(b.contiguous())
        new_layer.bias.requires_grad = True
    return new_layer


def prune_conv1d_layer(layer: Conv1D, index: torch.LongTensor, dim: int = 1) -> Conv1D:
    """
    Prune a Conv1D layer to keep only entries in index. A Conv1D work as a Linear layer (see e.g. BERT) but the weights
    are transposed.

    Used to remove heads.

    Args:
        layer (:class:`~transformers.modeling_utils.Conv1D`): The layer to prune.
        index (:obj:`torch.LongTensor`): The indices to keep in the layer.
        dim (:obj:`int`, `optional`, defaults to 1): The dimension on which to keep the indices.

    Returns:
        :class:`~transformers.modeling_utils.Conv1D`: The pruned layer as a new layer with :obj:`requires_grad=True`.
    """
    index = index.to(layer.weight.device)
    W = layer.weight.index_select(dim, index).clone().detach()
    if dim == 0:
        b = layer.bias.clone().detach()
    else:
        b = layer.bias[index].clone().detach()
    new_size = list(layer.weight.size())
    new_size[dim] = len(index)
    new_layer = Conv1D(new_size[1], new_size[0]).to(layer.weight.device)
    new_layer.weight.requires_grad = False
    new_layer.weight.copy_(W.contiguous())
    new_layer.weight.requires_grad = True
    new_layer.bias.requires_grad = False
    new_layer.bias.copy_(b.contiguous())
    new_layer.bias.requires_grad = True
    return new_layer


def prune_layer(
    layer: Union[nn.Linear, Conv1D], index: torch.LongTensor, dim: Optional[int] = None
) -> Union[nn.Linear, Conv1D]:
    """
    Prune a Conv1D or linear layer to keep only entries in index.

    Used to remove heads.

    Args:
        layer (:obj:`Union[torch.nn.Linear, Conv1D]`): The layer to prune.
        index (:obj:`torch.LongTensor`): The indices to keep in the layer.
        dim (:obj:`int`, `optional`): The dimension on which to keep the indices.

    Returns:
        :obj:`torch.nn.Linear` or :class:`~transformers.modeling_utils.Conv1D`: The pruned layer as a new layer with
        :obj:`requires_grad=True`.
    """
    if isinstance(layer, nn.Linear):
        return prune_linear_layer(layer, index, dim=0 if dim is None else dim)
    elif isinstance(layer, Conv1D):
        return prune_conv1d_layer(layer, index, dim=1 if dim is None else dim)
    else:
        raise ValueError(f"Can't prune layer of class {layer.__class__}")


def apply_chunking_to_forward(
    forward_fn: Callable[..., torch.Tensor], chunk_size: int, chunk_dim: int, *input_tensors
) -> torch.Tensor:
    """
    This function chunks the :obj:`input_tensors` into smaller input tensor parts of size :obj:`chunk_size` over the
    dimension :obj:`chunk_dim`. It then applies a layer :obj:`forward_fn` to each chunk independently to save memory.

    If the :obj:`forward_fn` is independent across the :obj:`chunk_dim` this function will yield the same result as
    directly applying :obj:`forward_fn` to :obj:`input_tensors`.

    Args:
        forward_fn (:obj:`Callable[..., torch.Tensor]`):
            The forward function of the model.
        chunk_size (:obj:`int`):
            The chunk size of a chunked tensor: :obj:`num_chunks = len(input_tensors[0]) / chunk_size`.
        chunk_dim (:obj:`int`):
            The dimension over which the :obj:`input_tensors` should be chunked.
        input_tensors (:obj:`Tuple[torch.Tensor]`):
            The input tensors of ``forward_fn`` which will be chunked

    Returns:
        :obj:`torch.Tensor`: A tensor with the same shape as the :obj:`forward_fn` would have given if applied`.


    Examples::

        # rename the usual forward() fn to forward_chunk()
        def forward_chunk(self, hidden_states):
            hidden_states = self.decoder(hidden_states)
            return hidden_states

        # implement a chunked forward function
        def forward(self, hidden_states):
            return apply_chunking_to_forward(self.forward_chunk, self.chunk_size_lm_head, self.seq_len_dim, hidden_states)
    """

    assert len(input_tensors) > 0, f"{input_tensors} has to be a tuple/list of tensors"
    tensor_shape = input_tensors[0].shape[chunk_dim]
    assert all(
        input_tensor.shape[chunk_dim] == tensor_shape for input_tensor in input_tensors
    ), "All input tenors have to be of the same shape"

    # inspect.signature exist since python 3.5 and is a python method -> no problem with backward compatibility
    num_args_in_forward_chunk_fn = len(inspect.signature(forward_fn).parameters)
    if num_args_in_forward_chunk_fn != len(input_tensors):
        raise ValueError(
            f"forward_chunk_fn expects {num_args_in_forward_chunk_fn} arguments, but only {len(input_tensors)} input "
            "tensors are given"
        )

    if chunk_size > 0:
        if input_tensors[0].shape[chunk_dim] % chunk_size != 0:
            raise ValueError(
                f"The dimension to be chunked {input_tensors[0].shape[chunk_dim]} has to be a multiple of the chunk "
                f"size {chunk_size}"
            )

        num_chunks = input_tensors[0].shape[chunk_dim] // chunk_size

        # chunk input tensor into tuples
        input_tensors_chunks = tuple(input_tensor.chunk(num_chunks, dim=chunk_dim) for input_tensor in input_tensors)
        # apply forward fn to every tuple
        output_chunks = tuple(forward_fn(*input_tensors_chunk) for input_tensors_chunk in zip(*input_tensors_chunks))
        # concatenate output at same dimension
        return torch.cat(output_chunks, dim=chunk_dim)

    return forward_fn(*input_tensors)