update model

config.json

@@ -0,0 +1,25 @@
+{
+  "_name_or_path": "temp/dummy/bert/BertModel",
+  "architectures": [
+    "BertModel"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 32,
+  "initializer_range": 0.02,
+  "intermediate_size": 37,
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 4,
+  "num_hidden_layers": 5,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.25.0.dev0",
+  "type_vocab_size": 16,
+  "use_cache": true,
+  "vocab_size": 1124
+}

create_model.py

@@ -0,0 +1,10 @@
+# from transformers import AutoConfig
+
+# from modeling.modeling_bert import BertCustomLMHeadModel
+
+# cfg = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-BertModel")
+
+# BertCustomLMHeadModel.register_for_auto_class("AutoModelForSequenceClassification")
+
+# model = BertCustomLMHeadModel(cfg)
+# model.save_pretrained("/home/Jingya/hf_internship/tiny-testing-gpt2-remote-code")

modeling_bert.py

@@ -0,0 +1,1894 @@
+# coding=utf-8
+# Copyright 2018 The Google AI Language Team 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.
+"""PyTorch BERT model."""
+
+
+import math
+import os
+import warnings
+from dataclasses import dataclass
+from typing import List, Optional, Tuple, Union
+
+import torch
+import torch.utils.checkpoint
+from torch import nn
+from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
+
+from ...activations import ACT2FN
+from ...modeling_outputs import (
+    BaseModelOutputWithPastAndCrossAttentions,
+    BaseModelOutputWithPoolingAndCrossAttentions,
+    CausalLMOutputWithCrossAttentions,
+    MaskedLMOutput,
+    MultipleChoiceModelOutput,
+    NextSentencePredictorOutput,
+    QuestionAnsweringModelOutput,
+    SequenceClassifierOutput,
+    TokenClassifierOutput,
+)
+from ...modeling_utils import PreTrainedModel
+from ...pytorch_utils import apply_chunking_to_forward, find_pruneable_heads_and_indices, prune_linear_layer
+from ...utils import (
+    ModelOutput,
+    add_code_sample_docstrings,
+    add_start_docstrings,
+    add_start_docstrings_to_model_forward,
+    logging,
+    replace_return_docstrings,
+)
+from .configuration_bert import BertConfig
+
+
+logger = logging.get_logger(__name__)
+
+_CHECKPOINT_FOR_DOC = "bert-base-uncased"
+_CONFIG_FOR_DOC = "BertConfig"
+
+# TokenClassification docstring
+_CHECKPOINT_FOR_TOKEN_CLASSIFICATION = "dbmdz/bert-large-cased-finetuned-conll03-english"
+_TOKEN_CLASS_EXPECTED_OUTPUT = (
+    "['O', 'I-ORG', 'I-ORG', 'I-ORG', 'O', 'O', 'O', 'O', 'O', 'I-LOC', 'O', 'I-LOC', 'I-LOC'] "
+)
+_TOKEN_CLASS_EXPECTED_LOSS = 0.01
+
+# QuestionAnswering docstring
+_CHECKPOINT_FOR_QA = "deepset/bert-base-cased-squad2"
+_QA_EXPECTED_OUTPUT = "'a nice puppet'"
+_QA_EXPECTED_LOSS = 7.41
+_QA_TARGET_START_INDEX = 14
+_QA_TARGET_END_INDEX = 15
+
+# SequenceClassification docstring
+_CHECKPOINT_FOR_SEQUENCE_CLASSIFICATION = "textattack/bert-base-uncased-yelp-polarity"
+_SEQ_CLASS_EXPECTED_OUTPUT = "'LABEL_1'"
+_SEQ_CLASS_EXPECTED_LOSS = 0.01
+
+
+BERT_PRETRAINED_MODEL_ARCHIVE_LIST = [
+    "bert-base-uncased",
+    "bert-large-uncased",
+    "bert-base-cased",
+    "bert-large-cased",
+    "bert-base-multilingual-uncased",
+    "bert-base-multilingual-cased",
+    "bert-base-chinese",
+    "bert-base-german-cased",
+    "bert-large-uncased-whole-word-masking",
+    "bert-large-cased-whole-word-masking",
+    "bert-large-uncased-whole-word-masking-finetuned-squad",
+    "bert-large-cased-whole-word-masking-finetuned-squad",
+    "bert-base-cased-finetuned-mrpc",
+    "bert-base-german-dbmdz-cased",
+    "bert-base-german-dbmdz-uncased",
+    "cl-tohoku/bert-base-japanese",
+    "cl-tohoku/bert-base-japanese-whole-word-masking",
+    "cl-tohoku/bert-base-japanese-char",
+    "cl-tohoku/bert-base-japanese-char-whole-word-masking",
+    "TurkuNLP/bert-base-finnish-cased-v1",
+    "TurkuNLP/bert-base-finnish-uncased-v1",
+    "wietsedv/bert-base-dutch-cased",
+    # See all BERT models at https://huggingface.co/models?filter=bert
+]
+
+
+def load_tf_weights_in_bert(model, config, tf_checkpoint_path):
+    """Load tf checkpoints in a pytorch model."""
+    try:
+        import re
+
+        import numpy as np
+        import tensorflow as tf
+    except ImportError:
+        logger.error(
+            "Loading a TensorFlow model in PyTorch, requires TensorFlow to be installed. Please see "
+            "https://www.tensorflow.org/install/ for installation instructions."
+        )
+        raise
+    tf_path = os.path.abspath(tf_checkpoint_path)
+    logger.info(f"Converting TensorFlow checkpoint from {tf_path}")
+    # Load weights from TF model
+    init_vars = tf.train.list_variables(tf_path)
+    names = []
+    arrays = []
+    for name, shape in init_vars:
+        logger.info(f"Loading TF weight {name} with shape {shape}")
+        array = tf.train.load_variable(tf_path, name)
+        names.append(name)
+        arrays.append(array)
+
+    for name, array in zip(names, arrays):
+        name = name.split("/")
+        # adam_v and adam_m are variables used in AdamWeightDecayOptimizer to calculated m and v
+        # which are not required for using pretrained model
+        if any(
+            n in ["adam_v", "adam_m", "AdamWeightDecayOptimizer", "AdamWeightDecayOptimizer_1", "global_step"]
+            for n in name
+        ):
+            logger.info(f"Skipping {'/'.join(name)}")
+            continue
+        pointer = model
+        for m_name in name:
+            if re.fullmatch(r"[A-Za-z]+_\d+", m_name):
+                scope_names = re.split(r"_(\d+)", m_name)
+            else:
+                scope_names = [m_name]
+            if scope_names[0] == "kernel" or scope_names[0] == "gamma":
+                pointer = getattr(pointer, "weight")
+            elif scope_names[0] == "output_bias" or scope_names[0] == "beta":
+                pointer = getattr(pointer, "bias")
+            elif scope_names[0] == "output_weights":
+                pointer = getattr(pointer, "weight")
+            elif scope_names[0] == "squad":
+                pointer = getattr(pointer, "classifier")
+            else:
+                try:
+                    pointer = getattr(pointer, scope_names[0])
+                except AttributeError:
+                    logger.info(f"Skipping {'/'.join(name)}")
+                    continue
+            if len(scope_names) >= 2:
+                num = int(scope_names[1])
+                pointer = pointer[num]
+        if m_name[-11:] == "_embeddings":
+            pointer = getattr(pointer, "weight")
+        elif m_name == "kernel":
+            array = np.transpose(array)
+        try:
+            if pointer.shape != array.shape:
+                raise ValueError(f"Pointer shape {pointer.shape} and array shape {array.shape} mismatched")
+        except AssertionError as e:
+            e.args += (pointer.shape, array.shape)
+            raise
+        logger.info(f"Initialize PyTorch weight {name}")
+        pointer.data = torch.from_numpy(array)
+    return model
+
+
+class BertEmbeddings(nn.Module):
+    """Construct the embeddings from word, position and token_type embeddings."""
+
+    def __init__(self, config):
+        super().__init__()
+        self.word_embeddings = nn.Embedding(config.vocab_size, config.hidden_size, padding_idx=config.pad_token_id)
+        self.position_embeddings = nn.Embedding(config.max_position_embeddings, config.hidden_size)
+        self.token_type_embeddings = nn.Embedding(config.type_vocab_size, config.hidden_size)
+
+        # self.LayerNorm is not snake-cased to stick with TensorFlow model variable name and be able to load
+        # any TensorFlow checkpoint file
+        self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
+        self.dropout = nn.Dropout(config.hidden_dropout_prob)
+        # position_ids (1, len position emb) is contiguous in memory and exported when serialized
+        self.position_embedding_type = getattr(config, "position_embedding_type", "absolute")
+        self.register_buffer("position_ids", torch.arange(config.max_position_embeddings).expand((1, -1)))
+        self.register_buffer(
+            "token_type_ids", torch.zeros(self.position_ids.size(), dtype=torch.long), persistent=False
+        )
+
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        token_type_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        past_key_values_length: int = 0,
+    ) -> torch.Tensor:
+        if input_ids is not None:
+            input_shape = input_ids.size()
+        else:
+            input_shape = inputs_embeds.size()[:-1]
+
+        seq_length = input_shape[1]
+
+        if position_ids is None:
+            position_ids = self.position_ids[:, past_key_values_length : seq_length + past_key_values_length]
+
+        # Setting the token_type_ids to the registered buffer in constructor where it is all zeros, which usually occurs
+        # when its auto-generated, registered buffer helps users when tracing the model without passing token_type_ids, solves
+        # issue #5664
+        if token_type_ids is None:
+            if hasattr(self, "token_type_ids"):
+                buffered_token_type_ids = self.token_type_ids[:, :seq_length]
+                buffered_token_type_ids_expanded = buffered_token_type_ids.expand(input_shape[0], seq_length)
+                token_type_ids = buffered_token_type_ids_expanded
+            else:
+                token_type_ids = torch.zeros(input_shape, dtype=torch.long, device=self.position_ids.device)
+
+        if inputs_embeds is None:
+            inputs_embeds = self.word_embeddings(input_ids)
+        token_type_embeddings = self.token_type_embeddings(token_type_ids)
+
+        embeddings = inputs_embeds + token_type_embeddings
+        if self.position_embedding_type == "absolute":
+            position_embeddings = self.position_embeddings(position_ids)
+            embeddings += position_embeddings
+        embeddings = self.LayerNorm(embeddings)
+        embeddings = self.dropout(embeddings)
+        return embeddings
+
+
+class BertSelfAttention(nn.Module):
+    def __init__(self, config, position_embedding_type=None):
+        super().__init__()
+        if config.hidden_size % config.num_attention_heads != 0 and not hasattr(config, "embedding_size"):
+            raise ValueError(
+                f"The hidden size ({config.hidden_size}) is not a multiple of the number of attention "
+                f"heads ({config.num_attention_heads})"
+            )
+
+        self.num_attention_heads = config.num_attention_heads
+        self.attention_head_size = int(config.hidden_size / config.num_attention_heads)
+        self.all_head_size = self.num_attention_heads * self.attention_head_size
+
+        self.query = nn.Linear(config.hidden_size, self.all_head_size)
+        self.key = nn.Linear(config.hidden_size, self.all_head_size)
+        self.value = nn.Linear(config.hidden_size, self.all_head_size)
+
+        self.dropout = nn.Dropout(config.attention_probs_dropout_prob)
+        self.position_embedding_type = position_embedding_type or getattr(
+            config, "position_embedding_type", "absolute"
+        )
+        if self.position_embedding_type == "relative_key" or self.position_embedding_type == "relative_key_query":
+            self.max_position_embeddings = config.max_position_embeddings
+            self.distance_embedding = nn.Embedding(2 * config.max_position_embeddings - 1, self.attention_head_size)
+
+        self.is_decoder = config.is_decoder
+
+    def transpose_for_scores(self, x: torch.Tensor) -> torch.Tensor:
+        new_x_shape = x.size()[:-1] + (self.num_attention_heads, self.attention_head_size)
+        x = x.view(new_x_shape)
+        return x.permute(0, 2, 1, 3)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.FloatTensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        past_key_value: Optional[Tuple[Tuple[torch.FloatTensor]]] = None,
+        output_attentions: Optional[bool] = False,
+    ) -> Tuple[torch.Tensor]:
+        mixed_query_layer = self.query(hidden_states)
+
+        # If this is instantiated as a cross-attention module, the keys
+        # and values come from an encoder; the attention mask needs to be
+        # such that the encoder's padding tokens are not attended to.
+        is_cross_attention = encoder_hidden_states is not None
+
+        if is_cross_attention and past_key_value is not None:
+            # reuse k,v, cross_attentions
+            key_layer = past_key_value[0]
+            value_layer = past_key_value[1]
+            attention_mask = encoder_attention_mask
+        elif is_cross_attention:
+            key_layer = self.transpose_for_scores(self.key(encoder_hidden_states))
+            value_layer = self.transpose_for_scores(self.value(encoder_hidden_states))
+            attention_mask = encoder_attention_mask
+        elif past_key_value is not None:
+            key_layer = self.transpose_for_scores(self.key(hidden_states))
+            value_layer = self.transpose_for_scores(self.value(hidden_states))
+            key_layer = torch.cat([past_key_value[0], key_layer], dim=2)
+            value_layer = torch.cat([past_key_value[1], value_layer], dim=2)
+        else:
+            key_layer = self.transpose_for_scores(self.key(hidden_states))
+            value_layer = self.transpose_for_scores(self.value(hidden_states))
+
+        query_layer = self.transpose_for_scores(mixed_query_layer)
+
+        use_cache = past_key_value is not None
+        if self.is_decoder:
+            # if cross_attention save Tuple(torch.Tensor, torch.Tensor) of all cross attention key/value_states.
+            # Further calls to cross_attention layer can then reuse all cross-attention
+            # key/value_states (first "if" case)
+            # if uni-directional self-attention (decoder) save Tuple(torch.Tensor, torch.Tensor) of
+            # all previous decoder key/value_states. Further calls to uni-directional self-attention
+            # can concat previous decoder key/value_states to current projected key/value_states (third "elif" case)
+            # if encoder bi-directional self-attention `past_key_value` is always `None`
+            past_key_value = (key_layer, value_layer)
+
+        # Take the dot product between "query" and "key" to get the raw attention scores.
+        attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))
+
+        if self.position_embedding_type == "relative_key" or self.position_embedding_type == "relative_key_query":
+            query_length, key_length = query_layer.shape[2], key_layer.shape[2]
+            if use_cache:
+                position_ids_l = torch.tensor(key_length - 1, dtype=torch.long, device=hidden_states.device).view(
+                    -1, 1
+                )
+            else:
+                position_ids_l = torch.arange(query_length, dtype=torch.long, device=hidden_states.device).view(-1, 1)
+            position_ids_r = torch.arange(key_length, dtype=torch.long, device=hidden_states.device).view(1, -1)
+            distance = position_ids_l - position_ids_r
+
+            positional_embedding = self.distance_embedding(distance + self.max_position_embeddings - 1)
+            positional_embedding = positional_embedding.to(dtype=query_layer.dtype)  # fp16 compatibility
+
+            if self.position_embedding_type == "relative_key":
+                relative_position_scores = torch.einsum("bhld,lrd->bhlr", query_layer, positional_embedding)
+                attention_scores = attention_scores + relative_position_scores
+            elif self.position_embedding_type == "relative_key_query":
+                relative_position_scores_query = torch.einsum("bhld,lrd->bhlr", query_layer, positional_embedding)
+                relative_position_scores_key = torch.einsum("bhrd,lrd->bhlr", key_layer, positional_embedding)
+                attention_scores = attention_scores + relative_position_scores_query + relative_position_scores_key
+
+        attention_scores = attention_scores / math.sqrt(self.attention_head_size)
+        if attention_mask is not None:
+            # Apply the attention mask is (precomputed for all layers in BertModel forward() function)
+            attention_scores = attention_scores + attention_mask
+
+        # Normalize the attention scores to probabilities.
+        attention_probs = nn.functional.softmax(attention_scores, dim=-1)
+
+        # This is actually dropping out entire tokens to attend to, which might
+        # seem a bit unusual, but is taken from the original Transformer paper.
+        attention_probs = self.dropout(attention_probs)
+
+        # Mask heads if we want to
+        if head_mask is not None:
+            attention_probs = attention_probs * head_mask
+
+        context_layer = torch.matmul(attention_probs, value_layer)
+
+        context_layer = context_layer.permute(0, 2, 1, 3).contiguous()
+        new_context_layer_shape = context_layer.size()[:-2] + (self.all_head_size,)
+        context_layer = context_layer.view(new_context_layer_shape)
+
+        outputs = (context_layer, attention_probs) if output_attentions else (context_layer,)
+
+        if self.is_decoder:
+            outputs = outputs + (past_key_value,)
+        return outputs
+
+
+class BertSelfOutput(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
+        self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
+        self.dropout = nn.Dropout(config.hidden_dropout_prob)
+
+    def forward(self, hidden_states: torch.Tensor, input_tensor: torch.Tensor) -> torch.Tensor:
+        hidden_states = self.dense(hidden_states)
+        hidden_states = self.dropout(hidden_states)
+        hidden_states = self.LayerNorm(hidden_states + input_tensor)
+        return hidden_states
+
+
+class BertAttention(nn.Module):
+    def __init__(self, config, position_embedding_type=None):
+        super().__init__()
+        self.self = BertSelfAttention(config, position_embedding_type=position_embedding_type)
+        self.output = BertSelfOutput(config)
+        self.pruned_heads = set()
+
+    def prune_heads(self, heads):
+        if len(heads) == 0:
+            return
+        heads, index = find_pruneable_heads_and_indices(
+            heads, self.self.num_attention_heads, self.self.attention_head_size, self.pruned_heads
+        )
+
+        # Prune linear layers
+        self.self.query = prune_linear_layer(self.self.query, index)
+        self.self.key = prune_linear_layer(self.self.key, index)
+        self.self.value = prune_linear_layer(self.self.value, index)
+        self.output.dense = prune_linear_layer(self.output.dense, index, dim=1)
+
+        # Update hyper params and store pruned heads
+        self.self.num_attention_heads = self.self.num_attention_heads - len(heads)
+        self.self.all_head_size = self.self.attention_head_size * self.self.num_attention_heads
+        self.pruned_heads = self.pruned_heads.union(heads)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.FloatTensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        past_key_value: Optional[Tuple[Tuple[torch.FloatTensor]]] = None,
+        output_attentions: Optional[bool] = False,
+    ) -> Tuple[torch.Tensor]:
+        self_outputs = self.self(
+            hidden_states,
+            attention_mask,
+            head_mask,
+            encoder_hidden_states,
+            encoder_attention_mask,
+            past_key_value,
+            output_attentions,
+        )
+        attention_output = self.output(self_outputs[0], hidden_states)
+        outputs = (attention_output,) + self_outputs[1:]  # add attentions if we output them
+        return outputs
+
+
+class BertIntermediate(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.dense = nn.Linear(config.hidden_size, config.intermediate_size)
+        if isinstance(config.hidden_act, str):
+            self.intermediate_act_fn = ACT2FN[config.hidden_act]
+        else:
+            self.intermediate_act_fn = config.hidden_act
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        hidden_states = self.dense(hidden_states)
+        hidden_states = self.intermediate_act_fn(hidden_states)
+        return hidden_states
+
+
+class BertOutput(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.dense = nn.Linear(config.intermediate_size, config.hidden_size)
+        self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
+        self.dropout = nn.Dropout(config.hidden_dropout_prob)
+
+    def forward(self, hidden_states: torch.Tensor, input_tensor: torch.Tensor) -> torch.Tensor:
+        hidden_states = self.dense(hidden_states)
+        hidden_states = self.dropout(hidden_states)
+        hidden_states = self.LayerNorm(hidden_states + input_tensor)
+        return hidden_states
+
+
+class BertLayer(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.chunk_size_feed_forward = config.chunk_size_feed_forward
+        self.seq_len_dim = 1
+        self.attention = BertAttention(config)
+        self.is_decoder = config.is_decoder
+        self.add_cross_attention = config.add_cross_attention
+        if self.add_cross_attention:
+            if not self.is_decoder:
+                raise ValueError(f"{self} should be used as a decoder model if cross attention is added")
+            self.crossattention = BertAttention(config, position_embedding_type="absolute")
+        self.intermediate = BertIntermediate(config)
+        self.output = BertOutput(config)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.FloatTensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        past_key_value: Optional[Tuple[Tuple[torch.FloatTensor]]] = None,
+        output_attentions: Optional[bool] = False,
+    ) -> Tuple[torch.Tensor]:
+        # decoder uni-directional self-attention cached key/values tuple is at positions 1,2
+        self_attn_past_key_value = past_key_value[:2] if past_key_value is not None else None
+        self_attention_outputs = self.attention(
+            hidden_states,
+            attention_mask,
+            head_mask,
+            output_attentions=output_attentions,
+            past_key_value=self_attn_past_key_value,
+        )
+        attention_output = self_attention_outputs[0]
+
+        # if decoder, the last output is tuple of self-attn cache
+        if self.is_decoder:
+            outputs = self_attention_outputs[1:-1]
+            present_key_value = self_attention_outputs[-1]
+        else:
+            outputs = self_attention_outputs[1:]  # add self attentions if we output attention weights
+
+        cross_attn_present_key_value = None
+        if self.is_decoder and encoder_hidden_states is not None:
+            if not hasattr(self, "crossattention"):
+                raise ValueError(
+                    f"If `encoder_hidden_states` are passed, {self} has to be instantiated with cross-attention layers"
+                    " by setting `config.add_cross_attention=True`"
+                )
+
+            # cross_attn cached key/values tuple is at positions 3,4 of past_key_value tuple
+            cross_attn_past_key_value = past_key_value[-2:] if past_key_value is not None else None
+            cross_attention_outputs = self.crossattention(
+                attention_output,
+                attention_mask,
+                head_mask,
+                encoder_hidden_states,
+                encoder_attention_mask,
+                cross_attn_past_key_value,
+                output_attentions,
+            )
+            attention_output = cross_attention_outputs[0]
+            outputs = outputs + cross_attention_outputs[1:-1]  # add cross attentions if we output attention weights
+
+            # add cross-attn cache to positions 3,4 of present_key_value tuple
+            cross_attn_present_key_value = cross_attention_outputs[-1]
+            present_key_value = present_key_value + cross_attn_present_key_value
+
+        layer_output = apply_chunking_to_forward(
+            self.feed_forward_chunk, self.chunk_size_feed_forward, self.seq_len_dim, attention_output
+        )
+        outputs = (layer_output,) + outputs
+
+        # if decoder, return the attn key/values as the last output
+        if self.is_decoder:
+            outputs = outputs + (present_key_value,)
+
+        return outputs
+
+    def feed_forward_chunk(self, attention_output):
+        intermediate_output = self.intermediate(attention_output)
+        layer_output = self.output(intermediate_output, attention_output)
+        return layer_output
+
+
+class BertEncoder(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.config = config
+        self.layer = nn.ModuleList([BertLayer(config) for _ in range(config.num_hidden_layers)])
+        self.gradient_checkpointing = False
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.FloatTensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.FloatTensor]]] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = False,
+        output_hidden_states: Optional[bool] = False,
+        return_dict: Optional[bool] = True,
+    ) -> Union[Tuple[torch.Tensor], BaseModelOutputWithPastAndCrossAttentions]:
+        all_hidden_states = () if output_hidden_states else None
+        all_self_attentions = () if output_attentions else None
+        all_cross_attentions = () if output_attentions and self.config.add_cross_attention else None
+
+        if self.gradient_checkpointing and self.training:
+            if use_cache:
+                logger.warning_once(
+                    "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
+                )
+                use_cache = False
+
+        next_decoder_cache = () if use_cache else None
+        for i, layer_module in enumerate(self.layer):
+            if output_hidden_states:
+                all_hidden_states = all_hidden_states + (hidden_states,)
+
+            layer_head_mask = head_mask[i] if head_mask is not None else None
+            past_key_value = past_key_values[i] if past_key_values is not None else None
+
+            if self.gradient_checkpointing and self.training:
+
+                def create_custom_forward(module):
+                    def custom_forward(*inputs):
+                        return module(*inputs, past_key_value, output_attentions)
+
+                    return custom_forward
+
+                layer_outputs = torch.utils.checkpoint.checkpoint(
+                    create_custom_forward(layer_module),
+                    hidden_states,
+                    attention_mask,
+                    layer_head_mask,
+                    encoder_hidden_states,
+                    encoder_attention_mask,
+                )
+            else:
+                layer_outputs = layer_module(
+                    hidden_states,
+                    attention_mask,
+                    layer_head_mask,
+                    encoder_hidden_states,
+                    encoder_attention_mask,
+                    past_key_value,
+                    output_attentions,
+                )
+
+            hidden_states = layer_outputs[0]
+            if use_cache:
+                next_decoder_cache += (layer_outputs[-1],)
+            if output_attentions:
+                all_self_attentions = all_self_attentions + (layer_outputs[1],)
+                if self.config.add_cross_attention:
+                    all_cross_attentions = all_cross_attentions + (layer_outputs[2],)
+
+        if output_hidden_states:
+            all_hidden_states = all_hidden_states + (hidden_states,)
+
+        if not return_dict:
+            return tuple(
+                v
+                for v in [
+                    hidden_states,
+                    next_decoder_cache,
+                    all_hidden_states,
+                    all_self_attentions,
+                    all_cross_attentions,
+                ]
+                if v is not None
+            )
+        return BaseModelOutputWithPastAndCrossAttentions(
+            last_hidden_state=hidden_states,
+            past_key_values=next_decoder_cache,
+            hidden_states=all_hidden_states,
+            attentions=all_self_attentions,
+            cross_attentions=all_cross_attentions,
+        )
+
+
+class BertPooler(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
+        self.activation = nn.Tanh()
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        # We "pool" the model by simply taking the hidden state corresponding
+        # to the first token.
+        first_token_tensor = hidden_states[:, 0]
+        pooled_output = self.dense(first_token_tensor)
+        pooled_output = self.activation(pooled_output)
+        return pooled_output
+
+
+class BertPredictionHeadTransform(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
+        if isinstance(config.hidden_act, str):
+            self.transform_act_fn = ACT2FN[config.hidden_act]
+        else:
+            self.transform_act_fn = config.hidden_act
+        self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        hidden_states = self.dense(hidden_states)
+        hidden_states = self.transform_act_fn(hidden_states)
+        hidden_states = self.LayerNorm(hidden_states)
+        return hidden_states
+
+
+class BertLMPredictionHead(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.transform = BertPredictionHeadTransform(config)
+
+        # The output weights are the same as the input embeddings, but there is
+        # an output-only bias for each token.
+        self.decoder = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
+
+        self.bias = nn.Parameter(torch.zeros(config.vocab_size))
+
+        # Need a link between the two variables so that the bias is correctly resized with `resize_token_embeddings`
+        self.decoder.bias = self.bias
+
+    def forward(self, hidden_states):
+        hidden_states = self.transform(hidden_states)
+        hidden_states = self.decoder(hidden_states)
+        return hidden_states
+
+
+class BertOnlyMLMHead(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.predictions = BertLMPredictionHead(config)
+
+    def forward(self, sequence_output: torch.Tensor) -> torch.Tensor:
+        prediction_scores = self.predictions(sequence_output)
+        return prediction_scores
+
+
+class BertOnlyNSPHead(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.seq_relationship = nn.Linear(config.hidden_size, 2)
+
+    def forward(self, pooled_output):
+        seq_relationship_score = self.seq_relationship(pooled_output)
+        return seq_relationship_score
+
+
+class BertPreTrainingHeads(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.predictions = BertLMPredictionHead(config)
+        self.seq_relationship = nn.Linear(config.hidden_size, 2)
+
+    def forward(self, sequence_output, pooled_output):
+        prediction_scores = self.predictions(sequence_output)
+        seq_relationship_score = self.seq_relationship(pooled_output)
+        return prediction_scores, seq_relationship_score
+
+
+class BertPreTrainedModel(PreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
+    models.
+    """
+
+    config_class = BertConfig
+    load_tf_weights = load_tf_weights_in_bert
+    base_model_prefix = "bert"
+    supports_gradient_checkpointing = True
+    _keys_to_ignore_on_load_missing = [r"position_ids"]
+
+    def _init_weights(self, module):
+        """Initialize the weights"""
+        if isinstance(module, nn.Linear):
+            # Slightly different from the TF version which uses truncated_normal for initialization
+            # cf https://github.com/pytorch/pytorch/pull/5617
+            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
+            if module.bias is not None:
+                module.bias.data.zero_()
+        elif isinstance(module, nn.Embedding):
+            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
+            if module.padding_idx is not None:
+                module.weight.data[module.padding_idx].zero_()
+        elif isinstance(module, nn.LayerNorm):
+            module.bias.data.zero_()
+            module.weight.data.fill_(1.0)
+
+    def _set_gradient_checkpointing(self, module, value=False):
+        if isinstance(module, BertEncoder):
+            module.gradient_checkpointing = value
+
+
+@dataclass
+class BertForPreTrainingOutput(ModelOutput):
+    """
+    Output type of [`BertForPreTraining`].
+
+    Args:
+        loss (*optional*, returned when `labels` is provided, `torch.FloatTensor` of shape `(1,)`):
+            Total loss as the sum of the masked language modeling loss and the next sequence prediction
+            (classification) loss.
+        prediction_logits (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`):
+            Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
+        seq_relationship_logits (`torch.FloatTensor` of shape `(batch_size, 2)`):
+            Prediction scores of the next sequence prediction (classification) head (scores of True/False continuation
+            before SoftMax).
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings + one for the output of each layer) of
+            shape `(batch_size, sequence_length, hidden_size)`.
+
+            Hidden-states of the model at the output of each layer plus the initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`.
+
+            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
+            heads.
+    """
+
+    loss: Optional[torch.FloatTensor] = None
+    prediction_logits: torch.FloatTensor = None
+    seq_relationship_logits: torch.FloatTensor = None
+    hidden_states: Optional[Tuple[torch.FloatTensor]] = None
+    attentions: Optional[Tuple[torch.FloatTensor]] = None
+
+
+BERT_START_DOCSTRING = r"""
+
+    This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
+    library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
+    etc.)
+
+    This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
+    Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
+    and behavior.
+
+    Parameters:
+        config ([`BertConfig`]): Model configuration class with all the parameters of the model.
+            Initializing with a config file does not load the weights associated with the model, only the
+            configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights.
+"""
+
+BERT_INPUTS_DOCSTRING = r"""
+    Args:
+        input_ids (`torch.LongTensor` of shape `({0})`):
+            Indices of input sequence tokens in the vocabulary.
+
+            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            [What are input IDs?](../glossary#input-ids)
+        attention_mask (`torch.FloatTensor` of shape `({0})`, *optional*):
+            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
+
+            - 1 for tokens that are **not masked**,
+            - 0 for tokens that are **masked**.
+
+            [What are attention masks?](../glossary#attention-mask)
+        token_type_ids (`torch.LongTensor` of shape `({0})`, *optional*):
+            Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,
+            1]`:
+
+            - 0 corresponds to a *sentence A* token,
+            - 1 corresponds to a *sentence B* token.
+
+            [What are token type IDs?](../glossary#token-type-ids)
+        position_ids (`torch.LongTensor` of shape `({0})`, *optional*):
+            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
+            config.max_position_embeddings - 1]`.
+
+            [What are position IDs?](../glossary#position-ids)
+        head_mask (`torch.FloatTensor` of shape `(num_heads,)` or `(num_layers, num_heads)`, *optional*):
+            Mask to nullify selected heads of the self-attention modules. Mask values selected in `[0, 1]`:
+
+            - 1 indicates the head is **not masked**,
+            - 0 indicates the head is **masked**.
+
+        inputs_embeds (`torch.FloatTensor` of shape `({0}, hidden_size)`, *optional*):
+            Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
+            is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
+            model's internal embedding lookup matrix.
+        output_attentions (`bool`, *optional*):
+            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
+            tensors for more detail.
+        output_hidden_states (`bool`, *optional*):
+            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
+            more detail.
+        return_dict (`bool`, *optional*):
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+"""
+
+
+@add_start_docstrings(
+    "The bare Bert Model transformer outputting raw hidden-states without any specific head on top.",
+    BERT_START_DOCSTRING,
+)
+class BertModel(BertPreTrainedModel):
+    """
+
+    The model can behave as an encoder (with only self-attention) as well as a decoder, in which case a layer of
+    cross-attention is added between the self-attention layers, following the architecture described in [Attention is
+    all you need](https://arxiv.org/abs/1706.03762) by Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit,
+    Llion Jones, Aidan N. Gomez, Lukasz Kaiser and Illia Polosukhin.
+
+    To behave as an decoder the model needs to be initialized with the `is_decoder` argument of the configuration set
+    to `True`. To be used in a Seq2Seq model, the model needs to initialized with both `is_decoder` argument and
+    `add_cross_attention` set to `True`; an `encoder_hidden_states` is then expected as an input to the forward pass.
+    """
+
+    def __init__(self, config, add_pooling_layer=True):
+        super().__init__(config)
+        self.config = config
+
+        self.embeddings = BertEmbeddings(config)
+        self.encoder = BertEncoder(config)
+
+        self.pooler = BertPooler(config) if add_pooling_layer else None
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.embeddings.word_embeddings
+
+    def set_input_embeddings(self, value):
+        self.embeddings.word_embeddings = value
+
+    def _prune_heads(self, heads_to_prune):
+        """
+        Prunes heads of the model. heads_to_prune: dict of {layer_num: list of heads to prune in this layer} See base
+        class PreTrainedModel
+        """
+        for layer, heads in heads_to_prune.items():
+            self.encoder.layer[layer].attention.prune_heads(heads)
+
+    @add_start_docstrings_to_model_forward(BERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_DOC,
+        output_type=BaseModelOutputWithPoolingAndCrossAttentions,
+        config_class=_CONFIG_FOR_DOC,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        token_type_ids: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        head_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.Tensor] = None,
+        past_key_values: Optional[List[torch.FloatTensor]] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple[torch.Tensor], BaseModelOutputWithPoolingAndCrossAttentions]:
+        r"""
+        encoder_hidden_states  (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
+            Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention if
+            the model is configured as a decoder.
+        encoder_attention_mask (`torch.FloatTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used in
+            the cross-attention if the model is configured as a decoder. Mask values selected in `[0, 1]`:
+
+            - 1 for tokens that are **not masked**,
+            - 0 for tokens that are **masked**.
+        past_key_values (`tuple(tuple(torch.FloatTensor))` of length `config.n_layers` with each tuple having 4 tensors of shape `(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`):
+            Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding.
+
+            If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that
+            don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all
+            `decoder_input_ids` of shape `(batch_size, sequence_length)`.
+        use_cache (`bool`, *optional*):
+            If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see
+            `past_key_values`).
+        """
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        if self.config.is_decoder:
+            use_cache = use_cache if use_cache is not None else self.config.use_cache
+        else:
+            use_cache = False
+
+        if input_ids is not None and inputs_embeds is not None:
+            raise ValueError("You cannot specify both input_ids and inputs_embeds at the same time")
+        elif input_ids is not None:
+            input_shape = input_ids.size()
+        elif inputs_embeds is not None:
+            input_shape = inputs_embeds.size()[:-1]
+        else:
+            raise ValueError("You have to specify either input_ids or inputs_embeds")
+
+        batch_size, seq_length = input_shape
+        device = input_ids.device if input_ids is not None else inputs_embeds.device
+
+        # past_key_values_length
+        past_key_values_length = past_key_values[0][0].shape[2] if past_key_values is not None else 0
+
+        if attention_mask is None:
+            attention_mask = torch.ones(((batch_size, seq_length + past_key_values_length)), device=device)
+
+        if token_type_ids is None:
+            if hasattr(self.embeddings, "token_type_ids"):
+                buffered_token_type_ids = self.embeddings.token_type_ids[:, :seq_length]
+                buffered_token_type_ids_expanded = buffered_token_type_ids.expand(batch_size, seq_length)
+                token_type_ids = buffered_token_type_ids_expanded
+            else:
+                token_type_ids = torch.zeros(input_shape, dtype=torch.long, device=device)
+
+        # 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.
+        extended_attention_mask: torch.Tensor = self.get_extended_attention_mask(attention_mask, input_shape)
+
+        # If a 2D or 3D attention mask is provided for the cross-attention
+        # we need to make broadcastable to [batch_size, num_heads, seq_length, seq_length]
+        if self.config.is_decoder and encoder_hidden_states is not None:
+            encoder_batch_size, encoder_sequence_length, _ = encoder_hidden_states.size()
+            encoder_hidden_shape = (encoder_batch_size, encoder_sequence_length)
+            if encoder_attention_mask is None:
+                encoder_attention_mask = torch.ones(encoder_hidden_shape, device=device)
+            encoder_extended_attention_mask = self.invert_attention_mask(encoder_attention_mask)
+        else:
+            encoder_extended_attention_mask = None
+
+        # Prepare head mask if needed
+        # 1.0 in head_mask indicate we keep the head
+        # attention_probs has shape bsz x n_heads x N x N
+        # input head_mask has shape [num_heads] or [num_hidden_layers x num_heads]
+        # and head_mask is converted to shape [num_hidden_layers x batch x num_heads x seq_length x seq_length]
+        head_mask = self.get_head_mask(head_mask, self.config.num_hidden_layers)
+
+        embedding_output = self.embeddings(
+            input_ids=input_ids,
+            position_ids=position_ids,
+            token_type_ids=token_type_ids,
+            inputs_embeds=inputs_embeds,
+            past_key_values_length=past_key_values_length,
+        )
+        encoder_outputs = self.encoder(
+            embedding_output,
+            attention_mask=extended_attention_mask,
+            head_mask=head_mask,
+            encoder_hidden_states=encoder_hidden_states,
+            encoder_attention_mask=encoder_extended_attention_mask,
+            past_key_values=past_key_values,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        sequence_output = encoder_outputs[0]
+        pooled_output = self.pooler(sequence_output) if self.pooler is not None else None
+
+        if not return_dict:
+            return (sequence_output, pooled_output) + encoder_outputs[1:]
+
+        return BaseModelOutputWithPoolingAndCrossAttentions(
+            last_hidden_state=sequence_output,
+            pooler_output=pooled_output,
+            past_key_values=encoder_outputs.past_key_values,
+            hidden_states=encoder_outputs.hidden_states,
+            attentions=encoder_outputs.attentions,
+            cross_attentions=encoder_outputs.cross_attentions,
+        )
+
+
+@add_start_docstrings(
+    """
+    Bert Model with two heads on top as done during the pretraining: a `masked language modeling` head and a `next
+    sentence prediction (classification)` head.
+    """,
+    BERT_START_DOCSTRING,
+)
+class BertForPreTraining(BertPreTrainedModel):
+    _keys_to_ignore_on_load_missing = [r"position_ids", r"predictions.decoder.bias", r"cls.predictions.decoder.weight"]
+
+    def __init__(self, config):
+        super().__init__(config)
+
+        self.bert = BertModel(config)
+        self.cls = BertPreTrainingHeads(config)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_output_embeddings(self):
+        return self.cls.predictions.decoder
+
+    def set_output_embeddings(self, new_embeddings):
+        self.cls.predictions.decoder = new_embeddings
+
+    @add_start_docstrings_to_model_forward(BERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
+    @replace_return_docstrings(output_type=BertForPreTrainingOutput, config_class=_CONFIG_FOR_DOC)
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        token_type_ids: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        head_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        labels: Optional[torch.Tensor] = None,
+        next_sentence_label: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple[torch.Tensor], BertForPreTrainingOutput]:
+        r"""
+            labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+                Labels for computing the masked language modeling loss. Indices should be in `[-100, 0, ...,
+                config.vocab_size]` (see `input_ids` docstring) Tokens with indices set to `-100` are ignored (masked),
+                the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`
+            next_sentence_label (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+                Labels for computing the next sequence prediction (classification) loss. Input should be a sequence
+                pair (see `input_ids` docstring) Indices should be in `[0, 1]`:
+
+                - 0 indicates sequence B is a continuation of sequence A,
+                - 1 indicates sequence B is a random sequence.
+            kwargs (`Dict[str, any]`, optional, defaults to *{}*):
+                Used to hide legacy arguments that have been deprecated.
+
+        Returns:
+
+        Example:
+
+        ```python
+        >>> from transformers import AutoTokenizer, BertForPreTraining
+        >>> import torch
+
+        >>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
+        >>> model = BertForPreTraining.from_pretrained("bert-base-uncased")
+
+        >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
+        >>> outputs = model(**inputs)
+
+        >>> prediction_logits = outputs.prediction_logits
+        >>> seq_relationship_logits = outputs.seq_relationship_logits
+        ```
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs = self.bert(
+            input_ids,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        sequence_output, pooled_output = outputs[:2]
+        prediction_scores, seq_relationship_score = self.cls(sequence_output, pooled_output)
+
+        total_loss = None
+        if labels is not None and next_sentence_label is not None:
+            loss_fct = CrossEntropyLoss()
+            masked_lm_loss = loss_fct(prediction_scores.view(-1, self.config.vocab_size), labels.view(-1))
+            next_sentence_loss = loss_fct(seq_relationship_score.view(-1, 2), next_sentence_label.view(-1))
+            total_loss = masked_lm_loss + next_sentence_loss
+
+        if not return_dict:
+            output = (prediction_scores, seq_relationship_score) + outputs[2:]
+            return ((total_loss,) + output) if total_loss is not None else output
+
+        return BertForPreTrainingOutput(
+            loss=total_loss,
+            prediction_logits=prediction_scores,
+            seq_relationship_logits=seq_relationship_score,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+
+@add_start_docstrings(
+    """Bert Model with a `language modeling` head on top for CLM fine-tuning.""", BERT_START_DOCSTRING
+)
+class BertCustomLMHeadModel(BertPreTrainedModel):
+    _keys_to_ignore_on_load_unexpected = [r"pooler"]
+    _keys_to_ignore_on_load_missing = [r"position_ids", r"predictions.decoder.bias", r"cls.predictions.decoder.weight"]
+
+    def __init__(self, config):
+        super().__init__(config)
+
+        if not config.is_decoder:
+            logger.warning("If you want to use `BertLMHeadModel` as a standalone, add `is_decoder=True.`")
+
+        self.bert = BertModel(config, add_pooling_layer=False)
+        self.cls = BertOnlyMLMHead(config)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_output_embeddings(self):
+        return self.cls.predictions.decoder
+
+    def set_output_embeddings(self, new_embeddings):
+        self.cls.predictions.decoder = new_embeddings
+
+    @add_start_docstrings_to_model_forward(BERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_DOC,
+        output_type=CausalLMOutputWithCrossAttentions,
+        config_class=_CONFIG_FOR_DOC,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        token_type_ids: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        head_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.Tensor] = None,
+        labels: Optional[torch.Tensor] = None,
+        past_key_values: Optional[List[torch.Tensor]] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple[torch.Tensor], CausalLMOutputWithCrossAttentions]:
+        r"""
+        encoder_hidden_states  (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
+            Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention if
+            the model is configured as a decoder.
+        encoder_attention_mask (`torch.FloatTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used in
+            the cross-attention if the model is configured as a decoder. Mask values selected in `[0, 1]`:
+
+            - 1 for tokens that are **not masked**,
+            - 0 for tokens that are **masked**.
+        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Labels for computing the left-to-right language modeling loss (next word prediction). Indices should be in
+            `[-100, 0, ..., config.vocab_size]` (see `input_ids` docstring) Tokens with indices set to `-100` are
+            ignored (masked), the loss is only computed for the tokens with labels n `[0, ..., config.vocab_size]`
+        past_key_values (`tuple(tuple(torch.FloatTensor))` of length `config.n_layers` with each tuple having 4 tensors of shape `(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`):
+            Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding.
+
+            If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that
+            don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all
+            `decoder_input_ids` of shape `(batch_size, sequence_length)`.
+        use_cache (`bool`, *optional*):
+            If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see
+            `past_key_values`).
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+        if labels is not None:
+            use_cache = False
+
+        outputs = self.bert(
+            input_ids,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            encoder_hidden_states=encoder_hidden_states,
+            encoder_attention_mask=encoder_attention_mask,
+            past_key_values=past_key_values,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        sequence_output = outputs[0]
+        prediction_scores = self.cls(sequence_output)
+
+        lm_loss = None
+        if labels is not None:
+            # we are doing next-token prediction; shift prediction scores and input ids by one
+            shifted_prediction_scores = prediction_scores[:, :-1, :].contiguous()
+            labels = labels[:, 1:].contiguous()
+            loss_fct = CrossEntropyLoss()
+            lm_loss = loss_fct(shifted_prediction_scores.view(-1, self.config.vocab_size), labels.view(-1))
+
+        if not return_dict:
+            output = (prediction_scores,) + outputs[2:]
+            return ((lm_loss,) + output) if lm_loss is not None else output
+
+        return CausalLMOutputWithCrossAttentions(
+            loss=lm_loss,
+            logits=prediction_scores,
+            past_key_values=outputs.past_key_values,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+            cross_attentions=outputs.cross_attentions,
+        )
+
+    def prepare_inputs_for_generation(
+        self, input_ids, past_key_values=None, attention_mask=None, use_cache=True, **model_kwargs
+    ):
+        input_shape = input_ids.shape
+        # if model is used as a decoder in encoder-decoder model, the decoder attention mask is created on the fly
+        if attention_mask is None:
+            attention_mask = input_ids.new_ones(input_shape)
+
+        # cut decoder_input_ids if past_key_values is used
+        if past_key_values is not None:
+            input_ids = input_ids[:, -1:]
+
+        return {
+            "input_ids": input_ids,
+            "attention_mask": attention_mask,
+            "past_key_values": past_key_values,
+            "use_cache": use_cache,
+        }
+
+    def _reorder_cache(self, past_key_values, beam_idx):
+        reordered_past = ()
+        for layer_past in past_key_values:
+            reordered_past += (tuple(past_state.index_select(0, beam_idx) for past_state in layer_past),)
+        return reordered_past
+
+
+@add_start_docstrings("""Bert Model with a `language modeling` head on top.""", BERT_START_DOCSTRING)
+class BertForMaskedLM(BertPreTrainedModel):
+    _keys_to_ignore_on_load_unexpected = [r"pooler"]
+    _keys_to_ignore_on_load_missing = [r"position_ids", r"predictions.decoder.bias", r"cls.predictions.decoder.weight"]
+
+    def __init__(self, config):
+        super().__init__(config)
+
+        if config.is_decoder:
+            logger.warning(
+                "If you want to use `BertForMaskedLM` make sure `config.is_decoder=False` for "
+                "bi-directional self-attention."
+            )
+
+        self.bert = BertModel(config, add_pooling_layer=False)
+        self.cls = BertOnlyMLMHead(config)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_output_embeddings(self):
+        return self.cls.predictions.decoder
+
+    def set_output_embeddings(self, new_embeddings):
+        self.cls.predictions.decoder = new_embeddings
+
+    @add_start_docstrings_to_model_forward(BERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_DOC,
+        output_type=MaskedLMOutput,
+        config_class=_CONFIG_FOR_DOC,
+        expected_output="'paris'",
+        expected_loss=0.88,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        token_type_ids: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        head_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.Tensor] = None,
+        labels: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple[torch.Tensor], MaskedLMOutput]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Labels for computing the masked language modeling loss. Indices should be in `[-100, 0, ...,
+            config.vocab_size]` (see `input_ids` docstring) Tokens with indices set to `-100` are ignored (masked), the
+            loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`
+        """
+
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs = self.bert(
+            input_ids,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            encoder_hidden_states=encoder_hidden_states,
+            encoder_attention_mask=encoder_attention_mask,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        sequence_output = outputs[0]
+        prediction_scores = self.cls(sequence_output)
+
+        masked_lm_loss = None
+        if labels is not None:
+            loss_fct = CrossEntropyLoss()  # -100 index = padding token
+            masked_lm_loss = loss_fct(prediction_scores.view(-1, self.config.vocab_size), labels.view(-1))
+
+        if not return_dict:
+            output = (prediction_scores,) + outputs[2:]
+            return ((masked_lm_loss,) + output) if masked_lm_loss is not None else output
+
+        return MaskedLMOutput(
+            loss=masked_lm_loss,
+            logits=prediction_scores,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+    def prepare_inputs_for_generation(self, input_ids, attention_mask=None, **model_kwargs):
+        input_shape = input_ids.shape
+        effective_batch_size = input_shape[0]
+
+        #  add a dummy token
+        if self.config.pad_token_id is None:
+            raise ValueError("The PAD token should be defined for generation")
+
+        attention_mask = torch.cat([attention_mask, attention_mask.new_zeros((attention_mask.shape[0], 1))], dim=-1)
+        dummy_token = torch.full(
+            (effective_batch_size, 1), self.config.pad_token_id, dtype=torch.long, device=input_ids.device
+        )
+        input_ids = torch.cat([input_ids, dummy_token], dim=1)
+
+        return {"input_ids": input_ids, "attention_mask": attention_mask}
+
+
+@add_start_docstrings(
+    """Bert Model with a `next sentence prediction (classification)` head on top.""",
+    BERT_START_DOCSTRING,
+)
+class BertForNextSentencePrediction(BertPreTrainedModel):
+    def __init__(self, config):
+        super().__init__(config)
+
+        self.bert = BertModel(config)
+        self.cls = BertOnlyNSPHead(config)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(BERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
+    @replace_return_docstrings(output_type=NextSentencePredictorOutput, config_class=_CONFIG_FOR_DOC)
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        token_type_ids: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        head_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        labels: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        **kwargs,
+    ) -> Union[Tuple[torch.Tensor], NextSentencePredictorOutput]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for computing the next sequence prediction (classification) loss. Input should be a sequence pair
+            (see `input_ids` docstring). Indices should be in `[0, 1]`:
+
+            - 0 indicates sequence B is a continuation of sequence A,
+            - 1 indicates sequence B is a random sequence.
+
+        Returns:
+
+        Example:
+
+        ```python
+        >>> from transformers import AutoTokenizer, BertForNextSentencePrediction
+        >>> import torch
+
+        >>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
+        >>> model = BertForNextSentencePrediction.from_pretrained("bert-base-uncased")
+
+        >>> prompt = "In Italy, pizza served in formal settings, such as at a restaurant, is presented unsliced."
+        >>> next_sentence = "The sky is blue due to the shorter wavelength of blue light."
+        >>> encoding = tokenizer(prompt, next_sentence, return_tensors="pt")
+
+        >>> outputs = model(**encoding, labels=torch.LongTensor([1]))
+        >>> logits = outputs.logits
+        >>> assert logits[0, 0] < logits[0, 1]  # next sentence was random
+        ```
+        """
+
+        if "next_sentence_label" in kwargs:
+            warnings.warn(
+                "The `next_sentence_label` argument is deprecated and will be removed in a future version, use"
+                " `labels` instead.",
+                FutureWarning,
+            )
+            labels = kwargs.pop("next_sentence_label")
+
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs = self.bert(
+            input_ids,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        pooled_output = outputs[1]
+
+        seq_relationship_scores = self.cls(pooled_output)
+
+        next_sentence_loss = None
+        if labels is not None:
+            loss_fct = CrossEntropyLoss()
+            next_sentence_loss = loss_fct(seq_relationship_scores.view(-1, 2), labels.view(-1))
+
+        if not return_dict:
+            output = (seq_relationship_scores,) + outputs[2:]
+            return ((next_sentence_loss,) + output) if next_sentence_loss is not None else output
+
+        return NextSentencePredictorOutput(
+            loss=next_sentence_loss,
+            logits=seq_relationship_scores,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+
+@add_start_docstrings(
+    """
+    Bert Model transformer with a sequence classification/regression head on top (a linear layer on top of the pooled
+    output) e.g. for GLUE tasks.
+    """,
+    BERT_START_DOCSTRING,
+)
+class BertForSequenceClassification(BertPreTrainedModel):
+    def __init__(self, config):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+        self.config = config
+
+        self.bert = BertModel(config)
+        classifier_dropout = (
+            config.classifier_dropout if config.classifier_dropout is not None else config.hidden_dropout_prob
+        )
+        self.dropout = nn.Dropout(classifier_dropout)
+        self.classifier = nn.Linear(config.hidden_size, config.num_labels)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(BERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_SEQUENCE_CLASSIFICATION,
+        output_type=SequenceClassifierOutput,
+        config_class=_CONFIG_FOR_DOC,
+        expected_output=_SEQ_CLASS_EXPECTED_OUTPUT,
+        expected_loss=_SEQ_CLASS_EXPECTED_LOSS,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        token_type_ids: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        head_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        labels: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple[torch.Tensor], SequenceClassifierOutput]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
+            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
+            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs = self.bert(
+            input_ids,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        pooled_output = outputs[1]
+
+        pooled_output = self.dropout(pooled_output)
+        logits = self.classifier(pooled_output)
+
+        loss = None
+        if labels is not None:
+            if self.config.problem_type is None:
+                if self.num_labels == 1:
+                    self.config.problem_type = "regression"
+                elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int):
+                    self.config.problem_type = "single_label_classification"
+                else:
+                    self.config.problem_type = "multi_label_classification"
+
+            if self.config.problem_type == "regression":
+                loss_fct = MSELoss()
+                if self.num_labels == 1:
+                    loss = loss_fct(logits.squeeze(), labels.squeeze())
+                else:
+                    loss = loss_fct(logits, labels)
+            elif self.config.problem_type == "single_label_classification":
+                loss_fct = CrossEntropyLoss()
+                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
+            elif self.config.problem_type == "multi_label_classification":
+                loss_fct = BCEWithLogitsLoss()
+                loss = loss_fct(logits, labels)
+        if not return_dict:
+            output = (logits,) + outputs[2:]
+            return ((loss,) + output) if loss is not None else output
+
+        return SequenceClassifierOutput(
+            loss=loss,
+            logits=logits,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+
+@add_start_docstrings(
+    """
+    Bert Model with a multiple choice classification head on top (a linear layer on top of the pooled output and a
+    softmax) e.g. for RocStories/SWAG tasks.
+    """,
+    BERT_START_DOCSTRING,
+)
+class BertForMultipleChoice(BertPreTrainedModel):
+    def __init__(self, config):
+        super().__init__(config)
+
+        self.bert = BertModel(config)
+        classifier_dropout = (
+            config.classifier_dropout if config.classifier_dropout is not None else config.hidden_dropout_prob
+        )
+        self.dropout = nn.Dropout(classifier_dropout)
+        self.classifier = nn.Linear(config.hidden_size, 1)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(BERT_INPUTS_DOCSTRING.format("batch_size, num_choices, sequence_length"))
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_DOC,
+        output_type=MultipleChoiceModelOutput,
+        config_class=_CONFIG_FOR_DOC,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        token_type_ids: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        head_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        labels: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple[torch.Tensor], MultipleChoiceModelOutput]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for computing the multiple choice classification loss. Indices should be in `[0, ...,
+            num_choices-1]` where `num_choices` is the size of the second dimension of the input tensors. (See
+            `input_ids` above)
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+        num_choices = input_ids.shape[1] if input_ids is not None else inputs_embeds.shape[1]
+
+        input_ids = input_ids.view(-1, input_ids.size(-1)) if input_ids is not None else None
+        attention_mask = attention_mask.view(-1, attention_mask.size(-1)) if attention_mask is not None else None
+        token_type_ids = token_type_ids.view(-1, token_type_ids.size(-1)) if token_type_ids is not None else None
+        position_ids = position_ids.view(-1, position_ids.size(-1)) if position_ids is not None else None
+        inputs_embeds = (
+            inputs_embeds.view(-1, inputs_embeds.size(-2), inputs_embeds.size(-1))
+            if inputs_embeds is not None
+            else None
+        )
+
+        outputs = self.bert(
+            input_ids,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        pooled_output = outputs[1]
+
+        pooled_output = self.dropout(pooled_output)
+        logits = self.classifier(pooled_output)
+        reshaped_logits = logits.view(-1, num_choices)
+
+        loss = None
+        if labels is not None:
+            loss_fct = CrossEntropyLoss()
+            loss = loss_fct(reshaped_logits, labels)
+
+        if not return_dict:
+            output = (reshaped_logits,) + outputs[2:]
+            return ((loss,) + output) if loss is not None else output
+
+        return MultipleChoiceModelOutput(
+            loss=loss,
+            logits=reshaped_logits,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+
+@add_start_docstrings(
+    """
+    Bert Model with a token classification head on top (a linear layer on top of the hidden-states output) e.g. for
+    Named-Entity-Recognition (NER) tasks.
+    """,
+    BERT_START_DOCSTRING,
+)
+class BertForTokenClassification(BertPreTrainedModel):
+    _keys_to_ignore_on_load_unexpected = [r"pooler"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+
+        self.bert = BertModel(config, add_pooling_layer=False)
+        classifier_dropout = (
+            config.classifier_dropout if config.classifier_dropout is not None else config.hidden_dropout_prob
+        )
+        self.dropout = nn.Dropout(classifier_dropout)
+        self.classifier = nn.Linear(config.hidden_size, config.num_labels)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(BERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_TOKEN_CLASSIFICATION,
+        output_type=TokenClassifierOutput,
+        config_class=_CONFIG_FOR_DOC,
+        expected_output=_TOKEN_CLASS_EXPECTED_OUTPUT,
+        expected_loss=_TOKEN_CLASS_EXPECTED_LOSS,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        token_type_ids: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        head_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        labels: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple[torch.Tensor], TokenClassifierOutput]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Labels for computing the token classification loss. Indices should be in `[0, ..., config.num_labels - 1]`.
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs = self.bert(
+            input_ids,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        sequence_output = outputs[0]
+
+        sequence_output = self.dropout(sequence_output)
+        logits = self.classifier(sequence_output)
+
+        loss = None
+        if labels is not None:
+            loss_fct = CrossEntropyLoss()
+            loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
+
+        if not return_dict:
+            output = (logits,) + outputs[2:]
+            return ((loss,) + output) if loss is not None else output
+
+        return TokenClassifierOutput(
+            loss=loss,
+            logits=logits,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+
+@add_start_docstrings(
+    """
+    Bert Model with a span classification head on top for extractive question-answering tasks like SQuAD (a linear
+    layers on top of the hidden-states output to compute `span start logits` and `span end logits`).
+    """,
+    BERT_START_DOCSTRING,
+)
+class BertForQuestionAnswering(BertPreTrainedModel):
+    _keys_to_ignore_on_load_unexpected = [r"pooler"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+
+        self.bert = BertModel(config, add_pooling_layer=False)
+        self.qa_outputs = nn.Linear(config.hidden_size, config.num_labels)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(BERT_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_QA,
+        output_type=QuestionAnsweringModelOutput,
+        config_class=_CONFIG_FOR_DOC,
+        qa_target_start_index=_QA_TARGET_START_INDEX,
+        qa_target_end_index=_QA_TARGET_END_INDEX,
+        expected_output=_QA_EXPECTED_OUTPUT,
+        expected_loss=_QA_EXPECTED_LOSS,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        token_type_ids: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.Tensor] = None,
+        head_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        start_positions: Optional[torch.Tensor] = None,
+        end_positions: Optional[torch.Tensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple[torch.Tensor], QuestionAnsweringModelOutput]:
+        r"""
+        start_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for position (index) of the start of the labelled span for computing the token classification loss.
+            Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence
+            are not taken into account for computing the loss.
+        end_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
+            Labels for position (index) of the end of the labelled span for computing the token classification loss.
+            Positions are clamped to the length of the sequence (`sequence_length`). Position outside of the sequence
+            are not taken into account for computing the loss.
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs = self.bert(
+            input_ids,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        sequence_output = outputs[0]
+
+        logits = self.qa_outputs(sequence_output)
+        start_logits, end_logits = logits.split(1, dim=-1)
+        start_logits = start_logits.squeeze(-1).contiguous()
+        end_logits = end_logits.squeeze(-1).contiguous()
+
+        total_loss = None
+        if start_positions is not None and end_positions is not None:
+            # If we are on multi-GPU, split add a dimension
+            if len(start_positions.size()) > 1:
+                start_positions = start_positions.squeeze(-1)
+            if len(end_positions.size()) > 1:
+                end_positions = end_positions.squeeze(-1)
+            # sometimes the start/end positions are outside our model inputs, we ignore these terms
+            ignored_index = start_logits.size(1)
+            start_positions = start_positions.clamp(0, ignored_index)
+            end_positions = end_positions.clamp(0, ignored_index)
+
+            loss_fct = CrossEntropyLoss(ignore_index=ignored_index)
+            start_loss = loss_fct(start_logits, start_positions)
+            end_loss = loss_fct(end_logits, end_positions)
+            total_loss = (start_loss + end_loss) / 2
+
+        if not return_dict:
+            output = (start_logits, end_logits) + outputs[2:]
+            return ((total_loss,) + output) if total_loss is not None else output
+
+        return QuestionAnsweringModelOutput(
+            loss=total_loss,
+            start_logits=start_logits,
+            end_logits=end_logits,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:545d8feae7cdaa752dfcecd8d480928b31a0f7a0b494877c9ab5ddf504906703
+size 383481

special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

tf_model.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8961e0116b64f7aa000cdee56f226922e47168126dfc846a85b935b259311edf
+size 472416

tokenizer.json

@@ -0,0 +1,1274 @@
+{
+  "version": "1.0",
+  "truncation": null,
+  "padding": null,
+  "added_tokens": [
+    {
+      "id": 0,
+      "content": "[PAD]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 1,
+      "content": "[UNK]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 2,
+      "content": "[CLS]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 3,
+      "content": "[SEP]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 4,
+      "content": "[MASK]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    }
+  ],
+  "normalizer": {
+    "type": "BertNormalizer",
+    "clean_text": true,
+    "handle_chinese_chars": true,
+    "strip_accents": null,
+    "lowercase": true
+  },
+  "pre_tokenizer": {
+    "type": "BertPreTokenizer"
+  },
+  "post_processor": {
+    "type": "TemplateProcessing",
+    "single": [
+      {
+        "SpecialToken": {
+          "id": "[CLS]",
+          "type_id": 0
+        }
+      },
+      {
+        "Sequence": {
+          "id": "A",
+          "type_id": 0
+        }
+      },
+      {
+        "SpecialToken": {
+          "id": "[SEP]",
+          "type_id": 0
+        }
+      }
+    ],
+    "pair": [
+      {
+        "SpecialToken": {
+          "id": "[CLS]",
+          "type_id": 0
+        }
+      },
+      {
+        "Sequence": {
+          "id": "A",
+          "type_id": 0
+        }
+      },
+      {
+        "SpecialToken": {
+          "id": "[SEP]",
+          "type_id": 0
+        }
+      },
+      {
+        "Sequence": {
+          "id": "B",
+          "type_id": 1
+        }
+      },
+      {
+        "SpecialToken": {
+          "id": "[SEP]",
+          "type_id": 1
+        }
+      }
+    ],
+    "special_tokens": {
+      "[CLS]": {
+        "id": "[CLS]",
+        "ids": [
+          2
+        ],
+        "tokens": [
+          "[CLS]"
+        ]
+      },
+      "[SEP]": {
+        "id": "[SEP]",
+        "ids": [
+          3
+        ],
+        "tokens": [
+          "[SEP]"
+        ]
+      }
+    }
+  },
+  "decoder": {
+    "type": "WordPiece",
+    "prefix": "##",
+    "cleanup": true
+  },
+  "model": {
+    "type": "WordPiece",
+    "unk_token": "[UNK]",
+    "continuing_subword_prefix": "##",
+    "max_input_chars_per_word": 100,
+    "vocab": {
+      "[PAD]": 0,
+      "[UNK]": 1,
+      "[CLS]": 2,
+      "[SEP]": 3,
+      "[MASK]": 4,
+      "!": 5,
+      "\"": 6,
+      "#": 7,
+      "$": 8,
+      "%": 9,
+      "&": 10,
+      "'": 11,
+      "(": 12,
+      ")": 13,
+      "*": 14,
+      "+": 15,
+      ",": 16,
+      "-": 17,
+      ".": 18,
+      "/": 19,
+      "0": 20,
+      "1": 21,
+      "2": 22,
+      "3": 23,
+      "4": 24,
+      "5": 25,
+      "6": 26,
+      "7": 27,
+      "8": 28,
+      "9": 29,
+      ":": 30,
+      ";": 31,
+      "<": 32,
+      "=": 33,
+      ">": 34,
+      "?": 35,
+      "@": 36,
+      "[": 37,
+      "\\": 38,
+      "]": 39,
+      "^": 40,
+      "_": 41,
+      "`": 42,
+      "a": 43,
+      "b": 44,
+      "c": 45,
+      "d": 46,
+      "e": 47,
+      "f": 48,
+      "g": 49,
+      "h": 50,
+      "i": 51,
+      "j": 52,
+      "k": 53,
+      "l": 54,
+      "m": 55,
+      "n": 56,
+      "o": 57,
+      "p": 58,
+      "q": 59,
+      "r": 60,
+      "s": 61,
+      "t": 62,
+      "u": 63,
+      "v": 64,
+      "w": 65,
+      "x": 66,
+      "y": 67,
+      "z": 68,
+      "|": 69,
+      "}": 70,
+      "~": 71,
+      "¡": 72,
+      "¢": 73,
+      "£": 74,
+      "¥": 75,
+      "§": 76,
+      "°": 77,
+      "±": 78,
+      "²": 79,
+      "³": 80,
+      "´": 81,
+      "µ": 82,
+      "·": 83,
+      "º": 84,
+      "½": 85,
+      "¿": 86,
+      "×": 87,
+      "ß": 88,
+      "æ": 89,
+      "ð": 90,
+      "ø": 91,
+      "þ": 92,
+      "đ": 93,
+      "ħ": 94,
+      "ı": 95,
+      "ł": 96,
+      "œ": 97,
+      "ɐ": 98,
+      "ɑ": 99,
+      "ɒ": 100,
+      "ɔ": 101,
+      "ə": 102,
+      "ɛ": 103,
+      "ɜ": 104,
+      "ɡ": 105,
+      "ɢ": 106,
+      "ɪ": 107,
+      "ɫ": 108,
+      "ɳ": 109,
+      "ɽ": 110,
+      "ɾ": 111,
+      "ʁ": 112,
+      "ʃ": 113,
+      "ʊ": 114,
+      "ʋ": 115,
+      "ʒ": 116,
+      "ʔ": 117,
+      "ʕ": 118,
+      "ʲ": 119,
+      "ʻ": 120,
+      "ʼ": 121,
+      "ʾ": 122,
+      "ʿ": 123,
+      "ˈ": 124,
+      "ˌ": 125,
+      "ː": 126,
+      "α": 127,
+      "β": 128,
+      "γ": 129,
+      "δ": 130,
+      "ε": 131,
+      "η": 132,
+      "θ": 133,
+      "ι": 134,
+      "κ": 135,
+      "λ": 136,
+      "μ": 137,
+      "��": 138,
+      "ξ": 139,
+      "ο": 140,
+      "π": 141,
+      "ρ": 142,
+      "ς": 143,
+      "σ": 144,
+      "τ": 145,
+      "υ": 146,
+      "φ": 147,
+      "χ": 148,
+      "ψ": 149,
+      "ω": 150,
+      "а": 151,
+      "б": 152,
+      "в": 153,
+      "г": 154,
+      "д": 155,
+      "е": 156,
+      "ж": 157,
+      "з": 158,
+      "и": 159,
+      "к": 160,
+      "л": 161,
+      "м": 162,
+      "н": 163,
+      "о": 164,
+      "п": 165,
+      "р": 166,
+      "с": 167,
+      "т": 168,
+      "у": 169,
+      "х": 170,
+      "ц": 171,
+      "ш": 172,
+      "ъ": 173,
+      "ы": 174,
+      "ь": 175,
+      "ю": 176,
+      "я": 177,
+      "є": 178,
+      "א": 179,
+      "ב": 180,
+      "ג": 181,
+      "ה": 182,
+      "ו": 183,
+      "ז": 184,
+      "ח": 185,
+      "י": 186,
+      "ל": 187,
+      "ם": 188,
+      "מ": 189,
+      "ן": 190,
+      "נ": 191,
+      "ס": 192,
+      "ף": 193,
+      "פ": 194,
+      "צ": 195,
+      "ר": 196,
+      "ש": 197,
+      "ת": 198,
+      "ء": 199,
+      "ا": 200,
+      "ب": 201,
+      "ة": 202,
+      "ت": 203,
+      "ث": 204,
+      "ج": 205,
+      "ح": 206,
+      "خ": 207,
+      "د": 208,
+      "ذ": 209,
+      "ر": 210,
+      "س": 211,
+      "ش": 212,
+      "ص": 213,
+      "ع": 214,
+      "ف": 215,
+      "ق": 216,
+      "ك": 217,
+      "ل": 218,
+      "م": 219,
+      "ن": 220,
+      "ه": 221,
+      "و": 222,
+      "ي": 223,
+      "ܐ": 224,
+      "ܕ": 225,
+      "ܗ": 226,
+      "ܝ": 227,
+      "ܠ": 228,
+      "ܢ": 229,
+      "ܬ": 230,
+      "अ": 231,
+      "ई": 232,
+      "क": 233,
+      "ग": 234,
+      "ण": 235,
+      "त": 236,
+      "द": 237,
+      "न": 238,
+      "प": 239,
+      "ब": 240,
+      "म": 241,
+      "य": 242,
+      "र": 243,
+      "ल": 244,
+      "व": 245,
+      "स": 246,
+      "ह": 247,
+      "ा": 248,
+      "ि": 249,
+      "আ": 250,
+      "ল": 251,
+      "হ": 252,
+      "া": 253,
+      "ਅ": 254,
+      "ਲ": 255,
+      "ਹ": 256,
+      "ਾ": 257,
+      "അ": 258,
+      "ള": 259,
+      "ഹ": 260,
+      "ാ": 261,
+      "ก": 262,
+      "ค": 263,
+      "ง": 264,
+      "ช": 265,
+      "ซ": 266,
+      "ญ": 267,
+      "ฐ": 268,
+      "ณ": 269,
+      "ด": 270,
+      "ต": 271,
+      "น": 272,
+      "บ": 273,
+      "ป": 274,
+      "พ": 275,
+      "ภ": 276,
+      "ม": 277,
+      "ย": 278,
+      "ร": 279,
+      "ล": 280,
+      "ว": 281,
+      "ศ": 282,
+      "ษ": 283,
+      "ส": 284,
+      "ห": 285,
+      "อ": 286,
+      "ฮ": 287,
+      "ะ": 288,
+      "า": 289,
+      "เ": 290,
+      "แ": 291,
+      "ไ": 292,
+      "ა": 293,
+      "ბ": 294,
+      "გ": 295,
+      "დ": 296,
+      "ე": 297,
+      "ვ": 298,
+      "ზ": 299,
+      "თ": 300,
+      "ი": 301,
+      "კ": 302,
+      "ლ": 303,
+      "მ": 304,
+      "ნ": 305,
+      "ო": 306,
+      "პ": 307,
+      "ჟ": 308,
+      "რ": 309,
+      "ს": 310,
+      "ტ": 311,
+      "უ": 312,
+      "ფ": 313,
+      "ქ": 314,
+      "ღ": 315,
+      "ყ": 316,
+      "შ": 317,
+      "ჩ": 318,
+      "ც": 319,
+      "ძ": 320,
+      "წ": 321,
+      "ჭ": 322,
+      "ხ": 323,
+      "ჯ": 324,
+      "ჰ": 325,
+      "ჱ": 326,
+      "ჲ": 327,
+      "ჳ": 328,
+      "ჴ": 329,
+      "ჵ": 330,
+      "ჶ": 331,
+      "ჷ": 332,
+      "ჸ": 333,
+      "ჹ": 334,
+      "ჺ": 335,
+      "჻": 336,
+      "ᄃ": 337,
+      "ᄅ": 338,
+      "ᄇ": 339,
+      "ᄋ": 340,
+      "ᄌ": 341,
+      "ᅡ": 342,
+      "ᅢ": 343,
+      "ᅦ": 344,
+      "ᅧ": 345,
+      "ᅩ": 346,
+      "ᅮ": 347,
+      "ᅵ": 348,
+      "ᆨ": 349,
+      "ᆫ": 350,
+      "ᆯ": 351,
+      "ᆸ": 352,
+      "ᆼ": 353,
+      "ᵻ": 354,
+      "‐": 355,
+      "‑": 356,
+      "–": 357,
+      "—": 358,
+      "―": 359,
+      "‘": 360,
+      "’": 361,
+      "“": 362,
+      "”": 363,
+      "„": 364,
+      "†": 365,
+      "‡": 366,
+      "•": 367,
+      "…": 368,
+      "′": 369,
+      "″": 370,
+      "⁄": 371,
+      "₣": 372,
+      "₤": 373,
+      "€": 374,
+      "₹": 375,
+      "⅓": 376,
+      "⅔": 377,
+      "→": 378,
+      "−": 379,
+      "≡": 380,
+      "≤": 381,
+      "①": 382,
+      "☉": 383,
+      "☫": 384,
+      "♀": 385,
+      "♭": 386,
+      "♯": 387,
+      "⚳": 388,
+      "ⴀ": 389,
+      "ⴂ": 390,
+      "ⴃ": 391,
+      "ⴈ": 392,
+      "ⴌ": 393,
+      "ⴕ": 394,
+      "ⴟ": 395,
+      "〈": 396,
+      "〉": 397,
+      "〜": 398,
+      "あ": 399,
+      "い": 400,
+      "う": 401,
+      "お": 402,
+      "か": 403,
+      "き": 404,
+      "く": 405,
+      "け": 406,
+      "こ": 407,
+      "さ": 408,
+      "し": 409,
+      "す": 410,
+      "せ": 411,
+      "た": 412,
+      "ち": 413,
+      "っ": 414,
+      "つ": 415,
+      "と": 416,
+      "な": 417,
+      "に": 418,
+      "の": 419,
+      "は": 420,
+      "ひ": 421,
+      "ふ": 422,
+      "ほ": 423,
+      "ま": 424,
+      "み": 425,
+      "め": 426,
+      "も": 427,
+      "ゃ": 428,
+      "ゆ": 429,
+      "ょ": 430,
+      "ら": 431,
+      "り": 432,
+      "る": 433,
+      "れ": 434,
+      "わ": 435,
+      "を": 436,
+      "ん": 437,
+      "ァ": 438,
+      "ア": 439,
+      "ィ": 440,
+      "イ": 441,
+      "ゥ": 442,
+      "ウ": 443,
+      "ェ": 444,
+      "エ": 445,
+      "ォ": 446,
+      "オ": 447,
+      "カ": 448,
+      "キ": 449,
+      "ク": 450,
+      "ケ": 451,
+      "コ": 452,
+      "サ": 453,
+      "シ": 454,
+      "ス": 455,
+      "セ": 456,
+      "タ": 457,
+      "チ": 458,
+      "ッ": 459,
+      "ツ": 460,
+      "テ": 461,
+      "ト": 462,
+      "ナ": 463,
+      "ニ": 464,
+      "ネ": 465,
+      "ノ": 466,
+      "ハ": 467,
+      "フ": 468,
+      "ヘ": 469,
+      "マ": 470,
+      "ミ": 471,
+      "ム": 472,
+      "モ": 473,
+      "ャ": 474,
+      "ュ": 475,
+      "ョ": 476,
+      "ラ": 477,
+      "リ": 478,
+      "ル": 479,
+      "レ": 480,
+      "ロ": 481,
+      "ン": 482,
+      "・": 483,
+      "ー": 484,
+      "一": 485,
+      "七": 486,
+      "下": 487,
+      "世": 488,
+      "丙": 489,
+      "中": 490,
+      "主": 491,
+      "乃": 492,
+      "之": 493,
+      "乙": 494,
+      "九": 495,
+      "二": 496,
+      "云": 497,
+      "人": 498,
+      "今": 499,
+      "付": 500,
+      "作": 501,
+      "侗": 502,
+      "依": 503,
+      "信": 504,
+      "傳": 505,
+      "儚": 506,
+      "充": 507,
+      "光": 508,
+      "全": 509,
+      "兵": 510,
+      "其": 511,
+      "具": 512,
+      "円": 513,
+      "再": 514,
+      "出": 515,
+      "判": 516,
+      "前": 517,
+      "剛": 518,
+      "劇": 519,
+      "劉": 520,
+      "動": 521,
+      "化": 522,
+      "北": 523,
+      "华": 524,
+      "厂": 525,
+      "去": 526,
+      "古": 527,
+      "可": 528,
+      "台": 529,
+      "史": 530,
+      "同": 531,
+      "名": 532,
+      "君": 533,
+      "吳": 534,
+      "周": 535,
+      "命": 536,
+      "和": 537,
+      "咲": 538,
+      "善": 539,
+      "四": 540,
+      "國": 541,
+      "園": 542,
+      "圣": 543,
+      "在": 544,
+      "坂": 545,
+      "堤": 546,
+      "場": 547,
+      "塘": 548,
+      "夕": 549,
+      "大": 550,
+      "天": 551,
+      "夫": 552,
+      "女": 553,
+      "妙": 554,
+      "姚": 555,
+      "子": 556,
+      "孟": 557,
+      "守": 558,
+      "安": 559,
+      "宋": 560,
+      "完": 561,
+      "宗": 562,
+      "宝": 563,
+      "宫": 564,
+      "寝": 565,
+      "寺": 566,
+      "小": 567,
+      "少": 568,
+      "尾": 569,
+      "山": 570,
+      "岳": 571,
+      "川": 572,
+      "州": 573,
+      "巳": 574,
+      "市": 575,
+      "師": 576,
+      "平": 577,
+      "广": 578,
+      "庆": 579,
+      "府": 580,
+      "座": 581,
+      "廬": 582,
+      "建": 583,
+      "式": 584,
+      "張": 585,
+      "彌": 586,
+      "彩": 587,
+      "彼": 588,
+      "後": 589,
+      "御": 590,
+      "德": 591,
+      "思": 592,
+      "愛": 593,
+      "憑": 594,
+      "憶": 595,
+      "應": 596,
+      "懷": 597,
+      "战": 598,
+      "戦": 599,
+      "扈": 600,
+      "技": 601,
+      "拉": 602,
+      "拳": 603,
+      "挑": 604,
+      "揺": 605,
+      "攻": 606,
+      "放": 607,
+      "政": 608,
+      "散": 609,
+      "斯": 610,
+      "方": 611,
+      "日": 612,
+      "旦": 613,
+      "旭": 614,
+      "昌": 615,
+      "明": 616,
+      "星": 617,
+      "春": 618,
+      "晋": 619,
+      "景": 620,
+      "曦": 621,
+      "月": 622,
+      "望": 623,
+      "未": 624,
+      "本": 625,
+      "李": 626,
+      "村": 627,
+      "杜": 628,
+      "束": 629,
+      "来": 630,
+      "林": 631,
+      "桜": 632,
+      "梶": 633,
+      "棘": 634,
+      "椎": 635,
+      "楊": 636,
+      "楚": 637,
+      "榮": 638,
+      "橘": 639,
+      "機": 640,
+      "正": 641,
+      "殻": 642,
+      "殿": 643,
+      "母": 644,
+      "水": 645,
+      "汉": 646,
+      "沂": 647,
+      "沙": 648,
+      "河": 649,
+      "泗": 650,
+      "波": 651,
+      "泣": 652,
+      "洪": 653,
+      "淹": 654,
+      "清": 655,
+      "湯": 656,
+      "漢": 657,
+      "澄": 658,
+      "澤": 659,
+      "火": 660,
+      "灯": 661,
+      "灵": 662,
+      "灼": 663,
+      "焼": 664,
+      "熱": 665,
+      "物": 666,
+      "狐": 667,
+      "狸": 668,
+      "玄": 669,
+      "王": 670,
+      "玩": 671,
+      "珂": 672,
+      "珙": 673,
+      "球": 674,
+      "理": 675,
+      "琦": 676,
+      "琪": 677,
+      "瓊": 678,
+      "生": 679,
+      "田": 680,
+      "畢": 681,
+      "番": 682,
+      "瘡": 683,
+      "白": 684,
+      "皮": 685,
+      "真": 686,
+      "砲": 687,
+      "礮": 688,
+      "祈": 689,
+      "神": 690,
+      "祠": 691,
+      "秋": 692,
+      "空": 693,
+      "立": 694,
+      "精": 695,
+      "約": 696,
+      "絵": 697,
+      "織": 698,
+      "義": 699,
+      "翠": 700,
+      "者": 701,
+      "耕": 702,
+      "肖": 703,
+      "胡": 704,
+      "膀": 705,
+      "臂": 706,
+      "興": 707,
+      "良": 708,
+      "花": 709,
+      "芳": 710,
+      "芽": 711,
+      "若": 712,
+      "英": 713,
+      "藕": 714,
+      "藥": 715,
+      "蘄": 716,
+      "蘇": 717,
+      "行": 718,
+      "裁": 719,
+      "規": 720,
+      "覺": 721,
+      "观": 722,
+      "解": 723,
+      "記": 724,
+      "誓": 725,
+      "誡": 726,
+      "誰": 727,
+      "謎": 728,
+      "许": 729,
+      "谭": 730,
+      "豪": 731,
+      "豫": 732,
+      "費": 733,
+      "贵": 734,
+      "赤": 735,
+      "趙": 736,
+      "足": 737,
+      "跡": 738,
+      "転": 739,
+      "辛": 740,
+      "逆": 741,
+      "遇": 742,
+      "運": 743,
+      "過": 744,
+      "遠": 745,
+      "選": 746,
+      "邦": 747,
+      "邱": 748,
+      "部": 749,
+      "郭": 750,
+      "都": 751,
+      "酈": 752,
+      "里": 753,
+      "野": 754,
+      "金": 755,
+      "銃": 756,
+      "鋼": 757,
+      "錄": 758,
+      "錡": 759,
+      "鍵": 760,
+      "鐵": 761,
+      "钱": 762,
+      "铁": 763,
+      "關": 764,
+      "防": 765,
+      "阿": 766,
+      "陈": 767,
+      "陳": 768,
+      "陽": 769,
+      "隊": 770,
+      "階": 771,
+      "集": 772,
+      "雪": 773,
+      "雲": 774,
+      "霖": 775,
+      "霹": 776,
+      "靂": 777,
+      "韓": 778,
+      "願": 779,
+      "顯": 780,
+      "颜": 781,
+      "马": 782,
+      "高": 783,
+      "龍": 784,
+      "ﷲ": 785,
+      "ﻋ": 786,
+      "／": 787,
+      "３": 788,
+      "～": 789,
+      "##i": 790,
+      "##y": 791,
+      "##o": 792,
+      "##r": 793,
+      "##g": 794,
+      "##a": 795,
+      "##w": 796,
+      "##l": 797,
+      "##b": 798,
+      "##z": 799,
+      "##t": 800,
+      "##n": 801,
+      "##c": 802,
+      "##h": 803,
+      "##s": 804,
+      "##u": 805,
+      "##d": 806,
+      "##e": 807,
+      "##k": 808,
+      "##v": 809,
+      "##f": 810,
+      "##x": 811,
+      "##q": 812,
+      "##p": 813,
+      "##æ": 814,
+      "##0": 815,
+      "##5": 816,
+      "##m": 817,
+      "##8": 818,
+      "##4": 819,
+      "##س": 820,
+      "##ت": 821,
+      "##ا": 822,
+      "##ن": 823,
+      "##6": 824,
+      "##1": 825,
+      "##7": 826,
+      "##j": 827,
+      "##つ": 828,
+      "##う": 829,
+      "##2": 830,
+      "##9": 831,
+      "##3": 832,
+      "##ø": 833,
+      "##ล": 834,
+      "##ว": 835,
+      "##ง": 836,
+      "##พ": 837,
+      "##ไ": 838,
+      "##ช": 839,
+      "##ย": 840,
+      "##า": 841,
+      "##ร": 842,
+      "##თ": 843,
+      "##ა": 844,
+      "##ვ": 845,
+      "##რ": 846,
+      "##ი": 847,
+      "##ള": 848,
+      "##あ": 849,
+      "##ん": 850,
+      "##α": 851,
+      "##ν": 852,
+      "##τ": 853,
+      "##ο": 854,
+      "##κ": 855,
+      "##ρ": 856,
+      "##ω": 857,
+      "##ς": 858,
+      "##の": 859,
+      "##な": 860,
+      "##ら": 861,
+      "##ð": 862,
+      "##œ": 863,
+      "##ɛ": 864,
+      "##ł": 865,
+      "##η": 866,
+      "##μ": 867,
+      "##ซ": 868,
+      "##ル": 869,
+      "##シ": 870,
+      "##ア": 871,
+      "##リ": 872,
+      "##ス": 873,
+      "##ʔ": 874,
+      "##ल": 875,
+      "##ᄇ": 876,
+      "##ᅮ": 877,
+      "##ᄃ": 878,
+      "##ᅢ": 879,
+      "##β": 880,
+      "##ß": 881,
+      "##か": 882,
+      "##た": 883,
+      "##ə": 884,
+      "##ʻ": 885,
+      "##ι": 886,
+      "##χ": 887,
+      "##о": 888,
+      "##л": 889,
+      "##с": 890,
+      "##а": 891,
+      "##т": 892,
+      "##ы": 893,
+      "##и": 894,
+      "##в": 895,
+      "##к": 896,
+      "##з": 897,
+      "##ッ": 898,
+      "##ク": 899,
+      "##マ": 900,
+      "##ン": 901,
+      "##გ": 902,
+      "##ლ": 903,
+      "##ო": 904,
+      "##ნ": 905,
+      "##ː": 906,
+      "##ל": 907,
+      "##ה": 908,
+      "##א": 909,
+      "##く": 910,
+      "##み": 911,
+      "##ε": 912,
+      "##ξ": 913,
+      "##ল": 914,
+      "##ˈ": 915,
+      "##ɡ": 916,
+      "##ɑ": 917,
+      "##ɒ": 918,
+      "##し": 919,
+      "##す": 920,
+      "##き": 921,
+      "##ひ": 922,
+      "##と": 923,
+      "##đ": 924,
+      "##ъ": 925,
+      "##н": 926,
+      "##е": 927,
+      "##י": 928,
+      "##פ": 929,
+      "##イ": 930,
+      "##λ": 931,
+      "##ق": 932,
+      "##ع": 933,
+      "##د": 934,
+      "##ᅡ": 935,
+      "##ᆯ": 936,
+      "##ᄅ": 937,
+      "##ɪ": 938,
+      "##ค": 939,
+      "##ต": 940,
+      "##व": 941,
+      "##��": 942,
+      "##द": 943,
+      "##は": 944,
+      "##り": 945,
+      "##レ": 946,
+      "##ー": 947,
+      "##ツ": 948,
+      "##ي": 949,
+      "##ش": 950,
+      "##و": 951,
+      "##م": 952,
+      "##º": 953,
+      "##ਲ": 954,
+      "##ਾ": 955,
+      "##ਹ": 956,
+      "##д": 957,
+      "##р": 958,
+      "##ل": 959,
+      "##ب": 960,
+      "##い": 961,
+      "##ち": 962,
+      "##ゃ": 963,
+      "##ʒ": 964,
+      "##ʃ": 965,
+      "##ɔ": 966,
+      "##ह": 967,
+      "##ニ": 968,
+      "##ウ": 969,
+      "##ァ": 970,
+      "##キ": 971,
+      "##ュ": 972,
+      "##３": 973,
+      "##ხ": 974,
+      "##ს": 975,
+      "##お": 976,
+      "##タ": 977,
+      "##ാ": 978,
+      "##ഹ": 979,
+      "##ɳ": 980,
+      "##ま": 981,
+      "##る": 982,
+      "##ะ": 983,
+      "##อ": 984,
+      "##น": 985,
+      "##ן": 986,
+      "##я": 987,
+      "##แ": 988,
+      "##ก": 989,
+      "##ɾ": 990,
+      "##ʲ": 991,
+      "##フ": 992,
+      "##უ": 993,
+      "##ภ": 994,
+      "##ด": 995,
+      "##ב": 996,
+      "##ת": 997,
+      "##خ": 998,
+      "##ラ": 999,
+      "##れ": 1000,
+      "##ण": 1001,
+      "##स": 1002,
+      "##न": 1003,
+      "##ه": 1004,
+      "##ف": 1005,
+      "##ر": 1006,
+      "##エ": 1007,
+      "##テ": 1008,
+      "##ษ": 1009,
+      "##ฐ": 1010,
+      "##ィ": 1011,
+      "##क": 1012,
+      "##ノ": 1013,
+      "##θ": 1014,
+      "##ネ": 1015,
+      "##ョ": 1016,
+      "##δ": 1017,
+      "##ɽ": 1018,
+      "##ʁ": 1019,
+      "##ტ": 1020,
+      "##ჱ": 1021,
+      "##ェ": 1022,
+      "##ハ": 1023,
+      "##υ": 1024,
+      "##र": 1025,
+      "##х": 1026,
+      "##も": 1027,
+      "##っ": 1028,
+      "##ょ": 1029,
+      "##に": 1030,
+      "##γ": 1031,
+      "##ც": 1032,
+      "##ე": 1033,
+      "##є": 1034,
+      "##м": 1035,
+      "##ܕ": 1036,
+      "##ܝ": 1037,
+      "##ܢ": 1038,
+      "##ܬ": 1039,
+      "##ณ": 1040,
+      "##ม": 1041,
+      "##ฮ": 1042,
+      "##ж": 1043,
+      "##ם": 1044,
+      "##ء": 1045,
+      "##ʊ": 1046,
+      "##ई": 1047,
+      "##め": 1048,
+      "##მ": 1049,
+      "##ム": 1050,
+      "##チ": 1051,
+      "##ᵻ": 1052,
+      "##ˌ": 1053,
+      "##ו": 1054,
+      "##ף": 1055,
+      "##წ": 1056,
+      "##ფ": 1057,
+      "##ャ": 1058,
+      "##モ": 1059,
+      "##ɐ": 1060,
+      "##ᅦ": 1061,
+      "##ᅩ": 1062,
+      "##ᆨ": 1063,
+      "##ᅵ": 1064,
+      "##ᆸ": 1065,
+      "##ᅧ": 1066,
+      "##ᆼ": 1067,
+      "##ᄋ": 1068,
+      "##ᆫ": 1069,
+      "##わ": 1070,
+      "##ı": 1071,
+      "##ქ": 1072,
+      "##დ": 1073,
+      "##ि": 1074,
+      "##ჲ": 1075,
+      "##ר": 1076,
+      "##セ": 1077,
+      "##オ": 1078,
+      "##ゆ": 1079,
+      "##せ": 1080,
+      "##ك": 1081,
+      "##ʿ": 1082,
+      "##ש": 1083,
+      "##מ": 1084,
+      "##צ": 1085,
+      "##п": 1086,
+      "##г": 1087,
+      "##カ": 1088,
+      "##ܠ": 1089,
+      "##ܗ": 1090,
+      "##ܐ": 1091,
+      "##ナ": 1092,
+      "##ミ": 1093,
+      "##こ": 1094,
+      "##を": 1095,
+      "##ψ": 1096,
+      "##サ": 1097,
+      "##ォ": 1098,
+      "##π": 1099,
+      "##ト": 1100,
+      "##у": 1101,
+      "##ح": 1102,
+      "##σ": 1103,
+      "##เ": 1104,
+      "##ป": 1105,
+      "##ш": 1106,
+      "##ゥ": 1107,
+      "##ロ": 1108,
+      "##া": 1109,
+      "##হ": 1110,
+      "##ɜ": 1111,
+      "##ة": 1112,
+      "##ص": 1113,
+      "##ס": 1114,
+      "##ث": 1115,
+      "##ჳ": 1116,
+      "##נ": 1117,
+      "##ذ": 1118,
+      "##ग": 1119,
+      "##ɫ": 1120,
+      "##ц": 1121,
+      "##ь": 1122,
+      "##ю": 1123
+    }
+  }
+}

tokenizer_config.json

@@ -0,0 +1,16 @@
+{
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "name_or_path": "temp/dummy/bert/processors",
+  "never_split": null,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "special_tokens_map_file": null,
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}

vocab.txt

@@ -0,0 +1,1124 @@
+[PAD]
+[UNK]
+[CLS]
+[SEP]
+[MASK]
+!
+"
+#
+$
+%
+&
+'
+(
+)
+*
++
+,
+-
+.
+/
+0
+1
+2
+3
+4
+5
+6
+7
+8
+9
+:
+;
+<
+=
+>
+?
+@
+[
+\
+]
+^
+_
+`
+a
+b
+c
+d
+e
+f
+g
+h
+i
+j
+k
+l
+m
+n
+o
+p
+q
+r
+s
+t
+u
+v
+w
+x
+y
+z
+|
+}
+~
+¡
+¢
+£
+¥
+§
+°
+±
+²
+³
+´
+µ
+·
+º
+½
+¿
+×
+ß
+æ
+ð
+ø
+þ
+đ
+ħ
+ı
+ł
+œ
+ɐ
+ɑ
+ɒ
+ɔ
+ə
+ɛ
+ɜ
+ɡ
+ɢ
+ɪ
+ɫ
+ɳ
+ɽ
+ɾ
+ʁ
+ʃ
+ʊ
+ʋ
+ʒ
+ʔ
+ʕ
+ʲ
+ʻ
+ʼ
+ʾ
+ʿ
+ˈ
+ˌ
+ː
+α
+β
+γ
+δ
+ε
+η
+θ
+ι
+κ
+λ
+μ
+ν
+ξ
+ο
+π
+ρ
+ς
+σ
+τ
+υ
+φ
+χ
+ψ
+ω
+а
+б
+в
+г
+д
+е
+ж
+з
+и
+к
+л
+м
+н
+о
+п
+р
+с
+т
+у
+х
+ц
+ш
+ъ
+ы
+ь
+ю
+я
+є
+א
+ב
+ג
+ה
+ו
+ז
+ח
+י
+ל
+ם
+מ
+ן
+נ
+ס
+ף
+פ
+צ
+ר
+ש
+ת
+ء
+ا
+ب
+ة
+ت
+ث
+ج
+ح
+خ
+د
+ذ
+ر
+س
+ش
+ص
+ع
+ف
+ق
+ك
+ل
+م
+ن
+ه
+و
+ي
+ܐ
+ܕ
+ܗ
+ܝ
+ܠ
+ܢ
+ܬ
+अ
+ई
+क
+ग
+ण
+त
+द
+न
+प
+ब
+म
+य
+र
+ल
+व
+स
+ह
+ा
+ि
+আ
+ল
+হ
+া
+ਅ
+ਲ
+ਹ
+ਾ
+അ
+ള
+ഹ
+ാ
+ก
+ค
+ง
+ช
+ซ
+ญ
+ฐ
+ณ
+ด
+ต
+น
+บ
+ป
+พ
+ภ
+ม
+ย
+ร
+ล
+ว
+ศ
+ษ
+ส
+ห
+อ
+ฮ
+ะ
+า
+เ
+แ
+ไ
+ა
+ბ
+გ
+დ
+ე
+ვ
+ზ
+თ
+ი
+კ
+ლ
+მ
+ნ
+ო
+პ
+ჟ
+რ
+ს
+ტ
+უ
+ფ
+ქ
+ღ
+ყ
+შ
+ჩ
+ც
+ძ
+წ
+ჭ
+ხ
+ჯ
+ჰ
+ჱ
+ჲ
+ჳ
+ჴ
+ჵ
+ჶ
+ჷ
+ჸ
+ჹ
+ჺ
+჻
+ᄃ
+ᄅ
+ᄇ
+ᄋ
+ᄌ
+ᅡ
+ᅢ
+ᅦ
+ᅧ
+ᅩ
+ᅮ
+ᅵ
+ᆨ
+ᆫ
+ᆯ
+ᆸ
+ᆼ
+ᵻ
+‐
+‑
+–
+—
+―
+‘
+’
+“
+”
+„
+†
+‡
+•
+…
+′
+″
+⁄
+₣
+₤
+€
+₹
+⅓
+⅔
+→
+−
+≡
+≤
+①
+☉
+☫
+♀
+♭
+♯
+⚳
+ⴀ
+ⴂ
+ⴃ
+ⴈ
+ⴌ
+ⴕ
+ⴟ
+〈
+〉
+〜
+あ
+い
+う
+お
+か
+き
+く
+け
+こ
+さ
+し
+す
+せ
+た
+ち
+っ
+つ
+と
+な
+に
+の
+は
+ひ
+ふ
+ほ
+ま
+み
+め
+も
+ゃ
+ゆ
+ょ
+ら
+り
+る
+れ
+わ
+を
+ん
+ァ
+ア
+ィ
+イ
+ゥ
+ウ
+ェ
+エ
+ォ
+オ
+カ
+キ
+ク
+ケ
+コ
+サ
+シ
+ス
+セ
+タ
+チ
+ッ
+ツ
+テ
+ト
+ナ
+ニ
+ネ
+ノ
+ハ
+フ
+ヘ
+マ
+ミ
+ム
+モ
+ャ
+ュ
+ョ
+ラ
+リ
+ル
+レ
+ロ
+ン
+・
+ー
+一
+七
+下
+世
+丙
+中
+主
+乃
+之
+乙
+九
+二
+云
+人
+今
+付
+作
+侗
+依
+信
+傳
+儚
+充
+光
+全
+兵
+其
+具
+円
+再
+出
+判
+前
+剛
+劇
+劉
+動
+化
+北
+华
+厂
+去
+古
+可
+台
+史
+同
+名
+君
+吳
+周
+命
+和
+咲
+善
+四
+國
+園
+圣
+在
+坂
+堤
+場
+塘
+夕
+大
+天
+夫
+女
+妙
+姚
+子
+孟
+守
+安
+宋
+完
+宗
+宝
+宫
+寝
+寺
+小
+少
+尾
+山
+岳
+川
+州
+巳
+市
+師
+平
+广
+庆
+府
+座
+廬
+建
+式
+張
+彌
+彩
+彼
+後
+御
+德
+思
+愛
+憑
+憶
+應
+懷
+战
+戦
+扈
+技
+拉
+拳
+挑
+揺
+攻
+放
+政
+散
+斯
+方
+日
+旦
+旭
+昌
+明
+星
+春
+晋
+景
+曦
+月
+望
+未
+本
+李
+村
+杜
+束
+来
+林
+桜
+梶
+棘
+椎
+楊
+楚
+榮
+橘
+機
+正
+殻
+殿
+母
+水
+汉
+沂
+沙
+河
+泗
+波
+泣
+洪
+淹
+清
+湯
+漢
+澄
+澤
+火
+灯
+灵
+灼
+焼
+熱
+物
+狐
+狸
+玄
+王
+玩
+珂
+珙
+球
+理
+琦
+琪
+瓊
+生
+田
+畢
+番
+瘡
+白
+皮
+真
+砲
+礮
+祈
+神
+祠
+秋
+空
+立
+精
+約
+絵
+織
+義
+翠
+者
+耕
+肖
+胡
+膀
+臂
+興
+良
+花
+芳
+芽
+若
+英
+藕
+藥
+蘄
+蘇
+行
+裁
+規
+覺
+观
+解
+記
+誓
+誡
+誰
+謎
+许
+谭
+豪
+豫
+費
+贵
+赤
+趙
+足
+跡
+転
+辛
+逆
+遇
+運
+過
+遠
+選
+邦
+邱
+部
+郭
+都
+酈
+里
+野
+金
+銃
+鋼
+錄
+錡
+鍵
+鐵
+钱
+铁
+關
+防
+阿
+陈
+陳
+陽
+隊
+階
+集
+雪
+雲
+霖
+霹
+靂
+韓
+願
+顯
+颜
+马
+高
+龍
+ﷲ
+ﻋ
+／
+３
+～
+##i
+##y
+##o
+##r
+##g
+##a
+##w
+##l
+##b
+##z
+##t
+##n
+##c
+##h
+##s
+##u
+##d
+##e
+##k
+##v
+##f
+##x
+##q
+##p
+##æ
+##0
+##5
+##m
+##8
+##4
+##س
+##ت
+##ا
+##ن
+##6
+##1
+##7
+##j
+##つ
+##う
+##2
+##9
+##3
+##ø
+##ล
+##ว
+##ง
+##พ
+##ไ
+##ช
+##ย
+##า
+##ร
+##თ
+##ა
+##ვ
+##რ
+##ი
+##ള
+##あ
+##ん
+##α
+##ν
+##τ
+##ο
+##κ
+##ρ
+##ω
+##ς
+##の
+##な
+##ら
+##ð
+##œ
+##ɛ
+##ł
+##η
+##μ
+##ซ
+##ル
+##シ
+##ア
+##リ
+##ス
+##ʔ
+##ल
+##ᄇ
+##ᅮ
+##ᄃ
+##ᅢ
+##β
+##ß
+##か
+##た
+##ə
+##ʻ
+##ι
+##χ
+##о
+##л
+##с
+##а
+##т
+##ы
+##и
+##в
+##к
+##з
+##ッ
+##ク
+##マ
+##ン
+##გ
+##ლ
+##ო
+##ნ
+##ː
+##ל
+##ה
+##א
+##く
+##み
+##ε
+##ξ
+##ল
+##ˈ
+##ɡ
+##ɑ
+##ɒ
+##し
+##す
+##き
+##ひ
+##と
+##đ
+##ъ
+##н
+##е
+##י
+##פ
+##イ
+##λ
+##ق
+##ع
+##د
+##ᅡ
+##ᆯ
+##ᄅ
+##ɪ
+##ค
+##ต
+##व
+##ा
+##द
+##は
+##り
+##レ
+##ー
+##ツ
+##ي
+##ش
+##و
+##م
+##º
+##ਲ
+##ਾ
+##ਹ
+##д
+##р
+##ل
+##ب
+##い
+##ち
+##ゃ
+##ʒ
+##ʃ
+##ɔ
+##ह
+##ニ
+##ウ
+##ァ
+##キ
+##ュ
+##３
+##ხ
+##ს
+##お
+##タ
+##ാ
+##ഹ
+##ɳ
+##ま
+##る
+##ะ
+##อ
+##น
+##ן
+##я
+##แ
+##ก
+##ɾ
+##ʲ
+##フ
+##უ
+##ภ
+##ด
+##ב
+##ת
+##خ
+##ラ
+##れ
+##ण
+##स
+##न
+##ه
+##ف
+##ر
+##エ
+##テ
+##ษ
+##ฐ
+##ィ
+##क
+##ノ
+##θ
+##ネ
+##��
+##δ
+##ɽ
+##ʁ
+##ტ
+##ჱ
+##ェ
+##ハ
+##υ
+##र
+##х
+##も
+##っ
+##ょ
+##に
+##γ
+##ც
+##ე
+##є
+##м
+##ܕ
+##ܝ
+##ܢ
+##ܬ
+##ณ
+##ม
+##ฮ
+##ж
+##ם
+##ء
+##ʊ
+##ई
+##め
+##მ
+##ム
+##チ
+##ᵻ
+##ˌ
+##ו
+##ף
+##წ
+##ფ
+##ャ
+##モ
+##ɐ
+##ᅦ
+##ᅩ
+##ᆨ
+##ᅵ
+##ᆸ
+##ᅧ
+##ᆼ
+##ᄋ
+##ᆫ
+##わ
+##ı
+##ქ
+##დ
+##ि
+##ჲ
+##ר
+##セ
+##オ
+##ゆ
+##せ
+##ك
+##ʿ
+##ש
+##מ
+##צ
+##п
+##г
+##カ
+##ܠ
+##ܗ
+##ܐ
+##ナ
+##ミ
+##こ
+##を
+##ψ
+##サ
+##ォ
+##π
+##ト
+##у
+##ح
+##σ
+##เ
+##ป
+##ш
+##ゥ
+##ロ
+##া
+##হ
+##ɜ
+##ة
+##ص
+##ס
+##ث
+##ჳ
+##נ
+##ذ
+##ग
+##ɫ
+##ц
+##ь
+##ю