add tiny random model

config.json

@@ -0,0 +1,43 @@
+{
+  "_name_or_path": "hf-internal-testing/tiny-random-gpt2",
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2CustomLMHeadModel"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "attn_pdrop": 0.1,
+  "auto_map": {
+    "AutoModelForCausalLM": "modeling_gpt2.GPT2CustomLMHeadModel"
+  },
+  "bos_token_id": 98,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 98,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "initializer_range": 0.02,
+  "intermediate_size": 37,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 512,
+  "n_embd": 32,
+  "n_head": 4,
+  "n_inner": null,
+  "n_layer": 5,
+  "n_positions": 512,
+  "pad_token_id": 98,
+  "reorder_and_upcast_attn": false,
+  "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.25.1",
+  "type_vocab_size": 16,
+  "use_cache": true,
+  "vocab_size": 1000
+}

create_model.py

@@ -0,0 +1,10 @@
+from transformers import AutoConfig
+
+from modeling.modeling_gpt2 import GPT2CustomLMHeadModel
+
+cfg = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-gpt2")
+
+GPT2CustomLMHeadModel.register_for_auto_class("AutoModelForCausalLM")
+
+model = GPT2CustomLMHeadModel(cfg)
+model.save_pretrained("/home/fxmarty/hf_internship/tiny-testing-gpt2-remote-code")

merges.txt

@@ -0,0 +1,807 @@
+#version: 0.2 - Trained by `huggingface/tokenizers`
+Ġ t
+h e
+Ġ a
+i n
+Ġt he
+e r
+o n
+Ġ ,
+r e
+Ġ s
+e d
+Ġ o
+Ġ w
+n d
+a t
+Ġ .
+o r
+i t
+Ġ c
+e n
+Ġ f
+i s
+e s
+a r
+Ġo f
+Ġ b
+a n
+Ġ in
+a l
+in g
+Ġ p
+Ġa nd
+a s
+Ġt o
+r o
+i c
+Ġ m
+Ġ d
+Ġ h
+i on
+l e
+o u
+Ġ T
+Ġ re
+Ġ =
+Ġ "
+Ġ A
+Ġ S
+en t
+i l
+Ġt h
+Ġ 1
+s t
+Ġ C
+e l
+o m
+Ġ l
+a m
+Ġ Ċ
+Ġ e
+Ġ n
+Ġ @
+a d
+a c
+Ġw as
+Ġ M
+u r
+ĠT he
+e c
+Ġ on
+l y
+Ġ B
+Ġ I
+Ġ g
+Ġ '
+e t
+o l
+i d
+i v
+i m
+Ġf or
+i r
+- @
+Ġ@ -@
+i g
+o t
+t er
+Ġa s
+Ġ H
+u s
+o w
+Ġs t
+u t
+it h
+a y
+Ġ 2
+Ġ P
+at ion
+v er
+Ġb e
+he r
+Ġth at
+Ġw ith
+Ġ R
+c e
+t h
+Ġ D
+Ġ is
+u n
+e m
+Ġ F
+Ġw h
+u l
+Ġb y
+Ġa l
+c h
+Ġ )
+Ġ (
+Ġ W
+Ġc on
+r a
+Ġ G
+o s
+Ġ L
+Ġ N
+Ġa t
+er s
+c t
+Ġ it
+Ġ1 9
+ro m
+a nd
+Ġa n
+u m
+es t
+Ġ J
+a g
+Ġ he
+0 0
+is t
+a in
+o d
+a v
+r i
+Ġ E
+Ġ O
+Ġf rom
+Ġc om
+Ġh is
+o p
+Ġp ro
+re s
+i es
+i f
+Ġ v
+or t
+er e
+il l
+l d
+Ġd e
+p p
+Ġs u
+o re
+ĠI n
+Ġ r
+Ġs e
+Ġw ere
+e w
+on g
+ig h
+ar d
+at e
+al l
+ar t
+a k
+ic h
+Ġc h
+Ġo r
+a b
+an t
+u d
+o c
+b er
+Ġe x
+g h
+it y
+at ed
+p t
+es s
+e ar
+Ġ K
+Ġp l
+am e
+q u
+iv e
+ro u
+Ġa re
+Ġ â
+Ġs h
+Ġ k
+ac k
+ec t
+Ġâ Ģ
+Ġ U
+Ġh ad
+s e
+Ġwh ich
+re d
+o v
+ĠS t
+as t
+Ġs p
+i an
+Ġ y
+m ent
+Ġ le
+Ġn ot
+g e
+or d
+r it
+i p
+in e
+el l
+al ly
+ou r
+o st
+igh t
+t her
+a p
+Ġ u
+is h
+ĠC h
+ou n
+i a
+Ġ 3
+av e
+ar y
+u st
+o g
+Ġ2 00
+Ġ un
+ou s
+ir st
+Ġ V
+c c
+Ġin c
+Ġ ;
+Ġcom p
+r u
+ion s
+Ġthe ir
+Ġb ut
+id e
+u re
+s o
+Ġcon t
+Ġin t
+f ter
+ic al
+i al
+Ġa r
+Ġf irst
+ou ld
+Ġit s
+he d
+ĠâĢ ĵ
+Ġw he
+w o
+ou t
+u b
+Ġ2 0
+f f
+Ġ :
+u e
+Ġ her
+ow n
+o k
+Ġal so
+Ġc l
+p er
+ig n
+at er
+r an
+or m
+i e
+om e
+or k
+as s
+i re
+e nd
+Ġre s
+Ġa b
+Ġa d
+Ġ us
+r y
+Ġre c
+Ġh ave
+ag e
+ĠH e
+Ġ 4
+Ġ ro
+m er
+Ġon e
+on d
+l ow
+Ġh as
+ĠT h
+d u
+Ġ 5
+Ġp er
+Ġbe en
+im e
+Ġt wo
+en ce
+l and
+Ġ1 8
+. @
+Ġ@ .@
+ul t
+re e
+ou gh
+i le
+Ġwh o
+ĠA l
+Ġs c
+ur ing
+p l
+or y
+it ion
+r ic
+ation s
+Ġd is
+Ġth is
+Ġb ec
+Ġa pp
+i z
+ĠI t
+a re
+ac h
+l ud
+ad e
+Ġpl ay
+Ġ j
+Ġm an
+ac t
+el y
+Ġp art
+Ġd es
+Ġa g
+Ġthe y
+Ġy ear
+oun t
+Ġ20 1
+Ġo ver
+Ġo ther
+ou nd
+Ġa fter
+i b
+o ver
+Ġs er
+Ġ en
+Ġof f
+Ġ im
+ct ion
+Ġ Y
+k e
+it e
+, @
+Ġ@ ,@
+t e
+ur n
+Ġinc lud
+res s
+an ce
+an g
+Ġat t
+ic e
+ac e
+ar k
+Ġo ut
+w n
+p h
+em ber
+Ġp re
+Ġu p
+en s
+m an
+Ġe v
+Ġt ime
+nd er
+rou gh
+c ed
+Ġf in
+Ġint o
+on e
+p ort
+rou nd
+w e
+re n
+l es
+in t
+ĠO n
+v el
+Ġcom m
+Ġs he
+as on
+am p
+Ġt e
+Ġw ould
+w ard
+Ġm ore
+Ġ 6
+i ed
+os e
+ri b
+ĠU n
+Ġal l
+ing s
+ter n
+c es
+ab le
+Ġw e
+it ed
+e ver
+ent s
+Ġh im
+as ed
+or s
+o y
+o od
+Ġc ent
+i x
+as e
+il d
+ĠA n
+Ġ 7
+Ġw ork
+at es
+i ous
+at h
+Ġp o
+ro p
+ol d
+al s
+is s
+e y
+ic t
+Ġf e
+Ġthe m
+g an
+Ġs ec
+Ġb et
+Ġwhe n
+Ġs ong
+Ġre m
+e p
+f orm
+a il
+f er
+Ġe ar
+ub l
+a w
+Ġk n
+ak e
+a us
+Ġm ost
+Ġcon s
+Ġd uring
+ĠA s
+or th
+Ġn ew
+er ed
+il m
+v ed
+at t
+Ġon ly
+Ġ 9
+Ġd ec
+Ġ 8
+ic k
+Ġg ame
+on s
+u g
+Ġt r
+f t
+ot h
+o ok
+ĠM ar
+re at
+w ay
+Ġc an
+ol low
+ou th
+we en
+ĠE n
+Ġ19 9
+ter s
+Ġre l
+in d
+Ġab out
+Ġse ason
+Ġag ain
+r al
+Ġth ree
+ation al
+Ġu nder
+ul ar
+Ġm e
+Ġth an
+ĠC om
+ĠA r
+h ip
+o b
+Ġn e
+Ġbet ween
+Ġf l
+h n
+v e
+Ġch ar
+Ġc ol
+Ġrec ord
+i ew
+r on
+f ore
+Ġth rough
+is ion
+or n
+Ġ 00
+oc k
+Ġ ver
+Ġl ater
+Ġn um
+Ġe nd
+ol og
+am es
+Ġp os
+Ġw rit
+Ġpro du
+Ġwh ile
+Ġa ct
+Ġre le
+Ġf ilm
+is hed
+Ġp r
+an s
+Ġre g
+Ġfor m
+Ġas s
+ĠS e
+ur y
+t ed
+t s
+Ġm ade
+Ġsu b
+Ġp e
+Ġs o
+or ld
+Ġre t
+ĠN ew
+Ġsp ec
+Ġa cc
+Ġ qu
+Ġwhe re
+en er
+Ġm ov
+he s
+mer ic
+at ing
+Ġin ter
+ĠL e
+ĠA meric
+Ġ ra
+Ġs ome
+Ġc o
+Ġl ar
+Ġb u
+Ġde f
+b um
+Ġa c
+Ġm us
+Ġf ollow
+ĠA t
+in s
+iv ed
+if ic
+u al
+Ġa m
+Ġsu ch
+Ġsec ond
+i ke
+Ġf our
+Ġin d
+an n
+he n
+Ġus ed
+ĠR e
+ic s
+le ct
+Ġd ay
+i el
+il y
+ĠTh is
+Ġ 0
+Ġp ubl
+Ġc all
+ĠJ o
+l l
+Ġal bum
+Ġ00 0
+ran s
+Ġd o
+an y
+Ġbe fore
+ro s
+ĠS h
+Ġs y
+a id
+ĠEn g
+Ġbe ing
+Ġ1 0
+u c
+Ġe p
+Ġsu pp
+Ġthe re
+Ġyear s
+ar s
+ow ever
+Ġ ent
+if e
+Ġh igh
+Ġf ound
+ir d
+Ġn o
+Ġs et
+in es
+iv er
+i o
+ot her
+j ect
+Ġs ur
+a j
+t en
+Ġt ra
+Ġ1 2
+is ed
+it ies
+vel op
+Ġb l
+al e
+Ġser ies
+Ġl oc
+Ġnum ber
+Ġp res
+an e
+aus e
+od e
+e k
+t on
+ĠS c
+i er
+is e
+Ġse ver
+in ce
+Ġb oth
+an k
+ro w
+ire ct
+s on
+Ġthe n
+ĠB rit
+i et
+Ġ1 6
+Ġep is
+Ġinclud ing
+it s
+ig in
+p r
+Ġ /
+Ġagain st
+Ġw ell
+Ġbec ame
+Ġex p
+Ġkn own
+Ġt rans
+Ġchar ac
+ĠâĢ Ķ
+r am
+Ġb ack
+Ġad d
+Ġp op
+Ġg o
+ur ch
+Ġdes c
+Ġs ing
+iel d
+Ġper form
+ain ed
+Ġre ce
+id ent
+Ġe m
+er t
+u res
+Ġin v
+Ġde p
+Ġ19 8
+a ir
+er n
+at her
+f ul
+Ġ Z
+Ġm on
+Ġman y
+Ġm ain
+Ġst ud
+Ġl ong
+in n
+th ough
+u p
+o ol
+ĠUn ited
+l ed
+em ent
+Ġ1 5
+ow er
+ĠJo hn
+Ġo p
+Ġ1 1
+in ed
+Ġm et
+o ber
+le y
+Ġ1 7
+Ġcent ury
+Ġte am
+Ġ est
+ĠA fter
+y l
+Ġm in
+u ch
+ut e
+Ġde velop
+ĠS he
+i am
+Ġsh ow
+el f
+Ġre p
+Ġcon c
+at ive
+Ġc re
+over n
+a red
+Ġ19 4
+Ġor igin
+Ġs m
+iv ers
+a z
+Ġle ad
+Ġsever al
+a h
+Ġo b
+Ġre v
+Ġm ill
+er m
+u ally
+o ot
+Ġbe gan
+Ġ19 6
+i red
+Ġd if
+Ġcont in
+Ġs ign
+i k
+ĠI nd
+ment s
+iz ed
+Ġ19 7
+Ġd irect
+a u
+Ġex t
+ros s
+em b
+d er
+Ġp ol
+Ġm ay
+a pt
+el s
+ĠW h
+Ġcomp le
+Ġar t
+ĠB r
+ĠI s
+un e
+t il
+Ġc rit
+Ġh ist
+Ġear ly
+Ġc ould
+ĠC on
+Ġd id
+Ġb el
+Ġcall ed
+u ed
+Ġn ear
+Ġepis ode
+y p
+Ġdesc rib

modeling_gpt2.py

@@ -0,0 +1,1548 @@
+# coding=utf-8
+# Copyright 2018 The OpenAI Team Authors and 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 OpenAI GPT-2 model."""
+
+import math
+import os
+from dataclasses import dataclass
+from typing import Optional, Tuple, Union
+from transformers import GenerationMixin
+
+import torch
+import torch.utils.checkpoint
+from torch import nn
+from torch.cuda.amp import autocast
+from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
+
+from transformers.activations import ACT2FN
+from transformers.modeling_outputs import (
+    BaseModelOutputWithPastAndCrossAttentions,
+    CausalLMOutputWithCrossAttentions,
+    SequenceClassifierOutputWithPast,
+    TokenClassifierOutput,
+)
+from transformers.modeling_utils import PreTrainedModel, SequenceSummary
+from transformers.pytorch_utils import Conv1D, find_pruneable_heads_and_indices, prune_conv1d_layer
+from transformers.utils import (
+    ModelOutput,
+    add_code_sample_docstrings,
+    add_start_docstrings,
+    add_start_docstrings_to_model_forward,
+    logging,
+    replace_return_docstrings,
+)
+from transformers.utils.model_parallel_utils import assert_device_map, get_device_map
+from transformers import GPT2Config
+
+
+logger = logging.get_logger(__name__)
+
+_CHECKPOINT_FOR_DOC = "gpt2"
+_CONFIG_FOR_DOC = "GPT2Config"
+_TOKENIZER_FOR_DOC = "GPT2Tokenizer"
+
+GPT2_PRETRAINED_MODEL_ARCHIVE_LIST = [
+    "gpt2",
+    "gpt2-medium",
+    "gpt2-large",
+    "gpt2-xl",
+    "distilgpt2",
+    # See all GPT-2 models at https://huggingface.co/models?filter=gpt2
+]
+
+
+def load_tf_weights_in_gpt2(model, config, gpt2_checkpoint_path):
+    """Load tf checkpoints in a pytorch model"""
+    try:
+        import re
+
+        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(gpt2_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.squeeze())
+
+    for name, array in zip(names, arrays):
+        name = name[6:]  # skip "model/"
+        name = name.split("/")
+        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] == "w" or scope_names[0] == "g":
+                pointer = getattr(pointer, "weight")
+            elif scope_names[0] == "b":
+                pointer = getattr(pointer, "bias")
+            elif scope_names[0] == "wpe" or scope_names[0] == "wte":
+                pointer = getattr(pointer, scope_names[0])
+                pointer = getattr(pointer, "weight")
+            else:
+                pointer = getattr(pointer, scope_names[0])
+            if len(scope_names) >= 2:
+                num = int(scope_names[1])
+                pointer = pointer[num]
+        try:
+            assert (
+                pointer.shape == array.shape
+            ), 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 GPT2Attention(nn.Module):
+    def __init__(self, config, is_cross_attention=False, layer_idx=None):
+        super().__init__()
+
+        max_positions = config.max_position_embeddings
+        self.register_buffer(
+            "bias",
+            torch.tril(torch.ones((max_positions, max_positions), dtype=torch.uint8)).view(
+                1, 1, max_positions, max_positions
+            ),
+        )
+        self.register_buffer("masked_bias", torch.tensor(-1e4))
+
+        self.embed_dim = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = self.embed_dim // self.num_heads
+        self.split_size = self.embed_dim
+        if self.head_dim * self.num_heads != self.embed_dim:
+            raise ValueError(
+                f"`embed_dim` must be divisible by num_heads (got `embed_dim`: {self.embed_dim} and `num_heads`:"
+                f" {self.num_heads})."
+            )
+
+        self.scale_attn_weights = config.scale_attn_weights
+        self.is_cross_attention = is_cross_attention
+
+        # Layer-wise attention scaling, reordering, and upcasting
+        self.scale_attn_by_inverse_layer_idx = config.scale_attn_by_inverse_layer_idx
+        self.layer_idx = layer_idx
+        self.reorder_and_upcast_attn = config.reorder_and_upcast_attn
+
+        if self.is_cross_attention:
+            self.c_attn = Conv1D(2 * self.embed_dim, self.embed_dim)
+            self.q_attn = Conv1D(self.embed_dim, self.embed_dim)
+        else:
+            self.c_attn = Conv1D(3 * self.embed_dim, self.embed_dim)
+        self.c_proj = Conv1D(self.embed_dim, self.embed_dim)
+
+        self.attn_dropout = nn.Dropout(config.attn_pdrop)
+        self.resid_dropout = nn.Dropout(config.resid_pdrop)
+
+        self.pruned_heads = set()
+
+    def prune_heads(self, heads):
+        if len(heads) == 0:
+            return
+        heads, index = find_pruneable_heads_and_indices(heads, self.num_heads, self.head_dim, self.pruned_heads)
+        index_attn = torch.cat([index, index + self.split_size, index + (2 * self.split_size)])
+
+        # Prune conv1d layers
+        self.c_attn = prune_conv1d_layer(self.c_attn, index_attn, dim=1)
+        self.c_proj = prune_conv1d_layer(self.c_proj, index, dim=0)
+
+        # Update hyper params
+        self.split_size = (self.split_size // self.num_heads) * (self.num_heads - len(heads))
+        self.num_heads = self.num_heads - len(heads)
+        self.pruned_heads = self.pruned_heads.union(heads)
+
+    def _attn(self, query, key, value, attention_mask=None, head_mask=None):
+        attn_weights = torch.matmul(query, key.transpose(-1, -2))
+
+        if self.scale_attn_weights:
+            attn_weights = attn_weights / torch.full(
+                [], value.size(-1) ** 0.5, dtype=attn_weights.dtype, device=attn_weights.device
+            )
+
+        # Layer-wise attention scaling
+        if self.scale_attn_by_inverse_layer_idx:
+            attn_weights = attn_weights / float(self.layer_idx + 1)
+
+        if not self.is_cross_attention:
+            # if only "normal" attention layer implements causal mask
+            query_length, key_length = query.size(-2), key.size(-2)
+            causal_mask = self.bias[:, :, key_length - query_length : key_length, :key_length].to(torch.bool)
+            mask_value = torch.finfo(attn_weights.dtype).min
+            # Need to be a tensor, otherwise we get error: `RuntimeError: expected scalar type float but found double`.
+            # Need to be on the same device, otherwise `RuntimeError: ..., x and y to be on the same device`
+            mask_value = torch.full([], mask_value, dtype=attn_weights.dtype).to(attn_weights.device)
+            attn_weights = torch.where(causal_mask, attn_weights, mask_value)
+
+        if attention_mask is not None:
+            # Apply the attention mask
+            attn_weights = attn_weights + attention_mask
+
+        attn_weights = nn.functional.softmax(attn_weights, dim=-1)
+
+        # Downcast (if necessary) back to V's dtype (if in mixed-precision) -- No-Op otherwise
+        attn_weights = attn_weights.type(value.dtype)
+        attn_weights = self.attn_dropout(attn_weights)
+
+        # Mask heads if we want to
+        if head_mask is not None:
+            attn_weights = attn_weights * head_mask
+
+        attn_output = torch.matmul(attn_weights, value)
+
+        return attn_output, attn_weights
+
+    def _upcast_and_reordered_attn(self, query, key, value, attention_mask=None, head_mask=None):
+        # Use `torch.baddbmm` (a bit more efficient w/ alpha param for scaling -- from Megatron-LM)
+        bsz, num_heads, q_seq_len, dk = query.size()
+        _, _, k_seq_len, _ = key.size()
+
+        # Preallocate attn_weights for `baddbmm`
+        attn_weights = torch.empty(bsz * num_heads, q_seq_len, k_seq_len, dtype=torch.float32, device=query.device)
+
+        # Compute Scale Factor
+        scale_factor = 1.0
+        if self.scale_attn_weights:
+            scale_factor /= float(value.size(-1)) ** 0.5
+
+        if self.scale_attn_by_inverse_layer_idx:
+            scale_factor /= float(self.layer_idx + 1)
+
+        # Upcast (turn off autocast) and reorder (Scale K by 1 / root(dk))
+        with autocast(enabled=False):
+            q, k = query.reshape(-1, q_seq_len, dk), key.transpose(-1, -2).reshape(-1, dk, k_seq_len)
+            attn_weights = torch.baddbmm(attn_weights, q.float(), k.float(), beta=0, alpha=scale_factor)
+            attn_weights = attn_weights.reshape(bsz, num_heads, q_seq_len, k_seq_len)
+
+        if not self.is_cross_attention:
+            # if only "normal" attention layer implements causal mask
+            query_length, key_length = query.size(-2), key.size(-2)
+            causal_mask = self.bias[:, :, key_length - query_length : key_length, :key_length].bool()
+            mask_value = torch.finfo(attn_weights.dtype).min
+            # Need to be a tensor, otherwise we get error: `RuntimeError: expected scalar type float but found double`.
+            # Need to be on the same device, otherwise `RuntimeError: ..., x and y to be on the same device`
+            mask_value = torch.tensor(mask_value, dtype=attn_weights.dtype).to(attn_weights.device)
+            attn_weights = torch.where(causal_mask, attn_weights, mask_value)
+
+        if attention_mask is not None:
+            # Apply the attention mask
+            attn_weights = attn_weights + attention_mask
+
+        attn_weights = nn.functional.softmax(attn_weights, dim=-1)
+
+        # Downcast (if necessary) back to V's dtype (if in mixed-precision) -- No-Op if otherwise
+        if attn_weights.dtype != torch.float32:
+            raise RuntimeError("Error with upcasting, attn_weights does not have dtype torch.float32")
+        attn_weights = attn_weights.type(value.dtype)
+        attn_weights = self.attn_dropout(attn_weights)
+
+        # Mask heads if we want to
+        if head_mask is not None:
+            attn_weights = attn_weights * head_mask
+
+        attn_output = torch.matmul(attn_weights, value)
+
+        return attn_output, attn_weights
+
+    def _split_heads(self, tensor, num_heads, attn_head_size):
+        """
+        Splits hidden_size dim into attn_head_size and num_heads
+        """
+        new_shape = tensor.size()[:-1] + (num_heads, attn_head_size)
+        tensor = tensor.view(new_shape)
+        return tensor.permute(0, 2, 1, 3)  # (batch, head, seq_length, head_features)
+
+    def _merge_heads(self, tensor, num_heads, attn_head_size):
+        """
+        Merges attn_head_size dim and num_attn_heads dim into hidden_size
+        """
+        tensor = tensor.permute(0, 2, 1, 3).contiguous()
+        new_shape = tensor.size()[:-2] + (num_heads * attn_head_size,)
+        return tensor.view(new_shape)
+
+    def forward(
+        self,
+        hidden_states: Optional[Tuple[torch.FloatTensor]],
+        layer_past: Optional[Tuple[torch.Tensor]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        use_cache: Optional[bool] = False,
+        output_attentions: Optional[bool] = False,
+    ) -> Tuple[Union[torch.Tensor, Tuple[torch.Tensor]], ...]:
+        if encoder_hidden_states is not None:
+            if not hasattr(self, "q_attn"):
+                raise ValueError(
+                    "If class is used as cross attention, the weights `q_attn` have to be defined. "
+                    "Please make sure to instantiate class with `GPT2Attention(..., is_cross_attention=True)`."
+                )
+
+            query = self.q_attn(hidden_states)
+            key, value = self.c_attn(encoder_hidden_states).split(self.split_size, dim=2)
+            attention_mask = encoder_attention_mask
+        else:
+            query, key, value = self.c_attn(hidden_states).split(self.split_size, dim=2)
+
+        query = self._split_heads(query, self.num_heads, self.head_dim)
+        key = self._split_heads(key, self.num_heads, self.head_dim)
+        value = self._split_heads(value, self.num_heads, self.head_dim)
+
+        value = value + 10
+        print("increased value")
+
+        if layer_past is not None:
+            past_key, past_value = layer_past
+            key = torch.cat((past_key, key), dim=-2)
+            value = torch.cat((past_value, value), dim=-2)
+
+        if use_cache is True:
+            present = (key, value)
+        else:
+            present = None
+
+        if self.reorder_and_upcast_attn:
+            attn_output, attn_weights = self._upcast_and_reordered_attn(query, key, value, attention_mask, head_mask)
+        else:
+            attn_output, attn_weights = self._attn(query, key, value, attention_mask, head_mask)
+
+        attn_output = self._merge_heads(attn_output, self.num_heads, self.head_dim)
+        attn_output = self.c_proj(attn_output)
+        attn_output = self.resid_dropout(attn_output)
+
+        outputs = (attn_output, present)
+        if output_attentions:
+            outputs += (attn_weights,)
+
+        return outputs  # a, present, (attentions)
+
+
+class GPT2MLP(nn.Module):
+    def __init__(self, intermediate_size, config):
+        super().__init__()
+        embed_dim = config.hidden_size
+        self.c_fc = Conv1D(intermediate_size, embed_dim)
+        self.c_proj = Conv1D(embed_dim, intermediate_size)
+        self.act = ACT2FN[config.activation_function]
+        self.dropout = nn.Dropout(config.resid_pdrop)
+
+    def forward(self, hidden_states: Optional[Tuple[torch.FloatTensor]]) -> torch.FloatTensor:
+        hidden_states = self.c_fc(hidden_states)
+        hidden_states = self.act(hidden_states)
+        hidden_states = self.c_proj(hidden_states)
+        hidden_states = self.dropout(hidden_states)
+        return hidden_states
+
+
+class GPT2Block(nn.Module):
+    def __init__(self, config, layer_idx=None):
+        super().__init__()
+        hidden_size = config.hidden_size
+        inner_dim = config.n_inner if config.n_inner is not None else 4 * hidden_size
+
+        self.ln_1 = nn.LayerNorm(hidden_size, eps=config.layer_norm_epsilon)
+        self.attn = GPT2Attention(config, layer_idx=layer_idx)
+        self.ln_2 = nn.LayerNorm(hidden_size, eps=config.layer_norm_epsilon)
+
+        if config.add_cross_attention:
+            self.crossattention = GPT2Attention(config, is_cross_attention=True, layer_idx=layer_idx)
+            self.ln_cross_attn = nn.LayerNorm(hidden_size, eps=config.layer_norm_epsilon)
+
+        self.mlp = GPT2MLP(inner_dim, config)
+
+    def forward(
+        self,
+        hidden_states: Optional[Tuple[torch.FloatTensor]],
+        layer_past: Optional[Tuple[torch.Tensor]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        use_cache: Optional[bool] = False,
+        output_attentions: Optional[bool] = False,
+    ) -> Union[Tuple[torch.Tensor], Optional[Tuple[torch.Tensor, Tuple[torch.FloatTensor, ...]]]]:
+        residual = hidden_states
+        hidden_states = self.ln_1(hidden_states)
+        attn_outputs = self.attn(
+            hidden_states,
+            layer_past=layer_past,
+            attention_mask=attention_mask,
+            head_mask=head_mask,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+        )
+        attn_output = attn_outputs[0]  # output_attn: a, present, (attentions)
+        outputs = attn_outputs[1:]
+        # residual connection
+        hidden_states = attn_output + residual
+
+        if encoder_hidden_states is not None:
+            # add one self-attention block for cross-attention
+            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`"
+                )
+            residual = hidden_states
+            hidden_states = self.ln_cross_attn(hidden_states)
+            cross_attn_outputs = self.crossattention(
+                hidden_states,
+                attention_mask=attention_mask,
+                head_mask=head_mask,
+                encoder_hidden_states=encoder_hidden_states,
+                encoder_attention_mask=encoder_attention_mask,
+                output_attentions=output_attentions,
+            )
+            attn_output = cross_attn_outputs[0]
+            # residual connection
+            hidden_states = residual + attn_output
+            outputs = outputs + cross_attn_outputs[2:]  # add cross attentions if we output attention weights
+
+        residual = hidden_states
+        hidden_states = self.ln_2(hidden_states)
+        feed_forward_hidden_states = self.mlp(hidden_states)
+        # residual connection
+        hidden_states = residual + feed_forward_hidden_states
+
+        if use_cache:
+            outputs = (hidden_states,) + outputs
+        else:
+            outputs = (hidden_states,) + outputs[1:]
+
+        return outputs  # hidden_states, present, (attentions, cross_attentions)
+
+
+class GPT2PreTrainedModel(PreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
+    models.
+    """
+
+    config_class = GPT2Config
+    load_tf_weights = load_tf_weights_in_gpt2
+    base_model_prefix = "transformer"
+    is_parallelizable = True
+    supports_gradient_checkpointing = True
+    _no_split_modules = ["GPT2Block"]
+
+    def __init__(self, *inputs, **kwargs):
+        super().__init__(*inputs, **kwargs)
+
+    def _init_weights(self, module):
+        """Initialize the weights."""
+        if isinstance(module, (nn.Linear, Conv1D)):
+            # 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)
+
+        # Reinitialize selected weights subject to the OpenAI GPT-2 Paper Scheme:
+        #   > A modified initialization which accounts for the accumulation on the residual path with model depth. Scale
+        #   > the weights of residual layers at initialization by a factor of 1/√N where N is the # of residual layers.
+        #   >   -- GPT-2 :: https://openai.com/blog/better-language-models/
+        #
+        # Reference (Megatron-LM): https://github.com/NVIDIA/Megatron-LM/blob/main/megatron/model/gpt_model.py
+        for name, p in module.named_parameters():
+            if name == "c_proj.weight":
+                # Special Scaled Initialization --> There are 2 Layer Norms per Transformer Block
+                p.data.normal_(mean=0.0, std=(self.config.initializer_range / math.sqrt(2 * self.config.n_layer)))
+
+    def _set_gradient_checkpointing(self, module, value=False):
+        if isinstance(module, GPT2Model):
+            module.gradient_checkpointing = value
+
+
+@dataclass
+class GPT2DoubleHeadsModelOutput(ModelOutput):
+    """
+    Base class for outputs of models predicting if two sentences are consecutive or not.
+
+    Args:
+        loss (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided):
+            Language modeling loss.
+        mc_loss (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `mc_labels` is provided):
+            Multiple choice classification loss.
+        logits (`torch.FloatTensor` of shape `(batch_size, num_choices, sequence_length, config.vocab_size)`):
+            Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
+        mc_logits (`torch.FloatTensor` of shape `(batch_size, num_choices)`):
+            Prediction scores of the multiple choice classification head (scores for each choice before SoftMax).
+        past_key_values (`Tuple[Tuple[torch.Tensor]]`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
+            Tuple of length `config.n_layers`, containing tuples of tensors of shape `(batch_size, num_heads,
+            sequence_length, embed_size_per_head)`).
+
+            Contains pre-computed hidden-states (key and values in the attention blocks) that can be used (see
+            `past_key_values` input) to speed up sequential decoding.
+        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)`.
+
+            GPT2Attentions weights after the attention softmax, used to compute the weighted average in the
+            self-attention heads.
+    """
+
+    loss: Optional[torch.FloatTensor] = None
+    mc_loss: Optional[torch.FloatTensor] = None
+    logits: torch.FloatTensor = None
+    mc_logits: torch.FloatTensor = None
+    past_key_values: Optional[Tuple[Tuple[torch.FloatTensor]]] = None
+    hidden_states: Optional[Tuple[torch.FloatTensor]] = None
+    attentions: Optional[Tuple[torch.FloatTensor]] = None
+
+
+GPT2_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 ([`GPT2Config`]): 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.
+"""
+
+GPT2_INPUTS_DOCSTRING = r"""
+    Args:
+        input_ids (`torch.LongTensor` of shape `(batch_size, input_ids_length)`):
+            `input_ids_length` = `sequence_length` if `past_key_values` is `None` else
+            `past_key_values[0][0].shape[-2]` (`sequence_length` of input past key value states). Indices of input
+            sequence tokens in the vocabulary.
+
+            If `past_key_values` is used, only `input_ids` that do not have their past calculated should be passed as
+            `input_ids`.
+
+            Indices can be obtained using [`GPT2Tokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            [What are input IDs?](../glossary#input-ids)
+        past_key_values (`Tuple[Tuple[torch.Tensor]]` of length `config.n_layers`):
+            Contains precomputed hidden-states (key and values in the attention blocks) as computed by the model (see
+            `past_key_values` output below). Can be used to speed up sequential decoding. The `input_ids` which have
+            their past given to this model should not be passed as `input_ids` as they have already been computed.
+        attention_mask (`torch.FloatTensor` of shape `(batch_size, sequence_length)`, *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**.
+
+            If `past_key_values` is used, `attention_mask` needs to contain the masking strategy that was used for
+            `past_key_values`. In other words, the `attention_mask` always has to have the length:
+            `len(past_key_values) + len(input_ids)`
+
+            [What are attention masks?](../glossary#attention-mask)
+        token_type_ids (`torch.LongTensor` of shape `(batch_size, input_ids_length)`, *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 `(batch_size, sequence_length)`, *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 `(batch_size, sequence_length, 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.
+
+            If `past_key_values` is used, optionally only the last `inputs_embeds` have to be input (see
+            `past_key_values`).
+        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 (`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.
+"""
+PARALLELIZE_DOCSTRING = r"""
+    This is an experimental feature and is a subject to change at a moment's notice.
+
+    Uses a device map to distribute attention modules of the model across several devices. If no device map is given,
+    it will evenly distribute blocks across all devices.
+
+    Args:
+        device_map (`Dict[int, list]`, optional, defaults to None):
+            A dictionary that maps attention modules to devices. Note that the embedding module and LMHead are always
+            automatically mapped to the first device (for esoteric reasons). That means that the first device should
+            have fewer attention modules mapped to it than other devices. For reference, the gpt2 models have the
+            following number of attention modules:
+
+                - gpt2: 12
+                - gpt2-medium: 24
+                - gpt2-large: 36
+                - gpt2-xl: 48
+
+    Example:
+
+    ```python
+    # Here is an example of a device map on a machine with 4 GPUs using gpt2-xl, which has a total of 48 attention modules:
+    model = GPT2LMHeadModel.from_pretrained("gpt2-xl")
+    device_map = {
+        0: [0, 1, 2, 3, 4, 5, 6, 7, 8],
+        1: [9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21],
+        2: [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34],
+        3: [35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47],
+    }
+    model.parallelize(device_map)
+    ```
+"""
+DEPARALLELIZE_DOCSTRING = r"""
+    Moves the model to cpu from a model parallel state.
+
+    Example:
+
+    ```python
+    # On a 4 GPU machine with gpt2-large:
+    model = GPT2LMHeadModel.from_pretrained("gpt2-large")
+    device_map = {
+        0: [0, 1, 2, 3, 4, 5, 6, 7],
+        1: [8, 9, 10, 11, 12, 13, 14, 15],
+        2: [16, 17, 18, 19, 20, 21, 22, 23],
+        3: [24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35],
+    }
+    model.parallelize(device_map)  # Splits the model across several devices
+    model.deparallelize()  # Put the model back on cpu and cleans memory by calling torch.cuda.empty_cache()
+    ```
+"""
+
+
+@add_start_docstrings(
+    "The bare GPT2 Model transformer outputting raw hidden-states without any specific head on top.",
+    GPT2_START_DOCSTRING,
+)
+class GPT2Model(GPT2PreTrainedModel):
+    _keys_to_ignore_on_load_missing = ["attn.masked_bias"]
+
+    def __init__(self, config):
+        super().__init__(config)
+
+        self.embed_dim = config.hidden_size
+
+        self.wte = nn.Embedding(config.vocab_size, self.embed_dim)
+        self.wpe = nn.Embedding(config.max_position_embeddings, self.embed_dim)
+
+        self.drop = nn.Dropout(config.embd_pdrop)
+        self.h = nn.ModuleList([GPT2Block(config, layer_idx=i) for i in range(config.num_hidden_layers)])
+        self.ln_f = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_epsilon)
+
+        # Model parallel
+        self.model_parallel = False
+        self.device_map = None
+        self.gradient_checkpointing = False
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings(PARALLELIZE_DOCSTRING)
+    def parallelize(self, device_map=None):
+        # Check validity of device_map
+        self.device_map = (
+            get_device_map(len(self.h), range(torch.cuda.device_count())) if device_map is None else device_map
+        )
+        assert_device_map(self.device_map, len(self.h))
+        self.model_parallel = True
+        self.first_device = "cpu" if "cpu" in self.device_map.keys() else "cuda:" + str(min(self.device_map.keys()))
+        self.last_device = "cuda:" + str(max(self.device_map.keys()))
+        self.wte = self.wte.to(self.first_device)
+        self.wpe = self.wpe.to(self.first_device)
+        # Load onto devices
+        for k, v in self.device_map.items():
+            for block in v:
+                cuda_device = "cuda:" + str(k)
+                self.h[block] = self.h[block].to(cuda_device)
+        # ln_f to last
+        self.ln_f = self.ln_f.to(self.last_device)
+
+    @add_start_docstrings(DEPARALLELIZE_DOCSTRING)
+    def deparallelize(self):
+        self.model_parallel = False
+        self.device_map = None
+        self.first_device = "cpu"
+        self.last_device = "cpu"
+        self.wte = self.wte.to("cpu")
+        self.wpe = self.wpe.to("cpu")
+        for index in range(len(self.h)):
+            self.h[index] = self.h[index].to("cpu")
+        self.ln_f = self.ln_f.to("cpu")
+        torch.cuda.empty_cache()
+
+    def get_input_embeddings(self):
+        return self.wte
+
+    def set_input_embeddings(self, new_embeddings):
+        self.wte = new_embeddings
+
+    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}
+        """
+        for layer, heads in heads_to_prune.items():
+            self.h[layer].attn.prune_heads(heads)
+
+    @add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING)
+    @add_code_sample_docstrings(
+        processor_class=_TOKENIZER_FOR_DOC,
+        checkpoint=_CHECKPOINT_FOR_DOC,
+        output_type=BaseModelOutputWithPastAndCrossAttentions,
+        config_class=_CONFIG_FOR_DOC,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        token_type_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[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, BaseModelOutputWithPastAndCrossAttentions]:
+        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
+        )
+        use_cache = use_cache if use_cache is not None else self.config.use_cache
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        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()
+            input_ids = input_ids.view(-1, input_shape[-1])
+            batch_size = input_ids.shape[0]
+        elif inputs_embeds is not None:
+            input_shape = inputs_embeds.size()[:-1]
+            batch_size = inputs_embeds.shape[0]
+        else:
+            raise ValueError("You have to specify either input_ids or inputs_embeds")
+
+        device = input_ids.device if input_ids is not None else inputs_embeds.device
+
+        if token_type_ids is not None:
+            token_type_ids = token_type_ids.view(-1, input_shape[-1])
+        if position_ids is not None:
+            position_ids = position_ids.view(-1, input_shape[-1])
+
+        if past_key_values is None:
+            past_length = 0
+            past_key_values = tuple([None] * len(self.h))
+        else:
+            past_length = past_key_values[0][0].size(-2)
+        if position_ids is None:
+            position_ids = torch.arange(past_length, input_shape[-1] + past_length, dtype=torch.long, device=device)
+            position_ids = position_ids.unsqueeze(0).view(-1, input_shape[-1])
+
+        # GPT2Attention mask.
+        if attention_mask is not None:
+            if batch_size <= 0:
+                raise ValueError("batch_size has to be defined and > 0")
+            attention_mask = attention_mask.view(batch_size, -1)
+            # We create a 3D attention mask from a 2D tensor mask.
+            # Sizes are [batch_size, 1, 1, to_seq_length]
+            # So we can broadcast to [batch_size, num_heads, from_seq_length, to_seq_length]
+            # this attention mask is more simple than the triangular masking of causal attention
+            # used in OpenAI GPT, we just need to prepare the broadcast dimension here.
+            attention_mask = attention_mask[:, None, None, :]
+
+            # Since attention_mask is 1.0 for positions we want to attend and 0.0 for
+            # masked positions, this operation will create a tensor which is 0.0 for
+            # positions we want to attend and the dtype's smallest value for masked positions.
+            # Since we are adding it to the raw scores before the softmax, this is
+            # effectively the same as removing these entirely.
+            attention_mask = attention_mask.to(dtype=self.dtype)  # fp16 compatibility
+            attention_mask = (1.0 - attention_mask) * torch.finfo(self.dtype).min
+
+        # 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.add_cross_attention 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_attention_mask = self.invert_attention_mask(encoder_attention_mask)
+        else:
+            encoder_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
+        # head_mask has shape n_layer x batch x n_heads x N x N
+        head_mask = self.get_head_mask(head_mask, self.config.n_layer)
+
+        if inputs_embeds is None:
+            inputs_embeds = self.wte(input_ids)
+        position_embeds = self.wpe(position_ids)
+        hidden_states = inputs_embeds + position_embeds
+
+        if token_type_ids is not None:
+            token_type_embeds = self.wte(token_type_ids)
+            hidden_states = hidden_states + token_type_embeds
+
+        hidden_states = self.drop(hidden_states)
+
+        output_shape = input_shape + (hidden_states.size(-1),)
+
+        presents = () if use_cache else None
+        all_self_attentions = () if output_attentions else None
+        all_cross_attentions = () if output_attentions and self.config.add_cross_attention else None
+        all_hidden_states = () if output_hidden_states else None
+        for i, (block, layer_past) in enumerate(zip(self.h, past_key_values)):
+
+            # Model parallel
+            if self.model_parallel:
+                torch.cuda.set_device(hidden_states.device)
+                # Ensure layer_past is on same device as hidden_states (might not be correct)
+                if layer_past is not None:
+                    layer_past = tuple(past_state.to(hidden_states.device) for past_state in layer_past)
+                # Ensure that attention_mask is always on the same device as hidden_states
+                if attention_mask is not None:
+                    attention_mask = attention_mask.to(hidden_states.device)
+                if isinstance(head_mask, torch.Tensor):
+                    head_mask = head_mask.to(hidden_states.device)
+            if output_hidden_states:
+                all_hidden_states = all_hidden_states + (hidden_states,)
+
+            if self.gradient_checkpointing and self.training:
+
+                if use_cache:
+                    logger.warning(
+                        "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
+                    )
+                    use_cache = False
+
+                def create_custom_forward(module):
+                    def custom_forward(*inputs):
+                        # None for past_key_value
+                        return module(*inputs, use_cache, output_attentions)
+
+                    return custom_forward
+
+                outputs = torch.utils.checkpoint.checkpoint(
+                    create_custom_forward(block),
+                    hidden_states,
+                    None,
+                    attention_mask,
+                    head_mask[i],
+                    encoder_hidden_states,
+                    encoder_attention_mask,
+                )
+            else:
+                outputs = block(
+                    hidden_states,
+                    layer_past=layer_past,
+                    attention_mask=attention_mask,
+                    head_mask=head_mask[i],
+                    encoder_hidden_states=encoder_hidden_states,
+                    encoder_attention_mask=encoder_attention_mask,
+                    use_cache=use_cache,
+                    output_attentions=output_attentions,
+                )
+
+            hidden_states = outputs[0]
+            if use_cache is True:
+                presents = presents + (outputs[1],)
+
+            if output_attentions:
+                all_self_attentions = all_self_attentions + (outputs[2 if use_cache else 1],)
+                if self.config.add_cross_attention:
+                    all_cross_attentions = all_cross_attentions + (outputs[3 if use_cache else 2],)
+
+            # Model Parallel: If it's the last layer for that device, put things on the next device
+            if self.model_parallel:
+                for k, v in self.device_map.items():
+                    if i == v[-1] and "cuda:" + str(k) != self.last_device:
+                        hidden_states = hidden_states.to("cuda:" + str(k + 1))
+
+        hidden_states = self.ln_f(hidden_states)
+
+        hidden_states = hidden_states.view(output_shape)
+        # Add last hidden state
+        if output_hidden_states:
+            all_hidden_states = all_hidden_states + (hidden_states,)
+
+        if not return_dict:
+            return tuple(
+                v
+                for v in [hidden_states, presents, all_hidden_states, all_self_attentions, all_cross_attentions]
+                if v is not None
+            )
+
+        return BaseModelOutputWithPastAndCrossAttentions(
+            last_hidden_state=hidden_states,
+            past_key_values=presents,
+            hidden_states=all_hidden_states,
+            attentions=all_self_attentions,
+            cross_attentions=all_cross_attentions,
+        )
+
+
+@add_start_docstrings(
+    """
+    The GPT2 Model transformer with a language modeling head on top (linear layer with weights tied to the input
+    embeddings).
+    """,
+    GPT2_START_DOCSTRING,
+)
+class GPT2CustomLMHeadModel(GPT2PreTrainedModel, GenerationMixin):
+    _keys_to_ignore_on_load_missing = [r"attn.masked_bias", r"attn.bias", r"lm_head.weight"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.transformer = GPT2Model(config)
+        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
+
+        # Model parallel
+        self.model_parallel = False
+        self.device_map = None
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings(PARALLELIZE_DOCSTRING)
+    def parallelize(self, device_map=None):
+        self.device_map = (
+            get_device_map(len(self.transformer.h), range(torch.cuda.device_count()))
+            if device_map is None
+            else device_map
+        )
+        assert_device_map(self.device_map, len(self.transformer.h))
+        self.transformer.parallelize(self.device_map)
+        self.lm_head = self.lm_head.to(self.transformer.first_device)
+        self.model_parallel = True
+
+    @add_start_docstrings(DEPARALLELIZE_DOCSTRING)
+    def deparallelize(self):
+        self.transformer.deparallelize()
+        self.transformer = self.transformer.to("cpu")
+        self.lm_head = self.lm_head.to("cpu")
+        self.model_parallel = False
+        torch.cuda.empty_cache()
+
+    def get_output_embeddings(self):
+        return self.lm_head
+
+    def set_output_embeddings(self, new_embeddings):
+        self.lm_head = new_embeddings
+
+    def prepare_inputs_for_generation(self, input_ids, past=None, **kwargs):
+        token_type_ids = kwargs.get("token_type_ids", None)
+        # only last token for inputs_ids if past is defined in kwargs
+        if past:
+            input_ids = input_ids[:, -1].unsqueeze(-1)
+            if token_type_ids is not None:
+                token_type_ids = token_type_ids[:, -1].unsqueeze(-1)
+
+        attention_mask = kwargs.get("attention_mask", None)
+        position_ids = kwargs.get("position_ids", None)
+
+        if attention_mask is not None and position_ids is None:
+            # create position_ids on the fly for batch generation
+            position_ids = attention_mask.long().cumsum(-1) - 1
+            position_ids.masked_fill_(attention_mask == 0, 1)
+            if past:
+                position_ids = position_ids[:, -1].unsqueeze(-1)
+        else:
+            position_ids = None
+        return {
+            "input_ids": input_ids,
+            "past_key_values": past,
+            "use_cache": kwargs.get("use_cache"),
+            "position_ids": position_ids,
+            "attention_mask": attention_mask,
+            "token_type_ids": token_type_ids,
+        }
+
+    @add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING)
+    @add_code_sample_docstrings(
+        processor_class=_TOKENIZER_FOR_DOC,
+        checkpoint=_CHECKPOINT_FOR_DOC,
+        output_type=CausalLMOutputWithCrossAttentions,
+        config_class=_CONFIG_FOR_DOC,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        token_type_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, CausalLMOutputWithCrossAttentions]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Labels for language modeling. Note that the labels **are shifted** inside the model, i.e. you can set
+            `labels = input_ids` Indices are selected in `[-100, 0, ..., config.vocab_size]` All labels set to `-100`
+            are ignored (masked), the loss is only computed for labels in `[0, ..., config.vocab_size]`
+        """
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        transformer_outputs = self.transformer(
+            input_ids,
+            past_key_values=past_key_values,
+            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,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        hidden_states = transformer_outputs[0]
+
+        # Set device for model parallelism
+        if self.model_parallel:
+            torch.cuda.set_device(self.transformer.first_device)
+            hidden_states = hidden_states.to(self.lm_head.weight.device)
+
+        lm_logits = self.lm_head(hidden_states)
+
+        loss = None
+        if labels is not None:
+            # Shift so that tokens < n predict n
+            shift_logits = lm_logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            # Flatten the tokens
+            loss_fct = CrossEntropyLoss()
+            loss = loss_fct(shift_logits.view(-1, shift_logits.size(-1)), shift_labels.view(-1))
+
+        if not return_dict:
+            output = (lm_logits,) + transformer_outputs[1:]
+            return ((loss,) + output) if loss is not None else output
+
+        return CausalLMOutputWithCrossAttentions(
+            loss=loss,
+            logits=lm_logits,
+            past_key_values=transformer_outputs.past_key_values,
+            hidden_states=transformer_outputs.hidden_states,
+            attentions=transformer_outputs.attentions,
+            cross_attentions=transformer_outputs.cross_attentions,
+        )
+
+    @staticmethod
+    def _reorder_cache(past: Tuple[Tuple[torch.Tensor]], beam_idx: torch.Tensor) -> Tuple[Tuple[torch.Tensor]]:
+        """
+        This function is used to re-order the `past_key_values` cache if [`~PreTrainedModel.beam_search`] or
+        [`~PreTrainedModel.beam_sample`] is called. This is required to match `past_key_values` with the correct
+        beam_idx at every generation step.
+        """
+        return tuple(
+            tuple(past_state.index_select(0, beam_idx.to(past_state.device)) for past_state in layer_past)
+            for layer_past in past
+        )
+
+
+@add_start_docstrings(
+    """
+The GPT2 Model transformer with a language modeling and a multiple-choice classification head on top e.g. for
+RocStories/SWAG tasks. The two heads are two linear layers. The language modeling head has its weights tied to the
+input embeddings, the classification head takes as input the input of a specified classification token index in the
+input sequence).
+""",
+    GPT2_START_DOCSTRING,
+)
+class GPT2DoubleHeadsModel(GPT2PreTrainedModel):
+    _keys_to_ignore_on_load_missing = [r"attn.masked_bias", r"attn.bias", r"lm_head.weight"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        config.num_labels = 1
+        self.transformer = GPT2Model(config)
+        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
+        self.multiple_choice_head = SequenceSummary(config)
+
+        # Model parallel
+        self.model_parallel = False
+        self.device_map = None
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings(PARALLELIZE_DOCSTRING)
+    def parallelize(self, device_map=None):
+        self.device_map = (
+            get_device_map(len(self.transformer.h), range(torch.cuda.device_count()))
+            if device_map is None
+            else device_map
+        )
+        assert_device_map(self.device_map, len(self.transformer.h))
+        self.transformer.parallelize(self.device_map)
+        self.lm_head = self.lm_head.to(self.transformer.first_device)
+        self.multiple_choice_head = self.multiple_choice_head.to(self.transformer.first_device)
+        self.model_parallel = True
+
+    @add_start_docstrings(DEPARALLELIZE_DOCSTRING)
+    def deparallelize(self):
+        self.transformer.deparallelize()
+        self.transformer = self.transformer.to("cpu")
+        self.lm_head = self.lm_head.to("cpu")
+        self.multiple_choice_head = self.multiple_choice_head.to("cpu")
+        self.model_parallel = False
+        torch.cuda.empty_cache()
+
+    def get_output_embeddings(self):
+        return self.lm_head
+
+    def set_output_embeddings(self, new_embeddings):
+        self.lm_head = new_embeddings
+
+    def prepare_inputs_for_generation(self, input_ids, past=None, **kwargs):
+        token_type_ids = kwargs.get("token_type_ids", None)
+        # only last token for inputs_ids if past is defined in kwargs
+        if past:
+            input_ids = input_ids[:, -1].unsqueeze(-1)
+            if token_type_ids is not None:
+                token_type_ids = token_type_ids[:, -1].unsqueeze(-1)
+
+        attention_mask = kwargs.get("attention_mask", None)
+        position_ids = kwargs.get("position_ids", None)
+
+        if attention_mask is not None and position_ids is None:
+            # create position_ids on the fly for batch generation
+            position_ids = attention_mask.long().cumsum(-1) - 1
+            position_ids.masked_fill_(attention_mask == 0, 1)
+            if past:
+                position_ids = position_ids[:, -1].unsqueeze(-1)
+        else:
+            position_ids = None
+
+        return {
+            "input_ids": input_ids,
+            "past_key_values": past,
+            "use_cache": kwargs.get("use_cache"),
+            "position_ids": position_ids,
+            "attention_mask": attention_mask,
+            "token_type_ids": token_type_ids,
+        }
+
+    @add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING)
+    @replace_return_docstrings(output_type=GPT2DoubleHeadsModelOutput, config_class=_CONFIG_FOR_DOC)
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        token_type_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        mc_token_ids: Optional[torch.LongTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        mc_labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        **kwargs,
+    ) -> Union[Tuple, GPT2DoubleHeadsModelOutput]:
+        r"""
+        mc_token_ids (`torch.LongTensor` of shape `(batch_size, num_choices)`, *optional*, default to index of the last token of the input):
+            Index of the classification token in each input sequence. Selected in the range `[0, input_ids.size(-1) -
+            1]`.
+        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Labels for language modeling. Note that the labels **are shifted** inside the model, i.e. you can set
+            `labels = input_ids`. Indices are selected in `[-100, 0, ..., config.vocab_size - 1]`. All labels set to
+            `-100` are ignored (masked), the loss is only computed for labels in `[0, ..., config.vocab_size - 1]`
+        mc_labels (`torch.LongTensor` of shape `(batch_size)`, *optional*):
+            Labels for computing the multiple choice classification loss. Indices should be in `[0, ..., num_choices]`
+            where *num_choices* is the size of the second dimension of the input tensors. (see *input_ids* above)
+
+        Return:
+
+        Example:
+
+        ```python
+        >>> import torch
+        >>> from transformers import GPT2Tokenizer, GPT2DoubleHeadsModel
+
+        >>> tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
+        >>> model = GPT2DoubleHeadsModel.from_pretrained("gpt2")
+
+        >>> # Add a [CLS] to the vocabulary (we should train it also!)
+        >>> num_added_tokens = tokenizer.add_special_tokens({"cls_token": "[CLS]"})
+        >>> # Update the model embeddings with the new vocabulary size
+        >>> embedding_layer = model.resize_token_embeddings(len(tokenizer))
+
+        >>> choices = ["Hello, my dog is cute [CLS]", "Hello, my cat is cute [CLS]"]
+        >>> encoded_choices = [tokenizer.encode(s) for s in choices]
+        >>> cls_token_location = [tokens.index(tokenizer.cls_token_id) for tokens in encoded_choices]
+
+        >>> input_ids = torch.tensor(encoded_choices).unsqueeze(0)  # Batch size: 1, number of choices: 2
+        >>> mc_token_ids = torch.tensor([cls_token_location])  # Batch size: 1
+
+        >>> outputs = model(input_ids, mc_token_ids=mc_token_ids)
+        >>> lm_logits = outputs.logits
+        >>> mc_logits = outputs.mc_logits
+        ```"""
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        transformer_outputs = self.transformer(
+            input_ids,
+            past_key_values=past_key_values,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        hidden_states = transformer_outputs[0]
+
+        # Set device for model parallelism
+        if self.model_parallel:
+            torch.cuda.set_device(self.transformer.first_device)
+            hidden_states = hidden_states.to(self.lm_head.weight.device)
+
+        lm_logits = self.lm_head(hidden_states)
+        mc_logits = self.multiple_choice_head(hidden_states, mc_token_ids).squeeze(-1)
+
+        mc_loss = None
+        if mc_labels is not None:
+            loss_fct = CrossEntropyLoss()
+            mc_loss = loss_fct(mc_logits.view(-1, mc_logits.size(-1)), mc_labels.view(-1))
+        lm_loss = None
+        if labels is not None:
+            shift_logits = lm_logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            loss_fct = CrossEntropyLoss()
+            lm_loss = loss_fct(shift_logits.view(-1, shift_logits.size(-1)), shift_labels.view(-1))
+
+        if not return_dict:
+            output = (lm_logits, mc_logits) + transformer_outputs[1:]
+            if mc_loss is not None:
+                output = (mc_loss,) + output
+            return ((lm_loss,) + output) if lm_loss is not None else output
+
+        return GPT2DoubleHeadsModelOutput(
+            loss=lm_loss,
+            mc_loss=mc_loss,
+            logits=lm_logits,
+            mc_logits=mc_logits,
+            past_key_values=transformer_outputs.past_key_values,
+            hidden_states=transformer_outputs.hidden_states,
+            attentions=transformer_outputs.attentions,
+        )
+
+    @staticmethod
+    def _reorder_cache(past: Tuple[Tuple[torch.Tensor]], beam_idx: torch.Tensor) -> Tuple[Tuple[torch.Tensor]]:
+        """
+        This function is used to re-order the `past_key_values` cache if [`~PreTrainedModel.beam_search`] or
+        [`~PreTrainedModel.beam_sample`] is called. This is required to match `past_key_values` with the correct
+        beam_idx at every generation step.
+        """
+        return tuple(
+            tuple(past_state.index_select(0, beam_idx.to(past_state.device)) for past_state in layer_past)
+            for layer_past in past
+        )
+
+
+@add_start_docstrings(
+    """
+    The GPT2 Model transformer with a sequence classification head on top (linear layer).
+
+    [`GPT2ForSequenceClassification`] uses the last token in order to do the classification, as other causal models
+    (e.g. GPT-1) do.
+
+    Since it does classification on the last token, it requires to know the position of the last token. If a
+    `pad_token_id` is defined in the configuration, it finds the last token that is not a padding token in each row. If
+    no `pad_token_id` is defined, it simply takes the last value in each row of the batch. Since it cannot guess the
+    padding tokens when `inputs_embeds` are passed instead of `input_ids`, it does the same (take the last value in
+    each row of the batch).
+    """,
+    GPT2_START_DOCSTRING,
+)
+class GPT2ForSequenceClassification(GPT2PreTrainedModel):
+    _keys_to_ignore_on_load_missing = [r"h\.\d+\.attn\.masked_bias", r"lm_head.weight"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+        self.transformer = GPT2Model(config)
+        self.score = nn.Linear(config.n_embd, self.num_labels, bias=False)
+
+        # Model parallel
+        self.model_parallel = False
+        self.device_map = None
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING)
+    @add_code_sample_docstrings(
+        processor_class=_TOKENIZER_FOR_DOC,
+        checkpoint="microsoft/DialogRPT-updown",
+        output_type=SequenceClassifierOutputWithPast,
+        config_class=_CONFIG_FOR_DOC,
+        expected_output="'LABEL_0'",
+        expected_loss=5.28,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        token_type_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, SequenceClassifierOutputWithPast]:
+        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
+
+        transformer_outputs = self.transformer(
+            input_ids,
+            past_key_values=past_key_values,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        hidden_states = transformer_outputs[0]
+        logits = self.score(hidden_states)
+
+        if input_ids is not None:
+            batch_size, sequence_length = input_ids.shape[:2]
+        else:
+            batch_size, sequence_length = inputs_embeds.shape[:2]
+
+        assert (
+            self.config.pad_token_id is not None or batch_size == 1
+        ), "Cannot handle batch sizes > 1 if no padding token is defined."
+        if self.config.pad_token_id is None:
+            sequence_lengths = -1
+        else:
+            if input_ids is not None:
+                sequence_lengths = torch.ne(input_ids, self.config.pad_token_id).sum(-1) - 1
+            else:
+                sequence_lengths = -1
+                logger.warning(
+                    f"{self.__class__.__name__} will not detect padding tokens in `inputs_embeds`. Results may be "
+                    "unexpected if using padding tokens in conjunction with `inputs_embeds.`"
+                )
+
+        pooled_logits = logits[torch.arange(batch_size, device=logits.device), sequence_lengths]
+
+        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(pooled_logits.squeeze(), labels.squeeze())
+                else:
+                    loss = loss_fct(pooled_logits, labels)
+            elif self.config.problem_type == "single_label_classification":
+                loss_fct = CrossEntropyLoss()
+                loss = loss_fct(pooled_logits.view(-1, self.num_labels), labels.view(-1))
+            elif self.config.problem_type == "multi_label_classification":
+                loss_fct = BCEWithLogitsLoss()
+                loss = loss_fct(pooled_logits, labels)
+        if not return_dict:
+            output = (pooled_logits,) + transformer_outputs[1:]
+            return ((loss,) + output) if loss is not None else output
+
+        return SequenceClassifierOutputWithPast(
+            loss=loss,
+            logits=pooled_logits,
+            past_key_values=transformer_outputs.past_key_values,
+            hidden_states=transformer_outputs.hidden_states,
+            attentions=transformer_outputs.attentions,
+        )
+
+
+@add_start_docstrings(
+    """
+    GPT2 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.
+    """,
+    GPT2_START_DOCSTRING,
+)
+class GPT2ForTokenClassification(GPT2PreTrainedModel):
+    def __init__(self, config):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+
+        self.transformer = GPT2Model(config)
+        if hasattr(config, "classifier_dropout") and config.classifier_dropout is not None:
+            classifier_dropout = config.classifier_dropout
+        elif hasattr(config, "hidden_dropout") and config.hidden_dropout is not None:
+            classifier_dropout = config.hidden_dropout
+        else:
+            classifier_dropout = 0.1
+        self.dropout = nn.Dropout(classifier_dropout)
+        self.classifier = nn.Linear(config.hidden_size, config.num_labels)
+
+        # Model parallel
+        self.model_parallel = False
+        self.device_map = None
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(GPT2_INPUTS_DOCSTRING)
+    # fmt: off
+    @add_code_sample_docstrings(
+        processor_class=_TOKENIZER_FOR_DOC,
+        checkpoint="brad1141/gpt2-finetuned-comp2",
+        output_type=TokenClassifierOutput,
+        config_class=_CONFIG_FOR_DOC,
+        expected_loss=0.25,
+        expected_output=["Lead", "Lead", "Lead", "Position", "Lead", "Lead", "Lead", "Lead", "Lead", "Lead", "Lead", "Lead"],
+    )
+    # fmt: on
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        token_type_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, TokenClassifierOutput]:
+        r"""
+        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *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
+
+        transformer_outputs = self.transformer(
+            input_ids,
+            past_key_values=past_key_values,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        hidden_states = transformer_outputs[0]
+        hidden_states = self.dropout(hidden_states)
+        logits = self.classifier(hidden_states)
+
+        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,) + transformer_outputs[2:]
+            return ((loss,) + output) if loss is not None else output
+
+        return TokenClassifierOutput(
+            loss=loss,
+            logits=logits,
+            hidden_states=transformer_outputs.hidden_states,
+            attentions=transformer_outputs.attentions,
+        )

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:677f711dfade47fe3bb712a5af6dc3b8bfeb1e22967a4f78777df319d2e15909
+size 1781199

special_tokens_map.json

@@ -0,0 +1,5 @@
+{
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "unk_token": "<|endoftext|>"
+}

tokenizer.json

@@ -0,0 +1,1853 @@
+{
+  "version": "1.0",
+  "truncation": null,
+  "padding": null,
+  "added_tokens": [
+    {
+      "id": 0,
+      "content": "<|endoftext|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    }
+  ],
+  "normalizer": null,
+  "pre_tokenizer": {
+    "type": "ByteLevel",
+    "add_prefix_space": false,
+    "trim_offsets": true,
+    "use_regex": true
+  },
+  "post_processor": {
+    "type": "ByteLevel",
+    "add_prefix_space": true,
+    "trim_offsets": false,
+    "use_regex": true
+  },
+  "decoder": {
+    "type": "ByteLevel",
+    "add_prefix_space": true,
+    "trim_offsets": true,
+    "use_regex": true
+  },
+  "model": {
+    "type": "BPE",
+    "dropout": null,
+    "unk_token": null,
+    "continuing_subword_prefix": "",
+    "end_of_word_suffix": "",
+    "fuse_unk": false,
+    "vocab": {
+      "<|endoftext|>": 0,
+      "!": 1,
+      "\"": 2,
+      "#": 3,
+      "$": 4,
+      "%": 5,
+      "&": 6,
+      "'": 7,
+      "(": 8,
+      ")": 9,
+      "*": 10,
+      "+": 11,
+      ",": 12,
+      "-": 13,
+      ".": 14,
+      "/": 15,
+      "0": 16,
+      "1": 17,
+      "2": 18,
+      "3": 19,
+      "4": 20,
+      "5": 21,
+      "6": 22,
+      "7": 23,
+      "8": 24,
+      "9": 25,
+      ":": 26,
+      ";": 27,
+      "<": 28,
+      "=": 29,
+      ">": 30,
+      "?": 31,
+      "@": 32,
+      "A": 33,
+      "B": 34,
+      "C": 35,
+      "D": 36,
+      "E": 37,
+      "F": 38,
+      "G": 39,
+      "H": 40,
+      "I": 41,
+      "J": 42,
+      "K": 43,
+      "L": 44,
+      "M": 45,
+      "N": 46,
+      "O": 47,
+      "P": 48,
+      "Q": 49,
+      "R": 50,
+      "S": 51,
+      "T": 52,
+      "U": 53,
+      "V": 54,
+      "W": 55,
+      "X": 56,
+      "Y": 57,
+      "Z": 58,
+      "[": 59,
+      "\\": 60,
+      "]": 61,
+      "^": 62,
+      "_": 63,
+      "`": 64,
+      "a": 65,
+      "b": 66,
+      "c": 67,
+      "d": 68,
+      "e": 69,
+      "f": 70,
+      "g": 71,
+      "h": 72,
+      "i": 73,
+      "j": 74,
+      "k": 75,
+      "l": 76,
+      "m": 77,
+      "n": 78,
+      "o": 79,
+      "p": 80,
+      "q": 81,
+      "r": 82,
+      "s": 83,
+      "t": 84,
+      "u": 85,
+      "v": 86,
+      "w": 87,
+      "x": 88,
+      "y": 89,
+      "z": 90,
+      "|": 91,
+      "}": 92,
+      "~": 93,
+      "¡": 94,
+      "¢": 95,
+      "£": 96,
+      "¤": 97,
+      "¥": 98,
+      "¦": 99,
+      "§": 100,
+      "¨": 101,
+      "©": 102,
+      "ª": 103,
+      "«": 104,
+      "¬": 105,
+      "®": 106,
+      "¯": 107,
+      "°": 108,
+      "±": 109,
+      "²": 110,
+      "³": 111,
+      "´": 112,
+      "µ": 113,
+      "¶": 114,
+      "·": 115,
+      "¸": 116,
+      "¹": 117,
+      "º": 118,
+      "»": 119,
+      "¼": 120,
+      "½": 121,
+      "¾": 122,
+      "¿": 123,
+      "Â": 124,
+      "Ã": 125,
+      "Ä": 126,
+      "Å": 127,
+      "Æ": 128,
+      "Ç": 129,
+      "È": 130,
+      "É": 131,
+      "Ê": 132,
+      "Ë": 133,
+      "Ì": 134,
+      "Í": 135,
+      "Î": 136,
+      "Ï": 137,
+      "Ð": 138,
+      "Ñ": 139,
+      "Ö": 140,
+      "×": 141,
+      "Ø": 142,
+      "Ù": 143,
+      "Ü": 144,
+      "à": 145,
+      "á": 146,
+      "â": 147,
+      "ã": 148,
+      "ä": 149,
+      "å": 150,
+      "æ": 151,
+      "ç": 152,
+      "è": 153,
+      "é": 154,
+      "ë": 155,
+      "ì": 156,
+      "ï": 157,
+      "Ċ": 158,
+      "Ġ": 159,
+      "Ģ": 160,
+      "ģ": 161,
+      "Ĥ": 162,
+      "ĥ": 163,
+      "Ħ": 164,
+      "ħ": 165,
+      "Ĩ": 166,
+      "ĩ": 167,
+      "Ī": 168,
+      "ī": 169,
+      "Ĭ": 170,
+      "ĭ": 171,
+      "Į": 172,
+      "į": 173,
+      "İ": 174,
+      "ı": 175,
+      "IJ": 176,
+      "ij": 177,
+      "Ĵ": 178,
+      "ĵ": 179,
+      "Ķ": 180,
+      "ķ": 181,
+      "ĸ": 182,
+      "Ĺ": 183,
+      "ĺ": 184,
+      "Ļ": 185,
+      "ļ": 186,
+      "Ľ": 187,
+      "ľ": 188,
+      "Ŀ": 189,
+      "ŀ": 190,
+      "Ł": 191,
+      "ł": 192,
+      "Ń": 193,
+      "Ġt": 194,
+      "he": 195,
+      "Ġa": 196,
+      "in": 197,
+      "Ġthe": 198,
+      "er": 199,
+      "on": 200,
+      "Ġ,": 201,
+      "re": 202,
+      "Ġs": 203,
+      "ed": 204,
+      "Ġo": 205,
+      "Ġw": 206,
+      "nd": 207,
+      "at": 208,
+      "Ġ.": 209,
+      "or": 210,
+      "it": 211,
+      "Ġc": 212,
+      "en": 213,
+      "Ġf": 214,
+      "is": 215,
+      "es": 216,
+      "ar": 217,
+      "Ġof": 218,
+      "Ġb": 219,
+      "an": 220,
+      "Ġin": 221,
+      "al": 222,
+      "ing": 223,
+      "Ġp": 224,
+      "Ġand": 225,
+      "as": 226,
+      "Ġto": 227,
+      "ro": 228,
+      "ic": 229,
+      "Ġm": 230,
+      "Ġd": 231,
+      "Ġh": 232,
+      "ion": 233,
+      "le": 234,
+      "ou": 235,
+      "ĠT": 236,
+      "Ġre": 237,
+      "Ġ=": 238,
+      "Ġ\"": 239,
+      "ĠA": 240,
+      "ĠS": 241,
+      "ent": 242,
+      "il": 243,
+      "Ġth": 244,
+      "Ġ1": 245,
+      "st": 246,
+      "ĠC": 247,
+      "el": 248,
+      "om": 249,
+      "Ġl": 250,
+      "am": 251,
+      "ĠĊ": 252,
+      "Ġe": 253,
+      "Ġn": 254,
+      "Ġ@": 255,
+      "ad": 256,
+      "ac": 257,
+      "Ġwas": 258,
+      "ĠM": 259,
+      "ur": 260,
+      "ĠThe": 261,
+      "ec": 262,
+      "Ġon": 263,
+      "ly": 264,
+      "ĠB": 265,
+      "ĠI": 266,
+      "Ġg": 267,
+      "Ġ'": 268,
+      "et": 269,
+      "ol": 270,
+      "id": 271,
+      "iv": 272,
+      "im": 273,
+      "Ġfor": 274,
+      "ir": 275,
+      "-@": 276,
+      "Ġ@-@": 277,
+      "ig": 278,
+      "ot": 279,
+      "ter": 280,
+      "Ġas": 281,
+      "ĠH": 282,
+      "us": 283,
+      "ow": 284,
+      "Ġst": 285,
+      "ut": 286,
+      "ith": 287,
+      "ay": 288,
+      "Ġ2": 289,
+      "ĠP": 290,
+      "ation": 291,
+      "ver": 292,
+      "Ġbe": 293,
+      "her": 294,
+      "Ġthat": 295,
+      "Ġwith": 296,
+      "ĠR": 297,
+      "ce": 298,
+      "th": 299,
+      "ĠD": 300,
+      "Ġis": 301,
+      "un": 302,
+      "em": 303,
+      "ĠF": 304,
+      "Ġwh": 305,
+      "ul": 306,
+      "Ġby": 307,
+      "Ġal": 308,
+      "ch": 309,
+      "Ġ)": 310,
+      "Ġ(": 311,
+      "ĠW": 312,
+      "Ġcon": 313,
+      "ra": 314,
+      "ĠG": 315,
+      "os": 316,
+      "ĠL": 317,
+      "ĠN": 318,
+      "Ġat": 319,
+      "ers": 320,
+      "ct": 321,
+      "Ġit": 322,
+      "Ġ19": 323,
+      "rom": 324,
+      "and": 325,
+      "Ġan": 326,
+      "um": 327,
+      "est": 328,
+      "ĠJ": 329,
+      "ag": 330,
+      "Ġhe": 331,
+      "00": 332,
+      "ist": 333,
+      "ain": 334,
+      "od": 335,
+      "av": 336,
+      "ri": 337,
+      "ĠE": 338,
+      "ĠO": 339,
+      "Ġfrom": 340,
+      "Ġcom": 341,
+      "Ġhis": 342,
+      "op": 343,
+      "Ġpro": 344,
+      "res": 345,
+      "ies": 346,
+      "if": 347,
+      "Ġv": 348,
+      "ort": 349,
+      "ere": 350,
+      "ill": 351,
+      "ld": 352,
+      "Ġde": 353,
+      "pp": 354,
+      "Ġsu": 355,
+      "ore": 356,
+      "ĠIn": 357,
+      "Ġr": 358,
+      "Ġse": 359,
+      "Ġwere": 360,
+      "ew": 361,
+      "ong": 362,
+      "igh": 363,
+      "ard": 364,
+      "ate": 365,
+      "all": 366,
+      "art": 367,
+      "ak": 368,
+      "ich": 369,
+      "Ġch": 370,
+      "Ġor": 371,
+      "ab": 372,
+      "ant": 373,
+      "ud": 374,
+      "oc": 375,
+      "ber": 376,
+      "Ġex": 377,
+      "gh": 378,
+      "ity": 379,
+      "ated": 380,
+      "pt": 381,
+      "ess": 382,
+      "ear": 383,
+      "ĠK": 384,
+      "Ġpl": 385,
+      "ame": 386,
+      "qu": 387,
+      "ive": 388,
+      "rou": 389,
+      "Ġare": 390,
+      "Ġâ": 391,
+      "Ġsh": 392,
+      "Ġk": 393,
+      "ack": 394,
+      "ect": 395,
+      "ĠâĢ": 396,
+      "ĠU": 397,
+      "Ġhad": 398,
+      "se": 399,
+      "Ġwhich": 400,
+      "red": 401,
+      "ov": 402,
+      "ĠSt": 403,
+      "ast": 404,
+      "Ġsp": 405,
+      "ian": 406,
+      "Ġy": 407,
+      "ment": 408,
+      "Ġle": 409,
+      "Ġnot": 410,
+      "ge": 411,
+      "ord": 412,
+      "rit": 413,
+      "ip": 414,
+      "ine": 415,
+      "ell": 416,
+      "ally": 417,
+      "our": 418,
+      "ost": 419,
+      "ight": 420,
+      "ther": 421,
+      "ap": 422,
+      "Ġu": 423,
+      "ish": 424,
+      "ĠCh": 425,
+      "oun": 426,
+      "ia": 427,
+      "Ġ3": 428,
+      "ave": 429,
+      "ary": 430,
+      "ust": 431,
+      "og": 432,
+      "Ġ200": 433,
+      "Ġun": 434,
+      "ous": 435,
+      "irst": 436,
+      "ĠV": 437,
+      "cc": 438,
+      "Ġinc": 439,
+      "Ġ;": 440,
+      "Ġcomp": 441,
+      "ru": 442,
+      "ions": 443,
+      "Ġtheir": 444,
+      "Ġbut": 445,
+      "ide": 446,
+      "ure": 447,
+      "so": 448,
+      "Ġcont": 449,
+      "Ġint": 450,
+      "fter": 451,
+      "ical": 452,
+      "ial": 453,
+      "Ġar": 454,
+      "Ġfirst": 455,
+      "ould": 456,
+      "Ġits": 457,
+      "hed": 458,
+      "ĠâĢĵ": 459,
+      "Ġwhe": 460,
+      "wo": 461,
+      "out": 462,
+      "ub": 463,
+      "Ġ20": 464,
+      "ff": 465,
+      "Ġ:": 466,
+      "ue": 467,
+      "Ġher": 468,
+      "own": 469,
+      "ok": 470,
+      "Ġalso": 471,
+      "Ġcl": 472,
+      "per": 473,
+      "ign": 474,
+      "ater": 475,
+      "ran": 476,
+      "orm": 477,
+      "ie": 478,
+      "ome": 479,
+      "ork": 480,
+      "ass": 481,
+      "ire": 482,
+      "end": 483,
+      "Ġres": 484,
+      "Ġab": 485,
+      "Ġad": 486,
+      "Ġus": 487,
+      "ry": 488,
+      "Ġrec": 489,
+      "Ġhave": 490,
+      "age": 491,
+      "ĠHe": 492,
+      "Ġ4": 493,
+      "Ġro": 494,
+      "mer": 495,
+      "Ġone": 496,
+      "ond": 497,
+      "low": 498,
+      "Ġhas": 499,
+      "ĠTh": 500,
+      "du": 501,
+      "Ġ5": 502,
+      "Ġper": 503,
+      "Ġbeen": 504,
+      "ime": 505,
+      "Ġtwo": 506,
+      "ence": 507,
+      "land": 508,
+      "Ġ18": 509,
+      ".@": 510,
+      "Ġ@.@": 511,
+      "ult": 512,
+      "ree": 513,
+      "ough": 514,
+      "ile": 515,
+      "Ġwho": 516,
+      "ĠAl": 517,
+      "Ġsc": 518,
+      "uring": 519,
+      "pl": 520,
+      "ory": 521,
+      "ition": 522,
+      "ric": 523,
+      "ations": 524,
+      "Ġdis": 525,
+      "Ġthis": 526,
+      "Ġbec": 527,
+      "Ġapp": 528,
+      "iz": 529,
+      "ĠIt": 530,
+      "are": 531,
+      "ach": 532,
+      "lud": 533,
+      "ade": 534,
+      "Ġplay": 535,
+      "Ġj": 536,
+      "Ġman": 537,
+      "act": 538,
+      "ely": 539,
+      "Ġpart": 540,
+      "Ġdes": 541,
+      "Ġag": 542,
+      "Ġthey": 543,
+      "Ġyear": 544,
+      "ount": 545,
+      "Ġ201": 546,
+      "Ġover": 547,
+      "Ġother": 548,
+      "ound": 549,
+      "Ġafter": 550,
+      "ib": 551,
+      "over": 552,
+      "Ġser": 553,
+      "Ġen": 554,
+      "Ġoff": 555,
+      "Ġim": 556,
+      "ction": 557,
+      "ĠY": 558,
+      "ke": 559,
+      "ite": 560,
+      ",@": 561,
+      "Ġ@,@": 562,
+      "te": 563,
+      "urn": 564,
+      "Ġinclud": 565,
+      "ress": 566,
+      "ance": 567,
+      "ang": 568,
+      "Ġatt": 569,
+      "ice": 570,
+      "ace": 571,
+      "ark": 572,
+      "Ġout": 573,
+      "wn": 574,
+      "ph": 575,
+      "ember": 576,
+      "Ġpre": 577,
+      "Ġup": 578,
+      "ens": 579,
+      "man": 580,
+      "Ġev": 581,
+      "Ġtime": 582,
+      "nder": 583,
+      "rough": 584,
+      "ced": 585,
+      "Ġfin": 586,
+      "Ġinto": 587,
+      "one": 588,
+      "port": 589,
+      "round": 590,
+      "we": 591,
+      "ren": 592,
+      "les": 593,
+      "int": 594,
+      "ĠOn": 595,
+      "vel": 596,
+      "Ġcomm": 597,
+      "Ġshe": 598,
+      "ason": 599,
+      "amp": 600,
+      "Ġte": 601,
+      "Ġwould": 602,
+      "ward": 603,
+      "Ġmore": 604,
+      "Ġ6": 605,
+      "ied": 606,
+      "ose": 607,
+      "rib": 608,
+      "ĠUn": 609,
+      "Ġall": 610,
+      "ings": 611,
+      "tern": 612,
+      "ces": 613,
+      "able": 614,
+      "Ġwe": 615,
+      "ited": 616,
+      "ever": 617,
+      "ents": 618,
+      "Ġhim": 619,
+      "ased": 620,
+      "ors": 621,
+      "oy": 622,
+      "ood": 623,
+      "Ġcent": 624,
+      "ix": 625,
+      "ase": 626,
+      "ild": 627,
+      "ĠAn": 628,
+      "Ġ7": 629,
+      "Ġwork": 630,
+      "ates": 631,
+      "ious": 632,
+      "ath": 633,
+      "Ġpo": 634,
+      "rop": 635,
+      "old": 636,
+      "als": 637,
+      "iss": 638,
+      "ey": 639,
+      "ict": 640,
+      "Ġfe": 641,
+      "Ġthem": 642,
+      "gan": 643,
+      "Ġsec": 644,
+      "Ġbet": 645,
+      "Ġwhen": 646,
+      "Ġsong": 647,
+      "Ġrem": 648,
+      "ep": 649,
+      "form": 650,
+      "ail": 651,
+      "fer": 652,
+      "Ġear": 653,
+      "ubl": 654,
+      "aw": 655,
+      "Ġkn": 656,
+      "ake": 657,
+      "aus": 658,
+      "Ġmost": 659,
+      "Ġcons": 660,
+      "Ġduring": 661,
+      "ĠAs": 662,
+      "orth": 663,
+      "Ġnew": 664,
+      "ered": 665,
+      "ilm": 666,
+      "ved": 667,
+      "att": 668,
+      "Ġonly": 669,
+      "Ġ9": 670,
+      "Ġdec": 671,
+      "Ġ8": 672,
+      "ick": 673,
+      "Ġgame": 674,
+      "ons": 675,
+      "ug": 676,
+      "Ġtr": 677,
+      "ft": 678,
+      "oth": 679,
+      "ook": 680,
+      "ĠMar": 681,
+      "reat": 682,
+      "way": 683,
+      "Ġcan": 684,
+      "ollow": 685,
+      "outh": 686,
+      "ween": 687,
+      "ĠEn": 688,
+      "Ġ199": 689,
+      "ters": 690,
+      "Ġrel": 691,
+      "ind": 692,
+      "Ġabout": 693,
+      "Ġseason": 694,
+      "Ġagain": 695,
+      "ral": 696,
+      "Ġthree": 697,
+      "ational": 698,
+      "Ġunder": 699,
+      "ular": 700,
+      "Ġme": 701,
+      "Ġthan": 702,
+      "ĠCom": 703,
+      "ĠAr": 704,
+      "hip": 705,
+      "ob": 706,
+      "Ġne": 707,
+      "Ġbetween": 708,
+      "Ġfl": 709,
+      "hn": 710,
+      "ve": 711,
+      "Ġchar": 712,
+      "Ġcol": 713,
+      "Ġrecord": 714,
+      "iew": 715,
+      "ron": 716,
+      "fore": 717,
+      "Ġthrough": 718,
+      "ision": 719,
+      "orn": 720,
+      "Ġ00": 721,
+      "ock": 722,
+      "Ġver": 723,
+      "Ġlater": 724,
+      "Ġnum": 725,
+      "Ġend": 726,
+      "olog": 727,
+      "ames": 728,
+      "Ġpos": 729,
+      "Ġwrit": 730,
+      "Ġprodu": 731,
+      "Ġwhile": 732,
+      "Ġact": 733,
+      "Ġrele": 734,
+      "Ġfilm": 735,
+      "ished": 736,
+      "Ġpr": 737,
+      "ans": 738,
+      "Ġreg": 739,
+      "Ġform": 740,
+      "Ġass": 741,
+      "ĠSe": 742,
+      "ury": 743,
+      "ted": 744,
+      "ts": 745,
+      "Ġmade": 746,
+      "Ġsub": 747,
+      "Ġpe": 748,
+      "Ġso": 749,
+      "orld": 750,
+      "Ġret": 751,
+      "ĠNew": 752,
+      "Ġspec": 753,
+      "Ġacc": 754,
+      "Ġqu": 755,
+      "Ġwhere": 756,
+      "ener": 757,
+      "Ġmov": 758,
+      "hes": 759,
+      "meric": 760,
+      "ating": 761,
+      "Ġinter": 762,
+      "ĠLe": 763,
+      "ĠAmeric": 764,
+      "Ġra": 765,
+      "Ġsome": 766,
+      "Ġco": 767,
+      "Ġlar": 768,
+      "Ġbu": 769,
+      "Ġdef": 770,
+      "bum": 771,
+      "Ġac": 772,
+      "Ġmus": 773,
+      "Ġfollow": 774,
+      "ĠAt": 775,
+      "ins": 776,
+      "ived": 777,
+      "ific": 778,
+      "ual": 779,
+      "Ġam": 780,
+      "Ġsuch": 781,
+      "Ġsecond": 782,
+      "ike": 783,
+      "Ġfour": 784,
+      "Ġind": 785,
+      "ann": 786,
+      "hen": 787,
+      "Ġused": 788,
+      "ĠRe": 789,
+      "ics": 790,
+      "lect": 791,
+      "Ġday": 792,
+      "iel": 793,
+      "ily": 794,
+      "ĠThis": 795,
+      "Ġ0": 796,
+      "Ġpubl": 797,
+      "Ġcall": 798,
+      "ĠJo": 799,
+      "ll": 800,
+      "Ġalbum": 801,
+      "Ġ000": 802,
+      "rans": 803,
+      "Ġdo": 804,
+      "any": 805,
+      "Ġbefore": 806,
+      "ros": 807,
+      "ĠSh": 808,
+      "Ġsy": 809,
+      "aid": 810,
+      "ĠEng": 811,
+      "Ġbeing": 812,
+      "Ġ10": 813,
+      "uc": 814,
+      "Ġep": 815,
+      "Ġsupp": 816,
+      "Ġthere": 817,
+      "Ġyears": 818,
+      "ars": 819,
+      "owever": 820,
+      "Ġent": 821,
+      "ife": 822,
+      "Ġhigh": 823,
+      "Ġfound": 824,
+      "ird": 825,
+      "Ġno": 826,
+      "Ġset": 827,
+      "ines": 828,
+      "iver": 829,
+      "io": 830,
+      "other": 831,
+      "ject": 832,
+      "Ġsur": 833,
+      "aj": 834,
+      "ten": 835,
+      "Ġtra": 836,
+      "Ġ12": 837,
+      "ised": 838,
+      "ities": 839,
+      "velop": 840,
+      "Ġbl": 841,
+      "ale": 842,
+      "Ġseries": 843,
+      "Ġloc": 844,
+      "Ġnumber": 845,
+      "Ġpres": 846,
+      "ane": 847,
+      "ause": 848,
+      "ode": 849,
+      "ek": 850,
+      "ton": 851,
+      "ĠSc": 852,
+      "ier": 853,
+      "ise": 854,
+      "Ġsever": 855,
+      "ince": 856,
+      "Ġboth": 857,
+      "ank": 858,
+      "row": 859,
+      "irect": 860,
+      "son": 861,
+      "Ġthen": 862,
+      "ĠBrit": 863,
+      "iet": 864,
+      "Ġ16": 865,
+      "Ġepis": 866,
+      "Ġincluding": 867,
+      "its": 868,
+      "igin": 869,
+      "pr": 870,
+      "Ġ/": 871,
+      "Ġagainst": 872,
+      "Ġwell": 873,
+      "Ġbecame": 874,
+      "Ġexp": 875,
+      "Ġknown": 876,
+      "Ġtrans": 877,
+      "Ġcharac": 878,
+      "ĠâĢĶ": 879,
+      "ram": 880,
+      "Ġback": 881,
+      "Ġadd": 882,
+      "Ġpop": 883,
+      "Ġgo": 884,
+      "urch": 885,
+      "Ġdesc": 886,
+      "Ġsing": 887,
+      "ield": 888,
+      "Ġperform": 889,
+      "ained": 890,
+      "Ġrece": 891,
+      "ident": 892,
+      "Ġem": 893,
+      "ert": 894,
+      "ures": 895,
+      "Ġinv": 896,
+      "Ġdep": 897,
+      "Ġ198": 898,
+      "air": 899,
+      "ern": 900,
+      "ather": 901,
+      "ful": 902,
+      "ĠZ": 903,
+      "Ġmon": 904,
+      "Ġmany": 905,
+      "Ġmain": 906,
+      "Ġstud": 907,
+      "Ġlong": 908,
+      "inn": 909,
+      "though": 910,
+      "up": 911,
+      "ool": 912,
+      "ĠUnited": 913,
+      "led": 914,
+      "ement": 915,
+      "Ġ15": 916,
+      "ower": 917,
+      "ĠJohn": 918,
+      "Ġop": 919,
+      "Ġ11": 920,
+      "ined": 921,
+      "Ġmet": 922,
+      "ober": 923,
+      "ley": 924,
+      "Ġ17": 925,
+      "Ġcentury": 926,
+      "Ġteam": 927,
+      "Ġest": 928,
+      "ĠAfter": 929,
+      "yl": 930,
+      "Ġmin": 931,
+      "uch": 932,
+      "ute": 933,
+      "Ġdevelop": 934,
+      "ĠShe": 935,
+      "iam": 936,
+      "Ġshow": 937,
+      "elf": 938,
+      "Ġrep": 939,
+      "Ġconc": 940,
+      "ative": 941,
+      "Ġcre": 942,
+      "overn": 943,
+      "ared": 944,
+      "Ġ194": 945,
+      "Ġorigin": 946,
+      "Ġsm": 947,
+      "ivers": 948,
+      "az": 949,
+      "Ġlead": 950,
+      "Ġseveral": 951,
+      "ah": 952,
+      "Ġob": 953,
+      "Ġrev": 954,
+      "Ġmill": 955,
+      "erm": 956,
+      "ually": 957,
+      "oot": 958,
+      "Ġbegan": 959,
+      "Ġ196": 960,
+      "ired": 961,
+      "Ġdif": 962,
+      "Ġcontin": 963,
+      "Ġsign": 964,
+      "ik": 965,
+      "ĠInd": 966,
+      "ments": 967,
+      "ized": 968,
+      "Ġ197": 969,
+      "Ġdirect": 970,
+      "au": 971,
+      "Ġext": 972,
+      "ross": 973,
+      "emb": 974,
+      "der": 975,
+      "Ġpol": 976,
+      "Ġmay": 977,
+      "apt": 978,
+      "els": 979,
+      "ĠWh": 980,
+      "Ġcomple": 981,
+      "Ġart": 982,
+      "ĠBr": 983,
+      "ĠIs": 984,
+      "une": 985,
+      "til": 986,
+      "Ġcrit": 987,
+      "Ġhist": 988,
+      "Ġearly": 989,
+      "Ġcould": 990,
+      "ĠCon": 991,
+      "Ġdid": 992,
+      "Ġbel": 993,
+      "Ġcalled": 994,
+      "ued": 995,
+      "Ġnear": 996,
+      "Ġepisode": 997,
+      "yp": 998,
+      "Ġdescrib": 999
+    },
+    "merges": [
+      "Ġ t",
+      "h e",
+      "Ġ a",
+      "i n",
+      "Ġt he",
+      "e r",
+      "o n",
+      "Ġ ,",
+      "r e",
+      "Ġ s",
+      "e d",
+      "Ġ o",
+      "Ġ w",
+      "n d",
+      "a t",
+      "Ġ .",
+      "o r",
+      "i t",
+      "Ġ c",
+      "e n",
+      "Ġ f",
+      "i s",
+      "e s",
+      "a r",
+      "Ġo f",
+      "Ġ b",
+      "a n",
+      "Ġ in",
+      "a l",
+      "in g",
+      "Ġ p",
+      "Ġa nd",
+      "a s",
+      "Ġt o",
+      "r o",
+      "i c",
+      "Ġ m",
+      "Ġ d",
+      "Ġ h",
+      "i on",
+      "l e",
+      "o u",
+      "Ġ T",
+      "Ġ re",
+      "Ġ =",
+      "Ġ \"",
+      "Ġ A",
+      "Ġ S",
+      "en t",
+      "i l",
+      "Ġt h",
+      "Ġ 1",
+      "s t",
+      "Ġ C",
+      "e l",
+      "o m",
+      "Ġ l",
+      "a m",
+      "Ġ Ċ",
+      "Ġ e",
+      "Ġ n",
+      "Ġ @",
+      "a d",
+      "a c",
+      "Ġw as",
+      "Ġ M",
+      "u r",
+      "ĠT he",
+      "e c",
+      "Ġ on",
+      "l y",
+      "Ġ B",
+      "Ġ I",
+      "Ġ g",
+      "Ġ '",
+      "e t",
+      "o l",
+      "i d",
+      "i v",
+      "i m",
+      "Ġf or",
+      "i r",
+      "- @",
+      "Ġ@ -@",
+      "i g",
+      "o t",
+      "t er",
+      "Ġa s",
+      "Ġ H",
+      "u s",
+      "o w",
+      "Ġs t",
+      "u t",
+      "it h",
+      "a y",
+      "Ġ 2",
+      "Ġ P",
+      "at ion",
+      "v er",
+      "Ġb e",
+      "he r",
+      "Ġth at",
+      "Ġw ith",
+      "Ġ R",
+      "c e",
+      "t h",
+      "Ġ D",
+      "Ġ is",
+      "u n",
+      "e m",
+      "Ġ F",
+      "Ġw h",
+      "u l",
+      "Ġb y",
+      "Ġa l",
+      "c h",
+      "Ġ )",
+      "Ġ (",
+      "Ġ W",
+      "Ġc on",
+      "r a",
+      "Ġ G",
+      "o s",
+      "Ġ L",
+      "Ġ N",
+      "Ġa t",
+      "er s",
+      "c t",
+      "Ġ it",
+      "Ġ1 9",
+      "ro m",
+      "a nd",
+      "Ġa n",
+      "u m",
+      "es t",
+      "Ġ J",
+      "a g",
+      "Ġ he",
+      "0 0",
+      "is t",
+      "a in",
+      "o d",
+      "a v",
+      "r i",
+      "Ġ E",
+      "Ġ O",
+      "Ġf rom",
+      "Ġc om",
+      "Ġh is",
+      "o p",
+      "Ġp ro",
+      "re s",
+      "i es",
+      "i f",
+      "Ġ v",
+      "or t",
+      "er e",
+      "il l",
+      "l d",
+      "Ġd e",
+      "p p",
+      "Ġs u",
+      "o re",
+      "ĠI n",
+      "Ġ r",
+      "Ġs e",
+      "Ġw ere",
+      "e w",
+      "on g",
+      "ig h",
+      "ar d",
+      "at e",
+      "al l",
+      "ar t",
+      "a k",
+      "ic h",
+      "Ġc h",
+      "Ġo r",
+      "a b",
+      "an t",
+      "u d",
+      "o c",
+      "b er",
+      "Ġe x",
+      "g h",
+      "it y",
+      "at ed",
+      "p t",
+      "es s",
+      "e ar",
+      "Ġ K",
+      "Ġp l",
+      "am e",
+      "q u",
+      "iv e",
+      "ro u",
+      "Ġa re",
+      "Ġ â",
+      "Ġs h",
+      "Ġ k",
+      "ac k",
+      "ec t",
+      "Ġâ Ģ",
+      "Ġ U",
+      "Ġh ad",
+      "s e",
+      "Ġwh ich",
+      "re d",
+      "o v",
+      "ĠS t",
+      "as t",
+      "Ġs p",
+      "i an",
+      "Ġ y",
+      "m ent",
+      "Ġ le",
+      "Ġn ot",
+      "g e",
+      "or d",
+      "r it",
+      "i p",
+      "in e",
+      "el l",
+      "al ly",
+      "ou r",
+      "o st",
+      "igh t",
+      "t her",
+      "a p",
+      "Ġ u",
+      "is h",
+      "ĠC h",
+      "ou n",
+      "i a",
+      "Ġ 3",
+      "av e",
+      "ar y",
+      "u st",
+      "o g",
+      "Ġ2 00",
+      "Ġ un",
+      "ou s",
+      "ir st",
+      "Ġ V",
+      "c c",
+      "Ġin c",
+      "Ġ ;",
+      "Ġcom p",
+      "r u",
+      "ion s",
+      "Ġthe ir",
+      "Ġb ut",
+      "id e",
+      "u re",
+      "s o",
+      "Ġcon t",
+      "Ġin t",
+      "f ter",
+      "ic al",
+      "i al",
+      "Ġa r",
+      "Ġf irst",
+      "ou ld",
+      "Ġit s",
+      "he d",
+      "ĠâĢ ĵ",
+      "Ġw he",
+      "w o",
+      "ou t",
+      "u b",
+      "Ġ2 0",
+      "f f",
+      "Ġ :",
+      "u e",
+      "Ġ her",
+      "ow n",
+      "o k",
+      "Ġal so",
+      "Ġc l",
+      "p er",
+      "ig n",
+      "at er",
+      "r an",
+      "or m",
+      "i e",
+      "om e",
+      "or k",
+      "as s",
+      "i re",
+      "e nd",
+      "Ġre s",
+      "Ġa b",
+      "Ġa d",
+      "Ġ us",
+      "r y",
+      "Ġre c",
+      "Ġh ave",
+      "ag e",
+      "ĠH e",
+      "Ġ 4",
+      "Ġ ro",
+      "m er",
+      "Ġon e",
+      "on d",
+      "l ow",
+      "Ġh as",
+      "ĠT h",
+      "d u",
+      "Ġ 5",
+      "Ġp er",
+      "Ġbe en",
+      "im e",
+      "Ġt wo",
+      "en ce",
+      "l and",
+      "Ġ1 8",
+      ". @",
+      "Ġ@ .@",
+      "ul t",
+      "re e",
+      "ou gh",
+      "i le",
+      "Ġwh o",
+      "ĠA l",
+      "Ġs c",
+      "ur ing",
+      "p l",
+      "or y",
+      "it ion",
+      "r ic",
+      "ation s",
+      "Ġd is",
+      "Ġth is",
+      "Ġb ec",
+      "Ġa pp",
+      "i z",
+      "ĠI t",
+      "a re",
+      "ac h",
+      "l ud",
+      "ad e",
+      "Ġpl ay",
+      "Ġ j",
+      "Ġm an",
+      "ac t",
+      "el y",
+      "Ġp art",
+      "Ġd es",
+      "Ġa g",
+      "Ġthe y",
+      "Ġy ear",
+      "oun t",
+      "Ġ20 1",
+      "Ġo ver",
+      "Ġo ther",
+      "ou nd",
+      "Ġa fter",
+      "i b",
+      "o ver",
+      "Ġs er",
+      "Ġ en",
+      "Ġof f",
+      "Ġ im",
+      "ct ion",
+      "Ġ Y",
+      "k e",
+      "it e",
+      ", @",
+      "Ġ@ ,@",
+      "t e",
+      "ur n",
+      "Ġinc lud",
+      "res s",
+      "an ce",
+      "an g",
+      "Ġat t",
+      "ic e",
+      "ac e",
+      "ar k",
+      "Ġo ut",
+      "w n",
+      "p h",
+      "em ber",
+      "Ġp re",
+      "Ġu p",
+      "en s",
+      "m an",
+      "Ġe v",
+      "Ġt ime",
+      "nd er",
+      "rou gh",
+      "c ed",
+      "Ġf in",
+      "Ġint o",
+      "on e",
+      "p ort",
+      "rou nd",
+      "w e",
+      "re n",
+      "l es",
+      "in t",
+      "ĠO n",
+      "v el",
+      "Ġcom m",
+      "Ġs he",
+      "as on",
+      "am p",
+      "Ġt e",
+      "Ġw ould",
+      "w ard",
+      "Ġm ore",
+      "Ġ 6",
+      "i ed",
+      "os e",
+      "ri b",
+      "ĠU n",
+      "Ġal l",
+      "ing s",
+      "ter n",
+      "c es",
+      "ab le",
+      "Ġw e",
+      "it ed",
+      "e ver",
+      "ent s",
+      "Ġh im",
+      "as ed",
+      "or s",
+      "o y",
+      "o od",
+      "Ġc ent",
+      "i x",
+      "as e",
+      "il d",
+      "ĠA n",
+      "Ġ 7",
+      "Ġw ork",
+      "at es",
+      "i ous",
+      "at h",
+      "Ġp o",
+      "ro p",
+      "ol d",
+      "al s",
+      "is s",
+      "e y",
+      "ic t",
+      "Ġf e",
+      "Ġthe m",
+      "g an",
+      "Ġs ec",
+      "Ġb et",
+      "Ġwhe n",
+      "Ġs ong",
+      "Ġre m",
+      "e p",
+      "f orm",
+      "a il",
+      "f er",
+      "Ġe ar",
+      "ub l",
+      "a w",
+      "Ġk n",
+      "ak e",
+      "a us",
+      "Ġm ost",
+      "Ġcon s",
+      "Ġd uring",
+      "ĠA s",
+      "or th",
+      "Ġn ew",
+      "er ed",
+      "il m",
+      "v ed",
+      "at t",
+      "Ġon ly",
+      "Ġ 9",
+      "Ġd ec",
+      "Ġ 8",
+      "ic k",
+      "Ġg ame",
+      "on s",
+      "u g",
+      "Ġt r",
+      "f t",
+      "ot h",
+      "o ok",
+      "ĠM ar",
+      "re at",
+      "w ay",
+      "Ġc an",
+      "ol low",
+      "ou th",
+      "we en",
+      "ĠE n",
+      "Ġ19 9",
+      "ter s",
+      "Ġre l",
+      "in d",
+      "Ġab out",
+      "Ġse ason",
+      "Ġag ain",
+      "r al",
+      "Ġth ree",
+      "ation al",
+      "Ġu nder",
+      "ul ar",
+      "Ġm e",
+      "Ġth an",
+      "ĠC om",
+      "ĠA r",
+      "h ip",
+      "o b",
+      "Ġn e",
+      "Ġbet ween",
+      "Ġf l",
+      "h n",
+      "v e",
+      "Ġch ar",
+      "Ġc ol",
+      "Ġrec ord",
+      "i ew",
+      "r on",
+      "f ore",
+      "Ġth rough",
+      "is ion",
+      "or n",
+      "Ġ 00",
+      "oc k",
+      "Ġ ver",
+      "Ġl ater",
+      "Ġn um",
+      "Ġe nd",
+      "ol og",
+      "am es",
+      "Ġp os",
+      "Ġw rit",
+      "Ġpro du",
+      "Ġwh ile",
+      "Ġa ct",
+      "Ġre le",
+      "Ġf ilm",
+      "is hed",
+      "Ġp r",
+      "an s",
+      "Ġre g",
+      "Ġfor m",
+      "Ġas s",
+      "ĠS e",
+      "ur y",
+      "t ed",
+      "t s",
+      "Ġm ade",
+      "Ġsu b",
+      "Ġp e",
+      "Ġs o",
+      "or ld",
+      "Ġre t",
+      "ĠN ew",
+      "Ġsp ec",
+      "Ġa cc",
+      "Ġ qu",
+      "Ġwhe re",
+      "en er",
+      "Ġm ov",
+      "he s",
+      "mer ic",
+      "at ing",
+      "Ġin ter",
+      "ĠL e",
+      "ĠA meric",
+      "Ġ ra",
+      "Ġs ome",
+      "Ġc o",
+      "Ġl ar",
+      "Ġb u",
+      "Ġde f",
+      "b um",
+      "Ġa c",
+      "Ġm us",
+      "Ġf ollow",
+      "ĠA t",
+      "in s",
+      "iv ed",
+      "if ic",
+      "u al",
+      "Ġa m",
+      "Ġsu ch",
+      "Ġsec ond",
+      "i ke",
+      "Ġf our",
+      "Ġin d",
+      "an n",
+      "he n",
+      "Ġus ed",
+      "ĠR e",
+      "ic s",
+      "le ct",
+      "Ġd ay",
+      "i el",
+      "il y",
+      "ĠTh is",
+      "Ġ 0",
+      "Ġp ubl",
+      "Ġc all",
+      "ĠJ o",
+      "l l",
+      "Ġal bum",
+      "Ġ00 0",
+      "ran s",
+      "Ġd o",
+      "an y",
+      "Ġbe fore",
+      "ro s",
+      "ĠS h",
+      "Ġs y",
+      "a id",
+      "ĠEn g",
+      "Ġbe ing",
+      "Ġ1 0",
+      "u c",
+      "Ġe p",
+      "Ġsu pp",
+      "Ġthe re",
+      "Ġyear s",
+      "ar s",
+      "ow ever",
+      "Ġ ent",
+      "if e",
+      "Ġh igh",
+      "Ġf ound",
+      "ir d",
+      "Ġn o",
+      "Ġs et",
+      "in es",
+      "iv er",
+      "i o",
+      "ot her",
+      "j ect",
+      "Ġs ur",
+      "a j",
+      "t en",
+      "Ġt ra",
+      "Ġ1 2",
+      "is ed",
+      "it ies",
+      "vel op",
+      "Ġb l",
+      "al e",
+      "Ġser ies",
+      "Ġl oc",
+      "Ġnum ber",
+      "Ġp res",
+      "an e",
+      "aus e",
+      "od e",
+      "e k",
+      "t on",
+      "ĠS c",
+      "i er",
+      "is e",
+      "Ġse ver",
+      "in ce",
+      "Ġb oth",
+      "an k",
+      "ro w",
+      "ire ct",
+      "s on",
+      "Ġthe n",
+      "ĠB rit",
+      "i et",
+      "Ġ1 6",
+      "Ġep is",
+      "Ġinclud ing",
+      "it s",
+      "ig in",
+      "p r",
+      "Ġ /",
+      "Ġagain st",
+      "Ġw ell",
+      "Ġbec ame",
+      "Ġex p",
+      "Ġkn own",
+      "Ġt rans",
+      "Ġchar ac",
+      "ĠâĢ Ķ",
+      "r am",
+      "Ġb ack",
+      "Ġad d",
+      "Ġp op",
+      "Ġg o",
+      "ur ch",
+      "Ġdes c",
+      "Ġs ing",
+      "iel d",
+      "Ġper form",
+      "ain ed",
+      "Ġre ce",
+      "id ent",
+      "Ġe m",
+      "er t",
+      "u res",
+      "Ġin v",
+      "Ġde p",
+      "Ġ19 8",
+      "a ir",
+      "er n",
+      "at her",
+      "f ul",
+      "Ġ Z",
+      "Ġm on",
+      "Ġman y",
+      "Ġm ain",
+      "Ġst ud",
+      "Ġl ong",
+      "in n",
+      "th ough",
+      "u p",
+      "o ol",
+      "ĠUn ited",
+      "l ed",
+      "em ent",
+      "Ġ1 5",
+      "ow er",
+      "ĠJo hn",
+      "Ġo p",
+      "Ġ1 1",
+      "in ed",
+      "Ġm et",
+      "o ber",
+      "le y",
+      "Ġ1 7",
+      "Ġcent ury",
+      "Ġte am",
+      "Ġ est",
+      "ĠA fter",
+      "y l",
+      "Ġm in",
+      "u ch",
+      "ut e",
+      "Ġde velop",
+      "ĠS he",
+      "i am",
+      "Ġsh ow",
+      "el f",
+      "Ġre p",
+      "Ġcon c",
+      "at ive",
+      "Ġc re",
+      "over n",
+      "a red",
+      "Ġ19 4",
+      "Ġor igin",
+      "Ġs m",
+      "iv ers",
+      "a z",
+      "Ġle ad",
+      "Ġsever al",
+      "a h",
+      "Ġo b",
+      "Ġre v",
+      "Ġm ill",
+      "er m",
+      "u ally",
+      "o ot",
+      "Ġbe gan",
+      "Ġ19 6",
+      "i red",
+      "Ġd if",
+      "Ġcont in",
+      "Ġs ign",
+      "i k",
+      "ĠI nd",
+      "ment s",
+      "iz ed",
+      "Ġ19 7",
+      "Ġd irect",
+      "a u",
+      "Ġex t",
+      "ros s",
+      "em b",
+      "d er",
+      "Ġp ol",
+      "Ġm ay",
+      "a pt",
+      "el s",
+      "ĠW h",
+      "Ġcomp le",
+      "Ġar t",
+      "ĠB r",
+      "ĠI s",
+      "un e",
+      "t il",
+      "Ġc rit",
+      "Ġh ist",
+      "Ġear ly",
+      "Ġc ould",
+      "ĠC on",
+      "Ġd id",
+      "Ġb el",
+      "Ġcall ed",
+      "u ed",
+      "Ġn ear",
+      "Ġepis ode",
+      "y p",
+      "Ġdesc rib"
+    ]
+  }
+}

tokenizer_config.json

@@ -0,0 +1,10 @@
+{
+  "add_prefix_space": false,
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 1024,
+  "name_or_path": "hf-internal-testing/tiny-random-gpt2",
+  "special_tokens_map_file": null,
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}

vocab.json

@@ -0,0 +1 @@
+{"<|endoftext|>":0,"!":1,"\"":2,"#":3,"$":4,"%":5,"&":6,"'":7,"(":8,")":9,"*":10,"+":11,",":12,"-":13,".":14,"/":15,"0":16,"1":17,"2":18,"3":19,"4":20,"5":21,"6":22,"7":23,"8":24,"9":25,":":26,";":27,"<":28,"=":29,">":30,"?":31,"@":32,"A":33,"B":34,"C":35,"D":36,"E":37,"F":38,"G":39,"H":40,"I":41,"J":42,"K":43,"L":44,"M":45,"N":46,"O":47,"P":48,"Q":49,"R":50,"S":51,"T":52,"U":53,"V":54,"W":55,"X":56,"Y":57,"Z":58,"[":59,"\\":60,"]":61,"^":62,"_":63,"`":64,"a":65,"b":66,"c":67,"d":68,"e":69,"f":70,"g":71,"h":72,"i":73,"j":74,"k":75,"l":76,"m":77,"n":78,"o":79,"p":80,"q":81,"r":82,"s":83,"t":84,"u":85,"v":86,"w":87,"x":88,"y":89,"z":90,"|":91,"}":92,"~":93,"¡":94,"¢":95,"£":96,"¤":97,"¥":98,"¦":99,"§":100,"¨":101,"©":102,"ª":103,"«":104,"¬":105,"®":106,"¯":107,"°":108,"±":109,"²":110,"³":111,"´":112,"µ":113,"¶":114,"·":115,"¸":116,"¹":117,"º":118,"»":119,"¼":120,"½":121,"¾":122,"¿":123,"Â":124,"Ã":125,"Ä":126,"Å":127,"Æ":128,"Ç":129,"È":130,"É":131,"Ê":132,"Ë":133,"Ì":134,"Í":135,"Î":136,"Ï":137,"Ð":138,"Ñ":139,"Ö":140,"×":141,"Ø":142,"Ù":143,"Ü":144,"à":145,"á":146,"â":147,"ã":148,"ä":149,"å":150,"æ":151,"ç":152,"è":153,"é":154,"ë":155,"ì":156,"ï":157,"Ċ":158,"Ġ":159,"Ģ":160,"ģ":161,"Ĥ":162,"ĥ":163,"Ħ":164,"ħ":165,"Ĩ":166,"ĩ":167,"Ī":168,"ī":169,"Ĭ":170,"ĭ":171,"Į":172,"į":173,"İ":174,"ı":175,"IJ":176,"ij":177,"Ĵ":178,"ĵ":179,"Ķ":180,"ķ":181,"ĸ":182,"Ĺ":183,"ĺ":184,"Ļ":185,"ļ":186,"Ľ":187,"ľ":188,"Ŀ":189,"ŀ":190,"Ł":191,"ł":192,"Ń":193,"Ġt":194,"he":195,"Ġa":196,"in":197,"Ġthe":198,"er":199,"on":200,"Ġ,":201,"re":202,"Ġs":203,"ed":204,"Ġo":205,"Ġw":206,"nd":207,"at":208,"Ġ.":209,"or":210,"it":211,"Ġc":212,"en":213,"Ġf":214,"is":215,"es":216,"ar":217,"Ġof":218,"Ġb":219,"an":220,"Ġin":221,"al":222,"ing":223,"Ġp":224,"Ġand":225,"as":226,"Ġto":227,"ro":228,"ic":229,"Ġm":230,"Ġd":231,"Ġh":232,"ion":233,"le":234,"ou":235,"ĠT":236,"Ġre":237,"Ġ=":238,"Ġ\"":239,"ĠA":240,"ĠS":241,"ent":242,"il":243,"Ġth":244,"Ġ1":245,"st":246,"ĠC":247,"el":248,"om":249,"Ġl":250,"am":251,"ĠĊ":252,"Ġe":253,"Ġn":254,"Ġ@":255,"ad":256,"ac":257,"Ġwas":258,"ĠM":259,"ur":260,"ĠThe":261,"ec":262,"Ġon":263,"ly":264,"ĠB":265,"ĠI":266,"Ġg":267,"Ġ'":268,"et":269,"ol":270,"id":271,"iv":272,"im":273,"Ġfor":274,"ir":275,"-@":276,"Ġ@-@":277,"ig":278,"ot":279,"ter":280,"Ġas":281,"ĠH":282,"us":283,"ow":284,"Ġst":285,"ut":286,"ith":287,"ay":288,"Ġ2":289,"ĠP":290,"ation":291,"ver":292,"Ġbe":293,"her":294,"Ġthat":295,"Ġwith":296,"ĠR":297,"ce":298,"th":299,"ĠD":300,"Ġis":301,"un":302,"em":303,"ĠF":304,"Ġwh":305,"ul":306,"Ġby":307,"Ġal":308,"ch":309,"Ġ)":310,"Ġ(":311,"ĠW":312,"Ġcon":313,"ra":314,"ĠG":315,"os":316,"ĠL":317,"ĠN":318,"Ġat":319,"ers":320,"ct":321,"Ġit":322,"Ġ19":323,"rom":324,"and":325,"Ġan":326,"um":327,"est":328,"ĠJ":329,"ag":330,"Ġhe":331,"00":332,"ist":333,"ain":334,"od":335,"av":336,"ri":337,"ĠE":338,"ĠO":339,"Ġfrom":340,"Ġcom":341,"Ġhis":342,"op":343,"Ġpro":344,"res":345,"ies":346,"if":347,"Ġv":348,"ort":349,"ere":350,"ill":351,"ld":352,"Ġde":353,"pp":354,"Ġsu":355,"ore":356,"ĠIn":357,"Ġr":358,"Ġse":359,"Ġwere":360,"ew":361,"ong":362,"igh":363,"ard":364,"ate":365,"all":366,"art":367,"ak":368,"ich":369,"Ġch":370,"Ġor":371,"ab":372,"ant":373,"ud":374,"oc":375,"ber":376,"Ġex":377,"gh":378,"ity":379,"ated":380,"pt":381,"ess":382,"ear":383,"ĠK":384,"Ġpl":385,"ame":386,"qu":387,"ive":388,"rou":389,"Ġare":390,"Ġâ":391,"Ġsh":392,"Ġk":393,"ack":394,"ect":395,"ĠâĢ":396,"ĠU":397,"Ġhad":398,"se":399,"Ġwhich":400,"red":401,"ov":402,"ĠSt":403,"ast":404,"Ġsp":405,"ian":406,"Ġy":407,"ment":408,"Ġle":409,"Ġnot":410,"ge":411,"ord":412,"rit":413,"ip":414,"ine":415,"ell":416,"ally":417,"our":418,"ost":419,"ight":420,"ther":421,"ap":422,"Ġu":423,"ish":424,"ĠCh":425,"oun":426,"ia":427,"Ġ3":428,"ave":429,"ary":430,"ust":431,"og":432,"Ġ200":433,"Ġun":434,"ous":435,"irst":436,"ĠV":437,"cc":438,"Ġinc":439,"Ġ;":440,"Ġcomp":441,"ru":442,"ions":443,"Ġtheir":444,"Ġbut":445,"ide":446,"ure":447,"so":448,"Ġcont":449,"Ġint":450,"fter":451,"ical":452,"ial":453,"Ġar":454,"Ġfirst":455,"ould":456,"Ġits":457,"hed":458,"ĠâĢĵ":459,"Ġwhe":460,"wo":461,"out":462,"ub":463,"Ġ20":464,"ff":465,"Ġ:":466,"ue":467,"Ġher":468,"own":469,"ok":470,"Ġalso":471,"Ġcl":472,"per":473,"ign":474,"ater":475,"ran":476,"orm":477,"ie":478,"ome":479,"ork":480,"ass":481,"ire":482,"end":483,"Ġres":484,"Ġab":485,"Ġad":486,"Ġus":487,"ry":488,"Ġrec":489,"Ġhave":490,"age":491,"ĠHe":492,"Ġ4":493,"Ġro":494,"mer":495,"Ġone":496,"ond":497,"low":498,"Ġhas":499,"ĠTh":500,"du":501,"Ġ5":502,"Ġper":503,"Ġbeen":504,"ime":505,"Ġtwo":506,"ence":507,"land":508,"Ġ18":509,".@":510,"Ġ@.@":511,"ult":512,"ree":513,"ough":514,"ile":515,"Ġwho":516,"ĠAl":517,"Ġsc":518,"uring":519,"pl":520,"ory":521,"ition":522,"ric":523,"ations":524,"Ġdis":525,"Ġthis":526,"Ġbec":527,"Ġapp":528,"iz":529,"ĠIt":530,"are":531,"ach":532,"lud":533,"ade":534,"Ġplay":535,"Ġj":536,"Ġman":537,"act":538,"ely":539,"Ġpart":540,"Ġdes":541,"Ġag":542,"Ġthey":543,"Ġyear":544,"ount":545,"Ġ201":546,"Ġover":547,"Ġother":548,"ound":549,"Ġafter":550,"ib":551,"over":552,"Ġser":553,"Ġen":554,"Ġoff":555,"Ġim":556,"ction":557,"ĠY":558,"ke":559,"ite":560,",@":561,"Ġ@,@":562,"te":563,"urn":564,"Ġinclud":565,"ress":566,"ance":567,"ang":568,"Ġatt":569,"ice":570,"ace":571,"ark":572,"Ġout":573,"wn":574,"ph":575,"ember":576,"Ġpre":577,"Ġup":578,"ens":579,"man":580,"Ġev":581,"Ġtime":582,"nder":583,"rough":584,"ced":585,"Ġfin":586,"Ġinto":587,"one":588,"port":589,"round":590,"we":591,"ren":592,"les":593,"int":594,"ĠOn":595,"vel":596,"Ġcomm":597,"Ġshe":598,"ason":599,"amp":600,"Ġte":601,"Ġwould":602,"ward":603,"Ġmore":604,"Ġ6":605,"ied":606,"ose":607,"rib":608,"ĠUn":609,"Ġall":610,"ings":611,"tern":612,"ces":613,"able":614,"Ġwe":615,"ited":616,"ever":617,"ents":618,"Ġhim":619,"ased":620,"ors":621,"oy":622,"ood":623,"Ġcent":624,"ix":625,"ase":626,"ild":627,"ĠAn":628,"Ġ7":629,"Ġwork":630,"ates":631,"ious":632,"ath":633,"Ġpo":634,"rop":635,"old":636,"als":637,"iss":638,"ey":639,"ict":640,"Ġfe":641,"Ġthem":642,"gan":643,"Ġsec":644,"Ġbet":645,"Ġwhen":646,"Ġsong":647,"Ġrem":648,"ep":649,"form":650,"ail":651,"fer":652,"Ġear":653,"ubl":654,"aw":655,"Ġkn":656,"ake":657,"aus":658,"Ġmost":659,"Ġcons":660,"Ġduring":661,"ĠAs":662,"orth":663,"Ġnew":664,"ered":665,"ilm":666,"ved":667,"att":668,"Ġonly":669,"Ġ9":670,"Ġdec":671,"Ġ8":672,"ick":673,"Ġgame":674,"ons":675,"ug":676,"Ġtr":677,"ft":678,"oth":679,"ook":680,"ĠMar":681,"reat":682,"way":683,"Ġcan":684,"ollow":685,"outh":686,"ween":687,"ĠEn":688,"Ġ199":689,"ters":690,"Ġrel":691,"ind":692,"Ġabout":693,"Ġseason":694,"Ġagain":695,"ral":696,"Ġthree":697,"ational":698,"Ġunder":699,"ular":700,"Ġme":701,"Ġthan":702,"ĠCom":703,"ĠAr":704,"hip":705,"ob":706,"Ġne":707,"Ġbetween":708,"Ġfl":709,"hn":710,"ve":711,"Ġchar":712,"Ġcol":713,"Ġrecord":714,"iew":715,"ron":716,"fore":717,"Ġthrough":718,"ision":719,"orn":720,"Ġ00":721,"ock":722,"Ġver":723,"Ġlater":724,"Ġnum":725,"Ġend":726,"olog":727,"ames":728,"Ġpos":729,"Ġwrit":730,"Ġprodu":731,"Ġwhile":732,"Ġact":733,"Ġrele":734,"Ġfilm":735,"ished":736,"Ġpr":737,"ans":738,"Ġreg":739,"Ġform":740,"Ġass":741,"ĠSe":742,"ury":743,"ted":744,"ts":745,"Ġmade":746,"Ġsub":747,"Ġpe":748,"Ġso":749,"orld":750,"Ġret":751,"ĠNew":752,"Ġspec":753,"Ġacc":754,"Ġqu":755,"Ġwhere":756,"ener":757,"Ġmov":758,"hes":759,"meric":760,"ating":761,"Ġinter":762,"ĠLe":763,"ĠAmeric":764,"Ġra":765,"Ġsome":766,"Ġco":767,"Ġlar":768,"Ġbu":769,"Ġdef":770,"bum":771,"Ġac":772,"Ġmus":773,"Ġfollow":774,"ĠAt":775,"ins":776,"ived":777,"ific":778,"ual":779,"Ġam":780,"Ġsuch":781,"Ġsecond":782,"ike":783,"Ġfour":784,"Ġind":785,"ann":786,"hen":787,"Ġused":788,"ĠRe":789,"ics":790,"lect":791,"Ġday":792,"iel":793,"ily":794,"ĠThis":795,"Ġ0":796,"Ġpubl":797,"Ġcall":798,"ĠJo":799,"ll":800,"Ġalbum":801,"Ġ000":802,"rans":803,"Ġdo":804,"any":805,"Ġbefore":806,"ros":807,"ĠSh":808,"Ġsy":809,"aid":810,"ĠEng":811,"Ġbeing":812,"Ġ10":813,"uc":814,"Ġep":815,"Ġsupp":816,"Ġthere":817,"Ġyears":818,"ars":819,"owever":820,"Ġent":821,"ife":822,"Ġhigh":823,"Ġfound":824,"ird":825,"Ġno":826,"Ġset":827,"ines":828,"iver":829,"io":830,"other":831,"ject":832,"Ġsur":833,"aj":834,"ten":835,"Ġtra":836,"Ġ12":837,"ised":838,"ities":839,"velop":840,"Ġbl":841,"ale":842,"Ġseries":843,"Ġloc":844,"Ġnumber":845,"Ġpres":846,"ane":847,"ause":848,"ode":849,"ek":850,"ton":851,"ĠSc":852,"ier":853,"ise":854,"Ġsever":855,"ince":856,"Ġboth":857,"ank":858,"row":859,"irect":860,"son":861,"Ġthen":862,"ĠBrit":863,"iet":864,"Ġ16":865,"Ġepis":866,"Ġincluding":867,"its":868,"igin":869,"pr":870,"Ġ/":871,"Ġagainst":872,"Ġwell":873,"Ġbecame":874,"Ġexp":875,"Ġknown":876,"Ġtrans":877,"Ġcharac":878,"ĠâĢĶ":879,"ram":880,"Ġback":881,"Ġadd":882,"Ġpop":883,"Ġgo":884,"urch":885,"Ġdesc":886,"Ġsing":887,"ield":888,"Ġperform":889,"ained":890,"Ġrece":891,"ident":892,"Ġem":893,"ert":894,"ures":895,"Ġinv":896,"Ġdep":897,"Ġ198":898,"air":899,"ern":900,"ather":901,"ful":902,"ĠZ":903,"Ġmon":904,"Ġmany":905,"Ġmain":906,"Ġstud":907,"Ġlong":908,"inn":909,"though":910,"up":911,"ool":912,"ĠUnited":913,"led":914,"ement":915,"Ġ15":916,"ower":917,"ĠJohn":918,"Ġop":919,"Ġ11":920,"ined":921,"Ġmet":922,"ober":923,"ley":924,"Ġ17":925,"Ġcentury":926,"Ġteam":927,"Ġest":928,"ĠAfter":929,"yl":930,"Ġmin":931,"uch":932,"ute":933,"Ġdevelop":934,"ĠShe":935,"iam":936,"Ġshow":937,"elf":938,"Ġrep":939,"Ġconc":940,"ative":941,"Ġcre":942,"overn":943,"ared":944,"Ġ194":945,"Ġorigin":946,"Ġsm":947,"ivers":948,"az":949,"Ġlead":950,"Ġseveral":951,"ah":952,"Ġob":953,"Ġrev":954,"Ġmill":955,"erm":956,"ually":957,"oot":958,"Ġbegan":959,"Ġ196":960,"ired":961,"Ġdif":962,"Ġcontin":963,"Ġsign":964,"ik":965,"ĠInd":966,"ments":967,"ized":968,"Ġ197":969,"Ġdirect":970,"au":971,"Ġext":972,"ross":973,"emb":974,"der":975,"Ġpol":976,"Ġmay":977,"apt":978,"els":979,"ĠWh":980,"Ġcomple":981,"Ġart":982,"ĠBr":983,"ĠIs":984,"une":985,"til":986,"Ġcrit":987,"Ġhist":988,"Ġearly":989,"Ġcould":990,"ĠCon":991,"Ġdid":992,"Ġbel":993,"Ġcalled":994,"ued":995,"Ġnear":996,"Ġepisode":997,"yp":998,"Ġdescrib":999}