modeling_gpt_jiang.py

# coding=utf-8
# Copyright 2023 EleutherAI The HuggingFace Inc. team. and JIANG.ai 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 GPTJiang model."""

from typing import Optional, Tuple, Union

import torch
import torch.utils.checkpoint
from torch import nn
from torch.nn import CrossEntropyLoss
import torch.nn.functional as F

from transformers.activations import ACT2FN
from transformers.file_utils import (
    add_code_sample_docstrings,
    add_start_docstrings,
    add_start_docstrings_to_model_forward,
    replace_return_docstrings,
)
from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast
from transformers.modeling_utils import PreTrainedModel
from transformers.utils import logging
from .configuration_gpt_jiang import GPTJiangConfig


logger = logging.get_logger(__name__)

_CONFIG_FOR_DOC = "GPTJiangConfig"
GPT_JIANG_PRETRAINED_MODEL_ARCHIVE_LIST = []


class RMSNorm(torch.nn.Module):
    def __init__(self, dim: int, eps: float=1e-5):
        super().__init__()
        self.eps = eps
        self.weight = nn.Parameter(torch.ones(dim))

    def _norm(self, x):
        return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps)

    def forward(self, x):
        output = self._norm(x.float()).type_as(x)
        return output * self.weight


class GPTJiangPreTrainedModel(PreTrainedModel):
    """
    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
    models.
    """
    config_class = GPTJiangConfig
    base_model_prefix = "gpt_jiang"
    supports_gradient_checkpointing = True
    _no_split_modules = ["GPTJiangLayer"]

    def _init_weights(self, module):
        """Initialize the weights"""
        if isinstance(module, GatedLinear):
            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
            if module.bias is not None:
                module.bias.data.fill_(1.0)
        elif isinstance(module, nn.Linear):
            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, RMSNorm):
            # module.bias.data.zero_()
            module.weight.data.fill_(1.0)

    def _set_gradient_checkpointing(self, module, value=False):
        if isinstance(module, GPTJiangModel):
            module.gradient_checkpointing = value


class GPTJiangAttention(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.max_position_embeddings = config.max_position_embeddings
        self.num_attention_heads = config.num_attention_heads
        self.hidden_size = config.hidden_size
        self.head_size = self.hidden_size // self.num_attention_heads
        self.rotary_ndims = int(self.head_size * config.rotary_pct)
        self.rotary_emb = RotaryEmbedding(
            self.rotary_ndims,
            config.max_position_embeddings,
            base=config.rotary_emb_base
        )
        self.query_key_value = nn.Linear(config.hidden_size, 3 * config.hidden_size)
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.causal_mask_cached = None

    def causal_mask(self, x, seq_len):
        if self.causal_mask_cached is None or seq_len > self.causal_mask_cached.shape[2]:
            cache_size = max(self.max_position_embeddings, seq_len)
            self.causal_mask_cached = torch.ones(
                cache_size,
                cache_size,
                dtype=torch.bool
            ).tril().view(1, 1, cache_size, cache_size)
        return self.causal_mask_cached[:, :, :seq_len, :seq_len].to(x.device)

    def forward(
        self,
        hidden_states,
        attention_mask,
        head_mask=None,
        layer_past=None,
        use_cache=False,
        output_attentions=False
    ):
        has_layer_past = layer_past is not None

        # Compute QKV
        # Attention heads [batch, seq_len, hidden_size]
        #   --> [batch, seq_len, (np * 3 * head_size)]
        qkv = self.query_key_value(hidden_states)

        # [batch, seq_len, (num_heads * 3 * head_size)]
        #   --> [batch, seq_len, num_heads, 3 * head_size]
        new_qkv_shape = qkv.size()[:-1] + (self.num_attention_heads, 3 * self.head_size)
        qkv = qkv.view(*new_qkv_shape)

        # [batch, seq_len, num_attention_heads, 3 * head_size] --> 3 [batch, num_attention_heads, seq_len, head_size]
        query = qkv[..., : self.head_size].permute(0, 2, 1, 3)
        key = qkv[..., self.head_size : 2 * self.head_size].permute(0, 2, 1, 3)
        value = qkv[..., 2 * self.head_size :].permute(0, 2, 1, 3)

        # Compute rotary embeddings on rotary_ndims
        # query_rot = query[..., : self.rotary_ndims]
        # query_pass = query[..., self.rotary_ndims :]
        # key_rot = key[..., : self.rotary_ndims]
        # key_pass = key[..., self.rotary_ndims :]

        # Compute token offset for rotary embeddings (when decoding)
        seq_len = key.shape[-2]
        offset = 0
        if has_layer_past:
            offset = layer_past[0].shape[-2]
            seq_len += offset
        cos, sin = self.rotary_emb(value, seq_len=seq_len)

        query, key = apply_rotary_pos_emb(query, key, cos, sin, offset=offset)
        # query, key = apply_rotary_pos_emb(query_rot, key_rot, cos, sin, offset=offset)
        # query = torch.cat((query, query_pass), dim=-1)
        # key = torch.cat((key, key_pass), dim=-1)

        # Cache QKV values
        if has_layer_past:
            past_key = layer_past[0]
            past_value = layer_past[1]
            key = torch.cat((past_key, key), dim=-2)
            value = torch.cat((past_value, value), dim=-2)
        present = (key, value,) if use_cache else None

        query = query.type_as(hidden_states)
        key = key.type_as(hidden_states)
        value = value.type_as(hidden_states)

        if output_attentions:
            # Use custom attention method to get attn_weights
            attn_output, attn_weights = self._attn(
                query, key, value,
                attention_mask=attention_mask,
                head_mask=head_mask
            )
        else:
            if layer_past is not None and attention_mask is None:
                # Must calculate attention_mask, or scaled_dot_product_attention will wrong
                batch_size = query.size(0)
                attention_mask = torch.ones(batch_size, seq_len, dtype=torch.bool)[:, None, None, :]

            if attention_mask is not None:
                attn_mask = attention_mask.transpose(2, 3) * attention_mask
                query_length = query.size(-2)
                key_length = key.size(-2)
                if query_length > 1:
                    causal_mask = self.causal_mask(query, seq_len)
                    causal_mask = causal_mask[:, :, -query_length:, :]
                    attn_mask = (attn_mask[:, :, -query_length:, :] * causal_mask).to(torch.bool)
                else:
                    attn_mask = attn_mask[:, :, -query_length:, :].to(torch.bool)

                attn_output = F.scaled_dot_product_attention(
                    query,
                    key,
                    value,
                    attn_mask=attn_mask,
                    is_causal=False
                )
            else:
                attn_output = F.scaled_dot_product_attention(
                    query,
                    key,
                    value,
                    attn_mask=None,
                    is_causal=True
                )
            attn_weights = None

        # Reshape outputs
        # attn_output == [bs, num_attention_heads, seq_len, attn_head_size]
        attn_output = self._merge_heads(attn_output, self.num_attention_heads, self.head_size)
        # tensor [bs, seq_len, num_attention_heads * attn_head_size]
        attn_output = self.dense(attn_output)

        outputs = (attn_output, present)
        if output_attentions:
            outputs += (attn_weights,)

        return outputs

    @classmethod
    def _calculate_attn_output_loss(self, attn_output):
        bs, num_attention_heads, seq_len, attn_head_size = attn_output.size()
        attn_output_out = attn_output.view(bs, num_attention_heads, -1)
        attn_output_out_norm = attn_output_out / torch.max(
            attn_output_out.norm(dim=2, keepdim=True),
            1e-8 * torch.ones_like(attn_output_out)
        )
        sim = torch.bmm(attn_output_out_norm, attn_output_out_norm.permute(0, 2, 1))
        attn_output_loss = sim.sum() / sim.numel()
        return attn_output_loss

    @classmethod
    def _split_heads(cls, tensor, num_attention_heads, attn_head_size):
        """
        Splits hidden dim into attn_head_size and num_attention_heads
        """
        # tensor: [bs, seq_len, hidden_size]
        new_shape = tensor.size()[:-1] + (num_attention_heads, attn_head_size)
        # -> [bs, seq_len, num_attention_heads, attn_head_size]
        tensor = tensor.view(new_shape)
        # -> [bs, num_attention_heads, seq_len, attn_head_size]
        tensor = tensor.permute(0, 2, 1, 3)
        return tensor

    @classmethod
    def _merge_heads(cls, tensor, num_attention_heads, attn_head_size):
        """
        Merges attn_head_size dim and num_attn_heads dim into hidden dim
        """
        # tensor [bs, num_attention_heads, seq_len, attn_head_size]
        tensor = tensor.permute(0, 2, 1, 3).contiguous()
        # -> [bs, seq_len, num_attention_heads, attn_head_size]
        tensor = tensor.view(tensor.size(0), tensor.size(1), num_attention_heads * attn_head_size)
        # -> [bs, seq_len, hidden_size]
        return tensor

    def create_upper_triangular_matrix(self, q, k):
        size = max(q, k)
        # 创建一个单位矩阵
        identity = torch.eye(size)
        # 创建一个矩阵，其中每个元素都是它的行索引
        row_indices = torch.arange(size).view(-1, 1).expand(size, size)
        # 创建一个矩阵，其中每个元素都是它的列索引
        col_indices = torch.arange(size).view(1, -1).expand(size, size)
        # 比较行和列索引，如果行索引小于列索引，则0，否则1
        upper_triangular_matrix = torch.where(row_indices < col_indices, 0, 1)
        return upper_triangular_matrix[-q:, -k:].to(torch.bool)
    
    def _attn(self, query, key, value, attention_mask=None, head_mask=None):
        # q, k, v: [bs, num_attention_heads, seq_len, attn_head_size]
        # compute causal mask from causal mask buffer
        batch_size, num_attention_heads, query_length, attn_head_size = query.size()
        key_length = key.size(-2)

        # 避免使用tril
        # causal_mask = torch.ones(
        #     query_length, key_length,
        #     dtype=torch.bool,
        #     device=query.device
        # ).tril(
        #     diagonal=key_length - query_length
        # ).view(1, 1, query_length, key_length)
        causal_mask = self.create_upper_triangular_matrix(
            query_length, key_length
        ).view(1, 1, query_length, key_length).to(query.device)

        query = query.view(batch_size * num_attention_heads, query_length, attn_head_size)
        key = key.view(batch_size * num_attention_heads, key_length, attn_head_size)
        attn_scores = torch.zeros(
            batch_size * num_attention_heads,
            query_length,
            key_length,
            dtype=query.dtype,
            device=key.device,
        )
        norm_factor = self.head_size ** 0.5
        attn_scores = torch.baddbmm(
            attn_scores,
            query,
            key.transpose(1, 2),
            beta=1.0,
            alpha=(torch.tensor(1.0, dtype=query.dtype, device=query.device) / norm_factor),
        )
        attn_scores = attn_scores.view(batch_size, num_attention_heads, query_length, key_length)

        mask_value = torch.finfo(attn_scores.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_scores.dtype).to(attn_scores.device)
        attn_scores = torch.where(causal_mask, attn_scores, mask_value)

        if attention_mask is not None:
            # Apply the attention mask
            attn_scores = attn_scores + attention_mask

        attn_weights = nn.functional.softmax(attn_scores.float(), dim=-1).type_as(value)

        # 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


class RotaryEmbedding(torch.nn.Module):
    def __init__(self, dim, max_position_embeddings, base=10000, device=None):
        super().__init__()
        inv_freq = 1.0 / (base ** (torch.arange(0, dim, 2).float().to(device) / dim))
        self.register_buffer("inv_freq", inv_freq)

        # Build here to make `torch.jit.trace` work.
        self.max_seq_len_cached = max_position_embeddings
        t = torch.arange(self.max_seq_len_cached, device=self.inv_freq.device, dtype=self.inv_freq.dtype)
        freqs = torch.einsum("i,j->ij", t, self.inv_freq)
        # Different from paper, but it uses a different permutation in order to obtain the same calculation
        emb = torch.cat((freqs, freqs), dim=-1)
        self.cos_cached = emb.cos()[None, None, :, :]
        self.sin_cached = emb.sin()[None, None, :, :]

    def forward(self, x, seq_len=None):
        # x: [bs, num_attention_heads, seq_len, head_size]
        # This `if` block is unlikely to be run after we build sin/cos in `__init__`. Keep the logic here just in case.
        if seq_len > self.max_seq_len_cached:
            self.max_seq_len_cached = seq_len
            t = torch.arange(self.max_seq_len_cached, device=x.device, dtype=self.inv_freq.dtype)
            freqs = torch.einsum("i,j->ij", t, self.inv_freq)
            # Different from paper, but it uses a different permutation in order to obtain the same calculation
            emb = torch.cat((freqs, freqs), dim=-1).to(x.device)
            self.cos_cached = emb.cos()[None, None, :, :]
            self.sin_cached = emb.sin()[None, None, :, :]
        return self.cos_cached[:seq_len, ...].to(x.device), self.sin_cached[:seq_len, ...].to(x.device)


def rotate_half(x):
    """Rotates half the hidden dims of the input."""
    x1 = x[..., : x.shape[-1] // 2]
    x2 = x[..., x.shape[-1] // 2 :]
    return torch.cat((-x2, x1), dim=-1)


def apply_rotary_pos_emb(q, k, cos, sin, offset: int = 0):
    cos = cos[..., offset : q.shape[-2] + offset, :]
    sin = sin[..., offset : q.shape[-2] + offset, :]
    q_embed = (q * cos) + (rotate_half(q) * sin)
    k_embed = (k * cos) + (rotate_half(k) * sin)
    return q_embed, k_embed


class GatedLinear(nn.Linear):
    pass


class GPTJiangMLP(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.dense_h_to_4h = nn.Linear(config.hidden_size, config.intermediate_size, bias=config.mlp_bias)
        self.dense_4h_to_h = nn.Linear(config.intermediate_size, config.hidden_size, bias=config.mlp_bias)
        self.gated = config.gated
        if config.gated:
            self.dense_h_to_4h_gate = GatedLinear(config.hidden_size, config.intermediate_size, bias=config.mlp_bias)
        self.act = ACT2FN[config.hidden_act]

    def forward(self, hidden_states):
        
        if self.gated:
            # pseudocode:
            # g is activation function, W and V are weights, * is element-wised product
            # x = g(Wx) * Vx
            hidden_states = self.act(self.dense_h_to_4h(hidden_states)) * self.dense_h_to_4h_gate(hidden_states)
        else:
            # pseudocode:
            # x = g(Wx)
            hidden_states = self.act(self.dense_h_to_4h(hidden_states))
        hidden_states = self.dense_4h_to_h(hidden_states)
        return hidden_states


class GPTJiangLayer(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.use_parallel_residual = config.use_parallel_residual
        self.input_layernorm = RMSNorm(config.hidden_size, eps=config.layer_norm_eps)
        self.post_attention_layernorm = RMSNorm(config.hidden_size, eps=config.layer_norm_eps)
        self.attention = GPTJiangAttention(config)
        self.mlp = GPTJiangMLP(config)

    def forward(
        self,
        hidden_states,
        attention_mask=None,
        head_mask=None,
        use_cache=False,
        layer_past=None,
        output_attentions=False,
    ):
        attention_layer_outputs = self.attention(
            self.input_layernorm(hidden_states),
            attention_mask=attention_mask,
            layer_past=layer_past,
            head_mask=head_mask,
            use_cache=use_cache,
            output_attentions=output_attentions,
        )
        attn_output = attention_layer_outputs[0]  # output_attn: attn_output, present, (attn_weights), (attentions_output_loss)
        outputs = attention_layer_outputs[1:]

        # Default True in multiple models, faster
        if self.use_parallel_residual:
            # pseudocode:
            # x = x + attn(ln1(x)) + mlp(ln2(x))
            mlp_output = self.mlp(self.post_attention_layernorm(hidden_states))
            hidden_states = mlp_output + attn_output + hidden_states
        else:
            # pseudocode:
            # x = x + attn(ln1(x))
            # x = x + mlp(ln2(x))
            attn_output = attn_output + hidden_states
            mlp_output = self.mlp(self.post_attention_layernorm(attn_output))
            hidden_states = mlp_output + attn_output

        if use_cache:
            outputs = (hidden_states,) + outputs  # hidden_states, present, (attn_weights), (attentions_output_loss)
        else:
            outputs = (hidden_states,) + outputs[1:]  # hidden_states, (attn_weights), (attentions_output_loss)

        return outputs


GPT_JIANG_START_DOCSTRING = r"""
    This model is a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) sub-class. Use
    it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and
    behavior.

    Parameters:
        config ([`~GPTJiangConfig`]): 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.
"""

GPT_JIANG_INPUTS_DOCSTRING = r"""
    Args:
        input_ids (`torch.LongTensor` of shape `({0})`):
            Indices of input sequence tokens in the vocabulary.

            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
            [`PreTrainedTokenizer.__call__`] for details.

            [What are input IDs?](../glossary#input-ids)
        attention_mask (`torch.FloatTensor` of shape `({0})`, *optional*):
            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:

            - 1 for tokens that are **not masked**,
            - 0 for tokens that are **masked**.

            [What are attention masks?](../glossary#attention-mask)
        head_mask (`torch.FloatTensor` of shape `(num_heads,)` or `(num_layers, num_heads)`, *optional*):
            Mask to nullify selected heads of the self-attention modules. Mask values selected in `[0, 1]`:

            - 1 indicates the head is **not masked**,
            - 0 indicates the head is **masked**.

        inputs_embeds (`torch.FloatTensor` of shape `({0}, hidden_size)`, *optional*):
            Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
            is useful if you want more control over how to convert *input_ids* indices into associated vectors than the
            model's internal embedding lookup matrix.
        output_attentions (`bool`, *optional*):
            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
            tensors for more detail.
        output_hidden_states (`bool`, *optional*):
            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
            more detail.
        return_dict (`bool`, *optional*):
            Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple.
"""


@add_start_docstrings(
    "The bare GPTJiang Model transformer outputting raw hidden-states without any specific head on top.",
    GPT_JIANG_START_DOCSTRING,
)
class GPTJiangModel(GPTJiangPreTrainedModel):
    def __init__(self, config):
        super().__init__(config)
        self.config = config

        self.embed_in = nn.Embedding(config.vocab_size, config.hidden_size)
        self.layers = nn.ModuleList([GPTJiangLayer(config) for _ in range(config.num_hidden_layers)])
        self.final_layer_norm = RMSNorm(config.hidden_size, eps=config.layer_norm_eps)

        self.gradient_checkpointing = False

        # Initialize weights and apply final processing
        self.post_init()

    def get_input_embeddings(self):
        return self.embed_in

    def set_input_embeddings(self, value):
        self.embed_in = value

    @add_start_docstrings_to_model_forward(GPT_JIANG_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
    @add_code_sample_docstrings(
        output_type=BaseModelOutputWithPast,
        config_class=_CONFIG_FOR_DOC,
    )
    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        attention_mask: Optional[torch.FloatTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        past_key_values: Optional[Tuple[Tuple[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, BaseModelOutputWithPast]:
        r"""
        past_key_values (`tuple(tuple(torch.FloatTensor))` of length `config.n_layers` with each tuple having 4 tensors of shape `(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`):
            Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding.
            If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that
            don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all
            `decoder_input_ids` of shape `(batch_size, sequence_length)`.
        use_cache (`bool`, *optional*):
            If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see `past_key_values`).
        """
        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
        output_hidden_states = (
            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
        )
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
        use_cache = use_cache if use_cache is not None else self.config.use_cache

        if input_ids is not None and inputs_embeds is not None:
            raise ValueError("You cannot specify both input_ids and inputs_embeds at the same time")
        elif input_ids is not None:
            input_shape = input_ids.size()
        elif inputs_embeds is not None:
            input_shape = inputs_embeds.size()[:-1]
        else:
            raise ValueError("You have to specify either input_ids or inputs_embeds")

        batch_size, seq_length = input_shape

        if past_key_values is None:
            past_key_values = tuple([None] * self.config.num_hidden_layers)

        # Attention mask.
        if attention_mask is not None:
            assert batch_size > 0, "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

        # Prepare head mask if needed
        # 1.0 in head_mask indicate we keep the head
        # attention_probs has shape bsz x n_heads x N x N
        # input head_mask has shape [num_heads] or [num_hidden_layers x num_heads]
        # and head_mask is converted to shape [num_hidden_layers x batch x num_heads x seq_length x seq_length]
        head_mask = self.get_head_mask(head_mask, self.config.num_hidden_layers)

        if inputs_embeds is None:
            inputs_embeds = self.embed_in(input_ids)

        hidden_states = inputs_embeds

        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

        presents = () if use_cache else None
        all_attentions = () if output_attentions else None
        all_hidden_states = () if output_hidden_states else None
        for i, (layer, layer_past) in enumerate(zip(self.layers, past_key_values)):

            if output_hidden_states:
                all_hidden_states = all_hidden_states + (hidden_states,)

            if self.gradient_checkpointing and self.training:

                def create_custom_forward(module):
                    def custom_forward(*inputs):
                        # None for layer_past
                        return module(*inputs, use_cache, None, output_attentions)

                    return custom_forward

                outputs = torch.utils.checkpoint.checkpoint(
                    create_custom_forward(layer),
                    hidden_states,
                    attention_mask,
                    head_mask[i],
                )
            else:
                outputs = layer(
                    hidden_states,
                    attention_mask=attention_mask,
                    head_mask=head_mask[i],
                    layer_past=layer_past,
                    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_attentions = all_attentions + (outputs[2 if use_cache else 1],)

        hidden_states = self.final_layer_norm(hidden_states)
        # 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_attentions] if v is not None)

        ret = BaseModelOutputWithPast(
            last_hidden_state=hidden_states,
            past_key_values=presents,
            hidden_states=all_hidden_states,
            attentions=all_attentions,
        )
        return ret


@add_start_docstrings(
    """GPTJiang Model with a `language modeling` head on top for CLM fine-tuning.""", GPT_JIANG_START_DOCSTRING
)
class GPTJiangForCausalLM(GPTJiangPreTrainedModel):
    _keys_to_ignore_on_load_missing = [r"position_ids", r"predictions.decoder.bias"]

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

        self.gpt_kdf = GPTJiangModel(config)
        self.embed_out = nn.Linear(config.hidden_size, config.vocab_size, bias=False)

        # Initialize weights and apply final processing
        self.post_init()

    def get_output_embeddings(self):
        return self.embed_out

    def set_output_embeddings(self, new_embeddings):
        self.embed_out = new_embeddings

    @add_start_docstrings_to_model_forward(GPT_JIANG_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
    @replace_return_docstrings(output_type=CausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC)
    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        attention_mask: Optional[torch.FloatTensor] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        past_key_values: Optional[Tuple[Tuple[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, CausalLMOutputWithPast]:
        r"""
        past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
            Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of shape
            `(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional tensors of shape
            `(batch_size, num_heads, encoder_sequence_length, embed_size_per_head)`. The two additional tensors are
            only required when the model is used as a decoder in a Sequence to Sequence model.

            Contains pre-computed hidden-states (key and values in the self-attention blocks that can be used (see
            `past_key_values` input) to speed up sequential decoding.

            If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that
            don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all
            `decoder_input_ids` of shape `(batch_size, sequence_length)`.
        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Labels for computing the left-to-right language modeling loss (next word prediction). Indices should be in
            `[-100, 0, ..., config.vocab_size]` (see `input_ids` docstring) Tokens with indices set to `-100` are
            ignored (masked), the loss is only computed for the tokens with labels n `[0, ..., config.vocab_size]`.
        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`).

        Returns:

        Example:

        ```python
        >>> from transformers import AutoTokenizer, GPTJiangForCausalLM, GPTJiangConfig
        >>> import torch

        >>> tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neox-20b")
        >>> config = GPTJiangConfig.from_pretrained("EleutherAI/gpt-neox-20b")
        >>> config.is_decoder = True
        >>> model = GPTJiangForCausalLM.from_pretrained("EleutherAI/gpt-neox-20b", config=config)

        >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
        >>> outputs = model(**inputs)

        >>> prediction_logits = outputs.logits
        ```"""
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict

        outputs = self.gpt_kdf(
            input_ids,
            attention_mask=attention_mask,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            past_key_values=past_key_values,
            use_cache=use_cache,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
        )

        hidden_states = outputs[0]
        lm_logits = self.embed_out(hidden_states)

        lm_loss = None
        attn_output_loss = None
        if labels is not None:
            # we are doing next-token prediction; shift prediction scores and input ids by one
            shift_logits = lm_logits[:, :-1, :].contiguous()
            labels = labels[:, 1:].contiguous()
            loss_fct = CrossEntropyLoss()
            lm_loss = loss_fct(shift_logits.view(-1, shift_logits.size(-1)), labels.view(-1))

        if not return_dict:
            output = (lm_logits,) + outputs[1:]
            return ((lm_loss,) + output) if lm_loss is not None else output

        ret = CausalLMOutputWithPast(
            loss=lm_loss,
            logits=lm_logits,
            past_key_values=outputs.past_key_values,
            hidden_states=outputs.hidden_states,
            attentions=outputs.attentions,
        )
        return ret

    def prepare_inputs_for_generation(self, input_ids, past_key_values=None, attention_mask=None, **model_kwargs):
        input_shape = input_ids.shape

        # if model is used as a decoder in encoder-decoder model, the decoder attention mask is created on the fly
        if attention_mask is None:
            attention_mask = input_ids.new_ones(input_shape)

        # cut decoder_input_ids if past is used
        if past_key_values and past_key_values[0] is not None:
            input_ids = input_ids[:, -1:]

        return {
            "input_ids": input_ids,
            "attention_mask": attention_mask,
            "past_key_values": past_key_values,
        }

    def _reorder_cache(self, past_key_values, beam_idx):
        reordered_past = ()
        for layer_past in past_key_values:
            reordered_past += (
                tuple(past_state.index_select(0, beam_idx) for past_state in layer_past[:2]) + layer_past[2:],
            )
        return reordered_past