modeling_flmaudio.py

# coding=utf-8
"""PyTorch FLM-Audio model, based on LLAMA implementation."""

import math
import warnings
from typing import List, Optional, Tuple, Union
from dataclasses import dataclass

import torch
import torch.nn as nn
import torch.nn.functional as F

from transformers.activations import ACT2FN
from transformers.modeling_rope_utils import ROPE_INIT_FUNCTIONS, dynamic_rope_update
from transformers.cache_utils import Cache, DynamicCache, StaticCache
from transformers.modeling_attn_mask_utils import AttentionMaskConverter
from transformers.modeling_outputs import (
    ModelOutput,
    BaseModelOutputWithPast,
    CausalLMOutputWithPast,
)
from transformers.modeling_utils import PreTrainedModel
from transformers.pytorch_utils import ALL_LAYERNORM_LAYERS
from transformers.utils import (
    add_start_docstrings,
    add_start_docstrings_to_model_forward,
    is_flash_attn_2_available,
    is_flash_attn_greater_or_equal_2_10,
    logging,
    replace_return_docstrings,
)
from .configuration_flmaudio import FLMAudioConfig
from .depth_gpt import DepthGPT, DepthGPTConfig

if is_flash_attn_2_available():
    from flash_attn import flash_attn_func, flash_attn_varlen_func
    from flash_attn.bert_padding import index_first_axis, pad_input, unpad_input  # noqa


logger = logging.get_logger(__name__)

_CONFIG_FOR_DOC = "FLMAudioConfig"


def _get_unpad_data(attention_mask):
    seqlens_in_batch = attention_mask.sum(dim=-1, dtype=torch.int32)
    indices = torch.nonzero(attention_mask.flatten(), as_tuple=False).flatten()
    max_seqlen_in_batch = seqlens_in_batch.max().item()
    cu_seqlens = F.pad(torch.cumsum(seqlens_in_batch, dim=0, dtype=torch.int32), (1, 0))
    return (
        indices,
        cu_seqlens,
        max_seqlen_in_batch,
    )


class FLMAudioRMSNorm(nn.Module):
    def __init__(self, hidden_size, eps=1e-6):
        """
        FLMAudioRMSNorm is equivalent to T5LayerNorm
        """
        super().__init__()
        self.weight = nn.Parameter(torch.ones(hidden_size))
        self.variance_epsilon = eps

    def forward(self, hidden_states):
        input_dtype = hidden_states.dtype
        hidden_states = hidden_states.to(torch.float32)
        variance = hidden_states.pow(2).mean(-1, keepdim=True)
        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
        return self.weight * hidden_states.to(input_dtype)


ALL_LAYERNORM_LAYERS.append(FLMAudioRMSNorm)

class FLMAudioRotaryEmbedding(nn.Module):
    def __init__(self, config, device=None):
        super().__init__()
        # BC: "rope_type" was originally "type"
        if hasattr(config, "rope_scaling") and config.rope_scaling is not None:
            self.rope_type = config.rope_scaling.get("rope_type", config.rope_scaling.get("type"))
        else:
            self.rope_type = "default"
        self.max_seq_len_cached = config.max_position_embeddings
        self.original_max_seq_len = config.max_position_embeddings

        self.config = config
        self.rope_init_fn = ROPE_INIT_FUNCTIONS[self.rope_type]

        inv_freq, self.attention_scaling = self.rope_init_fn(self.config, device)
        self.register_buffer("inv_freq", inv_freq, persistent=False)
        self.original_inv_freq = self.inv_freq

    @torch.no_grad()
    @dynamic_rope_update  # power user: used with advanced RoPE types (e.g. dynamic rope)
    def forward(self, x, position_ids):
        inv_freq_expanded = self.inv_freq[None, None, :, None].float().expand(3, position_ids.shape[1], -1, 1)
        position_ids_expanded = position_ids[:, :, None, :].float()  # shape (3, bs, 1, positions)

        device_type = x.device.type if isinstance(x.device.type, str) and x.device.type != "mps" else "cpu"
        with torch.autocast(device_type=device_type, enabled=False):  # Force float32
            freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(2, 3)
            emb = torch.cat((freqs, freqs), dim=-1)
            cos = emb.cos() * self.attention_scaling
            sin = emb.sin() * self.attention_scaling

        return cos.to(dtype=x.dtype), sin.to(dtype=x.dtype)


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_multimodal_rotary_pos_emb(q, k, cos, sin, mrope_section, unsqueeze_dim=1):
    mrope_section = mrope_section * 2
    cos = torch.cat([m[i % 3] for i, m in enumerate(cos.split(mrope_section, dim=-1))], dim=-1).unsqueeze(
        unsqueeze_dim
    )
    sin = torch.cat([m[i % 3] for i, m in enumerate(sin.split(mrope_section, dim=-1))], dim=-1).unsqueeze(
        unsqueeze_dim
    )

    q_embed = (q * cos) + (rotate_half(q) * sin)
    k_embed = (k * cos) + (rotate_half(k) * sin)
    return q_embed, k_embed


class FLMAudioMLP(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.config = config
        self.hidden_size = config.hidden_size
        self.intermediate_size = config.intermediate_size
        self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
        self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
        self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
        self.act_fn = ACT2FN[config.hidden_act]

    def forward(self, x):
        if self.config.pretraining_tp > 1:
            slice = self.intermediate_size // self.config.pretraining_tp
            gate_proj_slices = self.gate_proj.weight.split(slice, dim=0)
            up_proj_slices = self.up_proj.weight.split(slice, dim=0)
            down_proj_slices = self.down_proj.weight.split(slice, dim=1)

            gate_proj = torch.cat(
                [
                    F.linear(x, gate_proj_slices[i])
                    for i in range(self.config.pretraining_tp)
                ],
                dim=-1,
            )
            up_proj = torch.cat(
                [
                    F.linear(x, up_proj_slices[i])
                    for i in range(self.config.pretraining_tp)
                ],
                dim=-1,
            )

            intermediate_states = (self.act_fn(gate_proj) * up_proj).split(slice, dim=2)
            down_proj = [
                F.linear(intermediate_states[i], down_proj_slices[i])
                for i in range(self.config.pretraining_tp)
            ]
            down_proj = sum(down_proj)
        else:
            down_proj = self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))

        return down_proj


def repeat_kv(hidden_states: torch.Tensor, n_rep: int) -> torch.Tensor:
    """
    This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
    num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
    """
    batch, num_key_value_heads, slen, head_dim = hidden_states.shape
    if n_rep == 1:
        return hidden_states
    hidden_states = hidden_states[:, :, None, :, :].expand(
        batch, num_key_value_heads, n_rep, slen, head_dim
    )
    return hidden_states.reshape(batch, num_key_value_heads * n_rep, slen, head_dim)


class FLMAudioAttention(nn.Module):
    """Multi-headed attention from 'Attention Is All You Need' paper"""

    def __init__(self, config: FLMAudioConfig, layer_idx: Optional[int] = None):
        super().__init__()
        self.config = config
        self.layer_idx = layer_idx
        if layer_idx is None:
            logger.warning_once(
                f"Instantiating {self.__class__.__name__} without passing a `layer_idx` is not recommended and will "
                "lead to errors during the forward call if caching is used. Please make sure to provide a `layer_idx` "
                "when creating this class."
            )

        self.attention_dropout = config.attention_dropout
        self.hidden_size = config.hidden_size
        self.num_heads = config.num_attention_heads
        self.head_dim = self.hidden_size // self.num_heads
        self.num_key_value_heads = config.num_key_value_heads
        self.num_key_value_groups = self.num_heads // self.num_key_value_heads
        self.is_causal = True
        self.rope_scaling = config.rope_scaling

        if (self.head_dim * self.num_heads) != self.hidden_size:
            raise ValueError(
                f"hidden_size must be divisible by num_heads (got `hidden_size`: {self.hidden_size}"
                f" and `num_heads`: {self.num_heads})."
            )

        self.q_proj = nn.Linear(
            self.hidden_size, self.num_heads * self.head_dim, bias=config.attention_bias
        )
        self.k_proj = nn.Linear(
            self.hidden_size,
            self.num_key_value_heads * self.head_dim,
            bias=config.attention_bias,
        )
        self.v_proj = nn.Linear(
            self.hidden_size,
            self.num_key_value_heads * self.head_dim,
            bias=config.attention_bias,
        )
        self.o_proj = nn.Linear(
            self.hidden_size, self.hidden_size, bias=config.attention_bias and not config.disable_att_o_bias
        )


    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_value: Optional[Cache] = None,
        output_attentions: bool = False,
        use_cache: bool = False,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,  # necessary, but kept here for BC
        **kwargs,
    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
        bsz, q_len, _ = hidden_states.size()

        if self.config.pretraining_tp > 1:
            key_value_slicing = (
                self.num_key_value_heads * self.head_dim
            ) // self.config.pretraining_tp
            query_slices = self.q_proj.weight.split(
                (self.num_heads * self.head_dim) // self.config.pretraining_tp, dim=0
            )
            key_slices = self.k_proj.weight.split(key_value_slicing, dim=0)
            value_slices = self.v_proj.weight.split(key_value_slicing, dim=0)

            query_states = [
                F.linear(hidden_states, query_slices[i])
                for i in range(self.config.pretraining_tp)
            ]
            query_states = torch.cat(query_states, dim=-1)

            key_states = [
                F.linear(hidden_states, key_slices[i])
                for i in range(self.config.pretraining_tp)
            ]
            key_states = torch.cat(key_states, dim=-1)

            value_states = [
                F.linear(hidden_states, value_slices[i])
                for i in range(self.config.pretraining_tp)
            ]
            value_states = torch.cat(value_states, dim=-1)

        else:
            query_states = self.q_proj(hidden_states)
            key_states = self.k_proj(hidden_states)
            value_states = self.v_proj(hidden_states)

        query_states = query_states.view(
            bsz, q_len, self.num_heads, self.head_dim
        ).transpose(1, 2)
        key_states = key_states.view(
            bsz, q_len, self.num_key_value_heads, self.head_dim
        ).transpose(1, 2)
        value_states = value_states.view(
            bsz, q_len, self.num_key_value_heads, self.head_dim
        ).transpose(1, 2)

        cos, sin = position_embeddings

        query_states, key_states = apply_multimodal_rotary_pos_emb(
            query_states, key_states, cos, sin, self.rope_scaling["mrope_section"]
        )

        if past_key_value is not None:
            # sin and cos are specific to RoPE models; cache_position needed for the static cache
            cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
            key_states, value_states = past_key_value.update(
                key_states, value_states, self.layer_idx, cache_kwargs
            )

        key_states = repeat_kv(key_states, self.num_key_value_groups)
        value_states = repeat_kv(value_states, self.num_key_value_groups)

        attn_weights = torch.matmul(
            query_states, key_states.transpose(2, 3)
        ) / math.sqrt(self.head_dim)

        if attention_mask is not None:  # no matter the length, we just slice it
            causal_mask = attention_mask[:, :, :, : key_states.shape[-2]]
            attn_weights = attn_weights + causal_mask

        # upcast attention to fp32
        attn_weights = nn.functional.softmax(
            attn_weights, dim=-1, dtype=torch.float32
        ).to(query_states.dtype)
        attn_weights = nn.functional.dropout(
            attn_weights, p=self.attention_dropout, training=self.training
        )
        attn_output = torch.matmul(attn_weights, value_states)

        if attn_output.size() != (bsz, self.num_heads, q_len, self.head_dim):
            raise ValueError(
                f"`attn_output` should be of size {(bsz, self.num_heads, q_len, self.head_dim)}, but is"
                f" {attn_output.size()}"
            )

        attn_output = attn_output.transpose(1, 2).contiguous()

        attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)

        if self.config.pretraining_tp > 1:
            attn_output = attn_output.split(
                self.hidden_size // self.config.pretraining_tp, dim=2
            )
            o_proj_slices = self.o_proj.weight.split(
                self.hidden_size // self.config.pretraining_tp, dim=1
            )
            attn_output = sum(
                [
                    F.linear(attn_output[i], o_proj_slices[i])
                    for i in range(self.config.pretraining_tp)
                ]
            )
        else:
            attn_output = self.o_proj(attn_output)

        if not output_attentions:
            attn_weights = None

        return attn_output, attn_weights, past_key_value


class FLMAudioFlashAttention2(FLMAudioAttention):
    """
    FLM-Audio flash attention module. This module inherits from `FLMAudioAttention` as the weights of the module stays
    untouched. The only required change would be on the forward pass where it needs to correctly call the public API of
    flash attention and deal with padding tokens in case the input contains any of them.
    """

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

        # TODO: Should be removed once Flash Attention for RoCm is bumped to 2.1.
        # flash_attn<2.1 generates top-left aligned causal mask, while what is needed here is bottom-right alignement, that was made default for flash_attn>=2.1. This attribute is used to handle this difference. Reference: https://github.com/Dao-AILab/flash-attention/releases/tag/v2.1.0.
        # Beware that with flash_attn<2.1, using q_seqlen != k_seqlen (except for the case q_seqlen == 1) produces a wrong mask (top-left).
        self._flash_attn_uses_top_left_mask = not is_flash_attn_greater_or_equal_2_10()

    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.LongTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_value: Optional[Cache] = None,
        output_attentions: bool = False,
        use_cache: bool = False,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,  # necessary, but kept here for BC
        **kwargs,
    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
        output_attentions = False

        bsz, q_len, _ = hidden_states.size()

        query_states = self.q_proj(hidden_states)
        key_states = self.k_proj(hidden_states)
        value_states = self.v_proj(hidden_states)

        # Flash attention requires the input to have the shape
        # batch_size x seq_length x head_dim x hidden_dim
        # therefore we just need to keep the original shape
        query_states = query_states.view(
            bsz, q_len, self.num_heads, self.head_dim
        ).transpose(1, 2)
        key_states = key_states.view(
            bsz, q_len, self.num_key_value_heads, self.head_dim
        ).transpose(1, 2)
        value_states = value_states.view(
            bsz, q_len, self.num_key_value_heads, self.head_dim
        ).transpose(1, 2)

        # cos, sin = self.rotary_emb(value_states, position_ids)
        cos, sin = position_embeddings
        # query_states, key_states = apply_rotary_pos_emb(
        #     query_states, key_states, cos, sin
        # )
        query_states, key_states = apply_multimodal_rotary_pos_emb(
            query_states, key_states, cos, sin, self.rope_scaling["mrope_section"]
        )

        past_key_value = getattr(self, "past_key_value", past_key_value)

        if past_key_value is not None:
            # sin and cos are specific to RoPE models; cache_position needed for the static cache
            cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
            key_states, value_states = past_key_value.update(
                key_states, value_states, self.layer_idx, cache_kwargs
            )

        # TODO: These transpose are quite inefficient but Flash Attention requires the layout [batch_size, sequence_length, num_heads, head_dim]. We would need to refactor the KV cache
        # to be able to avoid many of these transpose/reshape/view.
        query_states = query_states.transpose(1, 2)
        key_states = key_states.transpose(1, 2)
        value_states = value_states.transpose(1, 2)

        dropout_rate = self.attention_dropout if self.training else 0.0

        # In PEFT, usually we cast the layer norms in float32 for training stability reasons
        # therefore the input hidden states gets silently casted in float32. Hence, we need
        # cast them back in the correct dtype just to be sure everything works as expected.
        # This might slowdown training & inference so it is recommended to not cast the LayerNorms
        # in fp32. (FLMAudioRMSNorm handles it correctly)

        input_dtype = query_states.dtype
        if input_dtype == torch.float32:
            if torch.is_autocast_enabled():
                target_dtype = torch.get_autocast_gpu_dtype()
            # Handle the case where the model is quantized
            elif hasattr(self.config, "_pre_quantization_dtype"):
                target_dtype = self.config._pre_quantization_dtype
            else:
                target_dtype = self.q_proj.weight.dtype

            logger.warning_once(
                f"The input hidden states seems to be silently casted in float32, this might be related to"
                f" the fact you have upcasted embedding or layer norm layers in float32. We will cast back the input in"
                f" {target_dtype}."
            )

            query_states = query_states.to(target_dtype)
            key_states = key_states.to(target_dtype)
            value_states = value_states.to(target_dtype)

        attn_output = self._flash_attention_forward(
            query_states,
            key_states,
            value_states,
            attention_mask,
            q_len,
            dropout=dropout_rate,
        )

        attn_output = attn_output.reshape(bsz, q_len, self.hidden_size).contiguous()
        attn_output = self.o_proj(attn_output)

        if not output_attentions:
            attn_weights = None

        return attn_output, attn_weights, past_key_value

    def _flash_attention_forward(
        self,
        query_states,
        key_states,
        value_states,
        attention_mask,
        query_length,
        dropout=0.0,
        softmax_scale=None,
    ):
        """
        Calls the forward method of Flash Attention - if the input hidden states contain at least one padding token
        first unpad the input, then computes the attention scores and pad the final attention scores.

        Args:
            query_states (`torch.Tensor`):
                Input query states to be passed to Flash Attention API
            key_states (`torch.Tensor`):
                Input key states to be passed to Flash Attention API
            value_states (`torch.Tensor`):
                Input value states to be passed to Flash Attention API
            attention_mask (`torch.Tensor`):
                The padding mask - corresponds to a tensor of size `(batch_size, seq_len)` where 0 stands for the
                position of padding tokens and 1 for the position of non-padding tokens.
            dropout (`float`):
                Attention dropout
            softmax_scale (`float`, *optional*):
                The scaling of QK^T before applying softmax. Default to 1 / sqrt(head_dim)
        """
        if not self._flash_attn_uses_top_left_mask:
            causal = self.is_causal
        else:
            # TODO: Remove the `query_length != 1` check once Flash Attention for RoCm is bumped to 2.1. For details, please see the comment in FLMAudioFlashAttention2 __init__.
            causal = self.is_causal and query_length != 1

        # Contains at least one padding token in the sequence
        if attention_mask is not None:
            batch_size = query_states.shape[0]
            (
                query_states,
                key_states,
                value_states,
                indices_q,
                cu_seq_lens,
                max_seq_lens,
            ) = self._upad_input(
                query_states, key_states, value_states, attention_mask, query_length
            )

            cu_seqlens_q, cu_seqlens_k = cu_seq_lens
            max_seqlen_in_batch_q, max_seqlen_in_batch_k = max_seq_lens

            attn_output_unpad = flash_attn_varlen_func(
                query_states,
                key_states,
                value_states,
                cu_seqlens_q=cu_seqlens_q,
                cu_seqlens_k=cu_seqlens_k,
                max_seqlen_q=max_seqlen_in_batch_q,
                max_seqlen_k=max_seqlen_in_batch_k,
                dropout_p=dropout,
                softmax_scale=softmax_scale,
                causal=causal,
            )

            attn_output = pad_input(
                attn_output_unpad, indices_q, batch_size, query_length
            )
        else:
            attn_output = flash_attn_func(
                query_states,
                key_states,
                value_states,
                dropout,
                softmax_scale=softmax_scale,
                causal=causal,
            )

        return attn_output

    def _upad_input(
        self, query_layer, key_layer, value_layer, attention_mask, query_length
    ):
        indices_k, cu_seqlens_k, max_seqlen_in_batch_k = _get_unpad_data(attention_mask)
        batch_size, kv_seq_len, num_key_value_heads, head_dim = key_layer.shape

        key_layer = index_first_axis(
            key_layer.reshape(batch_size * kv_seq_len, num_key_value_heads, head_dim),
            indices_k,
        )
        value_layer = index_first_axis(
            value_layer.reshape(batch_size * kv_seq_len, num_key_value_heads, head_dim),
            indices_k,
        )
        if query_length == kv_seq_len:
            query_layer = index_first_axis(
                query_layer.reshape(batch_size * kv_seq_len, self.num_heads, head_dim),
                indices_k,
            )
            cu_seqlens_q = cu_seqlens_k
            max_seqlen_in_batch_q = max_seqlen_in_batch_k
            indices_q = indices_k
        elif query_length == 1:
            max_seqlen_in_batch_q = 1
            cu_seqlens_q = torch.arange(
                batch_size + 1, dtype=torch.int32, device=query_layer.device
            )  # There is a memcpy here, that is very bad.
            indices_q = cu_seqlens_q[:-1]
            query_layer = query_layer.squeeze(1)
        else:
            # The -q_len: slice assumes left padding.
            attention_mask = attention_mask[:, -query_length:]
            query_layer, indices_q, cu_seqlens_q, max_seqlen_in_batch_q = unpad_input(
                query_layer, attention_mask
            )

        return (
            query_layer,
            key_layer,
            value_layer,
            indices_q,
            (cu_seqlens_q, cu_seqlens_k),
            (max_seqlen_in_batch_q, max_seqlen_in_batch_k),
        )


class FLMAudioSdpaAttention(FLMAudioAttention):
    """
    FLM-Audio attention module using torch.nn.functional.scaled_dot_product_attention. This module inherits from
    `FLMAudioAttention` as the weights of the module stays untouched. The only changes are on the forward pass to adapt to
    SDPA API.
    """

    # Adapted from FLMAudioAttention.forward
    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_value: Optional[Cache] = None,
        output_attentions: bool = False,
        use_cache: bool = False,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,  # necessary, but kept here for BC
    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
        if output_attentions:
            # TODO: Improve this warning with e.g. `model.config.attn_implementation = "manual"` once this is implemented.
            logger.warning_once(
                "FLMAudioModel is using FLMAudioSdpaAttention, but `torch.nn.functional.scaled_dot_product_attention` does not support `output_attentions=True`. Falling back to the manual attention implementation, "
                'but specifying the manual implementation will be required from Transformers version v5.0.0 onwards. This warning can be removed using the argument `attn_implementation="eager"` when loading the model.'
            )
            return super().forward(
                hidden_states=hidden_states,
                attention_mask=attention_mask,
                position_ids=position_ids,
                past_key_value=past_key_value,
                output_attentions=output_attentions,
                use_cache=use_cache,
                cache_position=cache_position,
            )

        bsz, q_len, _ = hidden_states.size()

        query_states = self.q_proj(hidden_states)
        key_states = self.k_proj(hidden_states)
        value_states = self.v_proj(hidden_states)

        query_states = query_states.view(
            bsz, q_len, self.num_heads, self.head_dim
        ).transpose(1, 2)
        key_states = key_states.view(
            bsz, q_len, self.num_key_value_heads, self.head_dim
        ).transpose(1, 2)
        value_states = value_states.view(
            bsz, q_len, self.num_key_value_heads, self.head_dim
        ).transpose(1, 2)

        cos, sin = position_embeddings

        query_states, key_states = apply_multimodal_rotary_pos_emb(
            query_states, key_states, cos, sin, self.rope_scaling["mrope_section"]
        )

        if past_key_value is not None:
            # sin and cos are specific to RoPE models; cache_position needed for the static cache
            cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
            key_states, value_states = past_key_value.update(
                key_states, value_states, self.layer_idx, cache_kwargs
            )

        key_states = repeat_kv(key_states, self.num_key_value_groups)
        value_states = repeat_kv(value_states, self.num_key_value_groups)

        causal_mask = attention_mask
        # if attention_mask is not None and cache_position is not None:
        if attention_mask is not None:
            causal_mask = causal_mask[:, :, :, : key_states.shape[-2]]

        # SDPA with memory-efficient backend is currently (torch==2.1.2) bugged with non-contiguous inputs with custom attn_mask,
        # Reference: https://github.com/pytorch/pytorch/issues/112577.
        if query_states.device.type == "cuda" and causal_mask is not None:
            query_states = query_states.contiguous()
            key_states = key_states.contiguous()
            value_states = value_states.contiguous()

        attn_output = F.scaled_dot_product_attention(
            query_states,
            key_states,
            value_states,
            attn_mask=causal_mask,
            dropout_p=self.attention_dropout if self.training else 0.0,
        )

        attn_output = attn_output.transpose(1, 2).contiguous()
        attn_output = attn_output.view(bsz, q_len, self.hidden_size)

        attn_output = self.o_proj(attn_output)

        return attn_output, None, past_key_value


FLMAUDIO_ATTENTION_CLASSES = {
    "eager": FLMAudioAttention,
    "flash_attention_2": FLMAudioFlashAttention2,
    "sdpa": FLMAudioSdpaAttention,
}


class FLMAudioDecoderLayer(nn.Module):
    def __init__(self, config: FLMAudioConfig, layer_idx: int):
        super().__init__()
        self.hidden_size = config.hidden_size
        self.self_attn = FLMAUDIO_ATTENTION_CLASSES.get(
            config._attn_implementation, FLMAudioAttention
        )(config=config, layer_idx=layer_idx)
        self.mlp = FLMAudioMLP(config)
        self.input_layernorm = FLMAudioRMSNorm(
            config.hidden_size, eps=config.rms_norm_eps
        )
        self.post_attention_layernorm = FLMAudioRMSNorm(
            config.hidden_size, eps=config.rms_norm_eps
        )

    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_value: Optional[Tuple[torch.Tensor]] = None,
        output_attentions: Optional[bool] = False,
        use_cache: Optional[bool] = False,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,  # necessary, but kept here for BC
        **kwargs,
    ) -> Tuple[
        torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]
    ]:
        """
        Args:
            hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)`
            attention_mask (`torch.FloatTensor`, *optional*):
                attention mask of size `(batch_size, sequence_length)` if flash attention is used or `(batch_size, 1,
                query_sequence_length, key_sequence_length)` if default attention is used.
            output_attentions (`bool`, *optional*):
                Whether or not to return the attentions tensors of all attention layers. See `attentions` under
                returned tensors for more detail.
            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`).
            past_key_value (`Tuple(torch.FloatTensor)`, *optional*): cached past key and value projection states
        """
        if "padding_mask" in kwargs:
            warnings.warn(
                "Passing `padding_mask` is deprecated and will be removed in v4.37. Please make sure use `attention_mask` instead.`"
            )

        residual = hidden_states

        hidden_states = self.input_layernorm(hidden_states)

        # Self Attention
        hidden_states, self_attn_weights, present_key_value = self.self_attn(
            hidden_states=hidden_states,
            attention_mask=attention_mask,
            position_ids=position_ids,
            past_key_value=past_key_value,
            output_attentions=output_attentions,
            use_cache=use_cache,
            cache_position=cache_position,
            position_embeddings=position_embeddings,
            **kwargs,
        )
        hidden_states = residual + hidden_states

        # Fully Connected
        residual = hidden_states
        hidden_states = self.post_attention_layernorm(hidden_states)
        hidden_states = self.mlp(hidden_states)
        hidden_states = residual + hidden_states

        outputs = (hidden_states,)

        if output_attentions:
            outputs += (self_attn_weights,)

        if use_cache:
            outputs += (present_key_value,)

        return outputs


FLMAUDIO_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 ([`FLMAudioConfig`]):
            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.
"""


@add_start_docstrings(
    "The bare FLM-Audio Model outputting raw hidden-states without any specific head on top.",
    FLMAUDIO_START_DOCSTRING,
)
class FLMAudioPreTrainedModel(PreTrainedModel):
    config_class = FLMAudioConfig
    base_model_prefix = "model"
    supports_gradient_checkpointing = True
    _no_split_modules = ["FLMAudioDecoderLayer"]
    _skip_keys_device_placement = ["past_key_values"]
    _supports_flash_attn_2 = True
    _supports_sdpa = True
    _supports_cache_class = True

    def _init_weights(self, module):
        std = self.config.initializer_range
        if isinstance(module, nn.Linear):
            module.weight.data.normal_(mean=0.0, std=std)
            if module.bias is not None:
                module.bias.data.zero_()
        elif isinstance(module, nn.Embedding):
            module.weight.data.normal_(mean=0.0, std=std)
            if module.padding_idx is not None:
                module.weight.data[module.padding_idx].zero_()

    def _setup_cache(
        self, cache_cls, max_batch_size, max_cache_len: Optional[int] = None
    ):
        if (
            self.config._attn_implementation == "flash_attention_2"
            and cache_cls == StaticCache
        ):
            raise ValueError(
                "`static` cache implementation is not compatible with `attn_implementation==flash_attention_2` "
                "make sure to use `sdpa` in the mean time, and open an issue at https://github.com/huggingface/transformers"
            )

        for layer in self.model.layers:
            device = layer.input_layernorm.weight.device
            if hasattr(self.config, "_pre_quantization_dtype"):
                dtype = self.config._pre_quantization_dtype
            else:
                dtype = layer.self_attn.o_proj.weight.dtype
            layer.self_attn.past_key_value = cache_cls(
                self.config, max_batch_size, max_cache_len, device=device, dtype=dtype
            )

    def _reset_cache(self):
        for layer in self.model.layers:
            layer.self_attn.past_key_value = None


class MultiModalEmbedding(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.config = config
        self.use_mup = config.use_mup
        self.input_mult = config.input_mult
        self.hidden_size = config.hidden_size

        self.vocab_size = config.vocab_size
        self.aud_vocab_size = config.aud_vocab_size

        self.aud_channel = config.aud_channel

        self.aud_emp_token_id = config.mm_token_info.aud_emp_token_id

        self.text_embeddings = nn.Embedding(self.vocab_size, self.hidden_size)

        self.aud_listen_embeddings = nn.ModuleList(
            [
                nn.Embedding(self.aud_vocab_size, self.hidden_size)
                for _ in range(self.aud_channel)
            ]
        )
        self.aud_speak_embeddings = nn.ModuleList(
            [
                nn.Embedding(self.aud_vocab_size, self.hidden_size)
                for _ in range(self.aud_channel)
            ]
        )

    @staticmethod
    def merge_multichannel_embeddings(
        token_ids, embedding_layer, emp_token_id, embeddings
    ):
        if token_ids is not None and embedding_layer is not None:
            assert token_ids.shape[2] == len(embedding_layer)
            for c in range(token_ids.shape[2]):
                _emb_state = embedding_layer[c](token_ids[:, :, c])
                _emb_state[token_ids[:, :, c] == emp_token_id] = 0.0
                embeddings += _emb_state
            _emb_state = None
            del _emb_state
        return embeddings

    def forward(
        self,
        text_ids,
        speak_ids,
        listen_ids,
    ):
        assert text_ids is not None
        embeddings = self.text_embeddings(text_ids)
        mask = ~(text_ids == self.config.pad_token_id)

        for aud_chn_idx in range(self.aud_channel):
            aud_speak_embed = self.aud_speak_embeddings[aud_chn_idx](
                speak_ids[..., aud_chn_idx]
            ).squeeze(0)
            aud_listen_embed = self.aud_listen_embeddings[aud_chn_idx](
                listen_ids[..., aud_chn_idx]
            ).squeeze(0)
            embeddings[mask] += aud_speak_embed + aud_listen_embed

        if self.use_mup:
            embeddings = embeddings * self.input_mult

        return embeddings


FLMAUDIO_INPUTS_DOCSTRING = r"""
    Args:
        input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
            Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
            it.

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

            [What are input IDs?](../glossary#input-ids)
        attention_mask (`torch.Tensor` 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**.

            [What are attention masks?](../glossary#attention-mask)

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

            If `past_key_values` is used, optionally only the last `input_ids` have to be input (see
            `past_key_values`).

            If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`]
            and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more
            information on the default strategy.

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

            [What are position IDs?](../glossary#position-ids)
        past_key_values (`Cache` or `tuple(tuple(torch.FloatTensor))`, *optional*):
            Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
            blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values`
            returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`.

            Two formats are allowed:
            - a [`~cache_utils.Cache`] instance;
            - 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)`). This is also known as the legacy
            cache format.

            The model will output the same cache format that is fed as input. If no `past_key_values` are passed, the
            legacy cache format will be returned.

            If `past_key_values` are used, the user can optionally input only the last `input_ids` (those that don't
            have their past key value states given to this model) of shape `(batch_size, 1)` instead of all `input_ids`
            of shape `(batch_size, sequence_length)`.
        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.
        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.
        cache_position (`torch.LongTensor` of shape `(sequence_length)`, *optional*):
            Indices depicting the position of the input sequence tokens in the sequence. Contrarily to `position_ids`,
            this tensor is not affected by padding. It is used to update the cache in the correct position and to infer
            the complete sequence length.
"""


@add_start_docstrings(
    "The bare FLM-Audio Model outputting raw hidden-states without any specific head on top.",
    FLMAUDIO_START_DOCSTRING,
)
class FLMAudioModel(FLMAudioPreTrainedModel):
    """
    Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`FLMAudioDecoderLayer`]

    Args:
        config: FLMAudioConfig
    """

    def __init__(self, config: FLMAudioConfig):
        super().__init__(config)
        self.padding_idx = config.pad_token_id
        self.vocab_size = config.vocab_size

        self.embed_tokens = MultiModalEmbedding(config)
        self.layers = nn.ModuleList(
            [
                FLMAudioDecoderLayer(config, layer_idx)
                for layer_idx in range(config.num_hidden_layers)
            ]
        )
        self.norm = FLMAudioRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.rotary_emb = FLMAudioRotaryEmbedding(config=config)
        self.gradient_checkpointing = False
        self.rope_deltas = None  # cache rope_deltas here

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

    def get_input_embeddings(self) -> MultiModalEmbedding:
        return self.embed_tokens

    def set_input_embeddings(self, value: MultiModalEmbedding):
        self.embed_tokens = value

    def get_rope_index(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        second_per_grid_ts: Optional[torch.Tensor] = None,
        attention_mask: Optional[torch.Tensor] = None,
    ) -> Tuple[torch.Tensor, torch.Tensor]:

        mrope_position_deltas = []

        if attention_mask is not None:
            position_ids = attention_mask.long().cumsum(-1) - 1
            position_ids.masked_fill_(attention_mask == 0, 1)
            position_ids = position_ids.unsqueeze(0).expand(3, -1, -1).to(attention_mask.device)
            max_position_ids = position_ids.max(0, keepdim=False)[0].max(-1, keepdim=True)[0]
            mrope_position_deltas = max_position_ids + 1 - attention_mask.shape[-1]
        else:
            position_ids = (
                torch.arange(input_ids.shape[1], device=input_ids.device)
                .view(1, 1, -1)
                .expand(3, input_ids.shape[0], -1)
            )
            mrope_position_deltas = torch.zeros(
                [input_ids.shape[0], 1],
                device=input_ids.device,
                dtype=input_ids.dtype,
            )

        return position_ids, mrope_position_deltas


    @add_start_docstrings_to_model_forward(FLMAUDIO_INPUTS_DOCSTRING)
    def forward(
        self,
        text_ids: torch.LongTensor = None,
        listen_ids: torch.LongTensor = None,
        speak_ids: torch.LongTensor = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[List[torch.FloatTensor]] = None,
        inputs_embeds: 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,
        rope_deltas: Optional[torch.LongTensor] = None,
        cache_position: Optional[torch.LongTensor] = None,
        second_per_grid_ts: Optional[torch.Tensor] = None,
        **kwargs,
    ) -> Union[Tuple, BaseModelOutputWithPast]:
        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 (text_ids is None) ^ (inputs_embeds is not None):
            raise ValueError(
                "You cannot specify both input_ids and inputs_embeds at the same time, and must specify either one"
            )

        if self.gradient_checkpointing and self.training and use_cache:
            logger.warning_once(
                "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`."
            )
            use_cache = False

        if inputs_embeds is None:
            inputs_embeds = self.embed_tokens(
                text_ids,
                speak_ids,
                listen_ids,
            )

        past_seen_tokens = 0
        if use_cache:  # kept for BC (cache positions)
            if not isinstance(past_key_values, StaticCache):
                past_key_values = DynamicCache.from_legacy_cache(past_key_values)
                past_seen_tokens = past_key_values.get_seq_length()

        if cache_position is None:
            if isinstance(past_key_values, StaticCache):
                raise ValueError(
                    "cache_position is a required argument when using StaticCache."
                )
            cache_position = torch.arange(
                past_seen_tokens,
                past_seen_tokens + inputs_embeds.shape[1],
                device=inputs_embeds.device,
            )

        # if we get 4D attention mask we cannot calculate rope deltas anymore. TODO @raushan fixme
        if position_ids is None:
            # calculate RoPE index once per generation in the pre-fill stage only
            if (
                (cache_position is not None and cache_position[0] == 0)
                or self.rope_deltas is None
                or (past_key_values is None or past_key_values.get_seq_length() == 0)
            ):
                position_ids, rope_deltas = self.get_rope_index(
                    text_ids,
                    second_per_grid_ts,
                    attention_mask,
                )
                self.rope_deltas = rope_deltas
            # then use the prev pre-calculated rope-deltas to get the correct position ids
            else:
                batch_size, seq_length, _ = inputs_embeds.shape
                delta = (
                    (cache_position[0] + self.rope_deltas).to(inputs_embeds.device)
                    if cache_position is not None
                    else 0
                )
                position_ids = torch.arange(seq_length, device=inputs_embeds.device)
                position_ids = position_ids.view(1, -1).expand(batch_size, -1)
                if cache_position is not None:  # otherwise `deltas` is an int `0`
                    delta = delta.repeat_interleave(batch_size // delta.shape[0], dim=0)
                position_ids = position_ids.add(delta)
                position_ids = position_ids.unsqueeze(0).expand(3, -1, -1)

        causal_mask = self._update_causal_mask(
            attention_mask, inputs_embeds, cache_position
        )

        # embed positions
        hidden_states = inputs_embeds

        position_embeddings = self.rotary_emb(hidden_states, position_ids)

        # decoder layers
        all_hidden_states = () if output_hidden_states else None
        all_self_attns = () if output_attentions else None
        next_decoder_cache = None

        for decoder_layer in self.layers:
            if output_hidden_states:
                all_hidden_states += (hidden_states,)

            if self.gradient_checkpointing and self.training:
                layer_outputs = self._gradient_checkpointing_func(
                    decoder_layer.__call__,
                    hidden_states,
                    causal_mask,
                    position_ids,
                    past_key_values,
                    output_attentions,
                    use_cache,
                    cache_position,
                    position_embeddings,
                )
            else:
                layer_outputs = decoder_layer(
                    hidden_states,
                    attention_mask=causal_mask,
                    position_ids=position_ids,
                    past_key_value=past_key_values,
                    output_attentions=output_attentions,
                    use_cache=use_cache,
                    cache_position=cache_position,
                    position_embeddings=position_embeddings,
                )

            hidden_states = layer_outputs[0]

            if use_cache:
                next_decoder_cache = layer_outputs[2 if output_attentions else 1]

            if output_attentions:
                all_self_attns += (layer_outputs[1],)

        hidden_states = self.norm(hidden_states)

        # add hidden states from the last decoder layer
        if output_hidden_states:
            all_hidden_states += (hidden_states,)

        next_cache = None
        if use_cache:
            next_cache = (
                next_decoder_cache.to_legacy_cache()
                if isinstance(next_decoder_cache, Cache)
                else next_decoder_cache
            )
        if not return_dict:
            return tuple(
                v
                for v in [hidden_states, next_cache, all_hidden_states, all_self_attns]
                if v is not None
            )
        return BaseModelOutputWithPast(
            last_hidden_state=hidden_states,
            past_key_values=next_cache,
            hidden_states=all_hidden_states,
            attentions=all_self_attns,
        )

    # TODO: As of torch==2.2.0, the `attention_mask` passed to the model in `generate` is 2D and of dynamic length even when the static
    # KV cache is used. This is an issue for torch.compile which then recaptures cudagraphs at each decode steps due to the dynamic shapes.
    # (`recording cudagraph tree for symint key 13`, etc.), which is VERY slow. A workaround is `@torch.compiler.disable`, but this prevents using
    # `fullgraph=True`. See more context in https://github.com/huggingface/transformers/pull/29114
    def _update_causal_mask(self, attention_mask, input_tensor, cache_position):
        if self.config._attn_implementation == "flash_attention_2":
            if attention_mask is not None and 0.0 in attention_mask:
                return attention_mask
            return None

        dtype, device = input_tensor.dtype, input_tensor.device
        min_dtype = torch.finfo(dtype).min
        sequence_length = input_tensor.shape[1]
        if hasattr(
            getattr(self.layers[0], "self_attn", {}), "past_key_value"
        ):  # static cache
            target_length = self.config.max_position_embeddings
        else:  # dynamic cache
            target_length = (
                attention_mask.shape[-1]
                if isinstance(attention_mask, torch.Tensor)
                else cache_position[-1] + 1
            )

        causal_mask = torch.full(
            (sequence_length, target_length),
            fill_value=min_dtype,
            dtype=dtype,
            device=device,
        )
        if sequence_length != 1:
            causal_mask = torch.triu(causal_mask, diagonal=1)
        causal_mask *= torch.arange(
            target_length, device=device
        ) > cache_position.reshape(-1, 1)
        causal_mask = causal_mask[None, None, :, :].expand(
            input_tensor.shape[0], 1, -1, -1
        )
        if attention_mask is not None:
            causal_mask = (
                causal_mask.clone()
            )  # copy to contiguous memory for in-place edit
            if attention_mask.dim() == 2:
                mask_length = attention_mask.shape[-1]
                padding_mask = causal_mask[..., :mask_length].eq(0.0) * attention_mask[
                    :, None, None, :
                ].eq(0.0)
                causal_mask[..., :mask_length] = causal_mask[
                    ..., :mask_length
                ].masked_fill(padding_mask, min_dtype)
            elif attention_mask.dim() == 4:
                # backwards compatibility: we allow passing a 4D attention mask shorter than the input length with
                # cache. In that case, the 4D attention mask attends to the newest tokens only.
                if attention_mask.shape[-2] < cache_position[0] + sequence_length:
                    offset = cache_position[0]
                else:
                    offset = 0
                mask_shape = attention_mask.shape
                mask_slice = (attention_mask.eq(0.0)).to(dtype=dtype) * min_dtype
                causal_mask[
                    : mask_shape[0],
                    : mask_shape[1],
                    offset : mask_shape[2] + offset,
                    : mask_shape[3],
                ] = mask_slice

        if (
            self.config._attn_implementation == "sdpa"
            and attention_mask is not None
            and attention_mask.device.type == "cuda"
        ):
            # Attend to all tokens in fully masked rows in the causal_mask, for example the relevant first rows when
            # using left padding. This is required by F.scaled_dot_product_attention memory-efficient attention path.
            # Details: https://github.com/pytorch/pytorch/issues/110213
            causal_mask = AttentionMaskConverter._unmask_unattended(
                causal_mask, min_dtype
            )

        return causal_mask


@dataclass
class FLMAudioCausalLMOutputWithPast(ModelOutput):
    loss: Optional[torch.FloatTensor] = None
    logits: torch.FloatTensor = None
    audio_logits: torch.FloatTensor = None
    past_key_values: Optional[List[torch.FloatTensor]] = None
    hidden_states: Optional[Tuple[torch.FloatTensor]] = None
    attentions: Optional[Tuple[torch.FloatTensor]] = None
    rope_deltas: Optional[torch.LongTensor] = None


class FLMAudioForCausalLM(FLMAudioPreTrainedModel):
    _tied_weights_keys = ["lm_head.weight"]

    def __init__(self, config):
        super().__init__(config)
        self.model = FLMAudioModel(config)
        self.vocab_size = config.vocab_size
        self.output_mult = config.output_mult

        depth_config = DepthGPTConfig(
            block_size=config.aud_channel,
            vocab_size=config.aud_vocab_size,
            n_layer=config.aud_depthgpt.n_layer,
            n_head=config.aud_depthgpt.n_head,
            n_embd=config.aud_depthgpt.n_embd,
            dropout=config.aud_depthgpt.dropout,
            bias=config.aud_depthgpt.bias,
            main_hidden_size=config.hidden_size,
            pad_token_id=config.mm_token_info.aud_emp_token_id,
            use_cmlp=config.aud_depthgpt.use_cmlp,
            use_rmsnorm=config.aud_depthgpt.use_rmsnorm,
            use_swiglu=config.aud_depthgpt.use_swiglu,
        )

        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)

        self.aud_output_layers = DepthGPT(depth_config)

        self.use_mup = config.use_mup
        if self.use_mup:
            self.mup_scale_factor = config.mup_scale_factor
            self.output_mult = config.output_mult / self.mup_scale_factor
        # Initialize weights and apply final processing
        self.post_init()

    def get_input_embeddings(self):
        return self.model.embed_tokens

    def set_input_embeddings(self, value):
        self.model.embed_tokens = value

    def get_output_embeddings(self):
        return self.lm_head

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

    def set_decoder(self, decoder):
        self.model = decoder

    def get_decoder(self):
        return self.model

    def _forward_text(self, outputs, labels, return_dict):

        logits = self.lm_head(outputs[0])
        logits = logits.float()
        # Mup
        if self.use_mup:
            logits = logits * self.output_mult

        loss = None
        if labels is not None:
            raise NotImplementedError

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

        return FLMAudioCausalLMOutputWithPast(
            loss=loss,
            logits=logits,
            past_key_values=outputs.past_key_values,
            hidden_states=outputs.last_hidden_state,
            attentions=outputs.attentions,
        )

    def forward_audio(self, transformer_output_states, audio_input_ids):
        return self.aud_output_layers(transformer_output_states, audio_input_ids)

    @add_start_docstrings_to_model_forward(FLMAUDIO_INPUTS_DOCSTRING)
    @replace_return_docstrings(
        output_type=CausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC
    )
    def forward(
        self,
        input_ids: torch.LongTensor = None,
        listen_ids: torch.LongTensor = None,
        speak_ids: torch.LongTensor = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[List[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,
        rope_deltas: Optional[torch.LongTensor] = None,
        cache_position: Optional[torch.LongTensor] = None,
        second_per_grid_ts: Optional[torch.Tensor] = None,
        **kwargs,
    ) -> Union[Tuple, CausalLMOutputWithPast]:
        r"""
        Args:
            labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
                Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
                config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
                (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.

        Returns:

        Example:

        ```python
        >>> from transformers import AutoTokenizer, FLMAudioForCausalLM

        >>> model = FLMAudioForCausalLM.from_pretrained("CofeAI/FLM-Audio")
        >>> tokenizer = AutoTokenizer.from_pretrained("CofeAI/FLM-Audio")

        >>> prompt = "Hey, are you conscious? Can you talk to me?"
        >>> inputs = tokenizer(prompt, return_tensors="pt")

        >>> # Generate
        >>> generate_ids = model.generate(inputs.input_ids, max_length=30)
        >>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
        "Hey, are you conscious? Can you talk to me?\nI'm not conscious, but I can talk to you."
        ```"""
        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
        )

        # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)

        if listen_ids is None and speak_ids is None:
            batch_size, seq_len = input_ids.shape[:2]
            listen_ids = torch.full((seq_len*batch_size, 8), self.model.config.mm_token_info.aud_pad_token_id, device=input_ids.device, dtype=input_ids.dtype)
            speak_ids = torch.full((seq_len*batch_size, 8), self.model.config.mm_token_info.aud_pad_token_id, device=input_ids.device, dtype=input_ids.dtype)
        outputs = self.model(
            text_ids=input_ids,
            listen_ids=listen_ids,
            speak_ids=speak_ids,
            attention_mask=attention_mask,
            position_ids=position_ids,
            past_key_values=past_key_values,
            inputs_embeds=inputs_embeds,
            use_cache=use_cache,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
            cache_position=cache_position,
            second_per_grid_ts=second_per_grid_ts
        )
        return self._forward_text(outputs, labels, return_dict)

    @staticmethod
    def _reorder_cache(past_key_values, beam_idx):
        reordered_past = ()
        for layer_past in past_key_values:
            reordered_past += (
                tuple(
                    past_state.index_select(0, beam_idx.to(past_state.device))
                    for past_state in layer_past
                ),
            )
        return reordered_past

    def _get_initial_token(self) -> torch.Tensor:
        # Returns the initial token that will be fed to the model to predict the very first timestep.
        # The output shape will be [B, K, 1].
        device = next(iter(self.parameters())).device
        zero = torch.full([1, 1, 1], 0, device=device, dtype=torch.long)
        special = torch.full_like(zero, self.config.mm_token_info.aud_pad_token_id)

        text_special = torch.full_like(
            zero, self.config.mm_token_info.text_wait_token_id
        )
        audio_token = special
        text_token = text_special
        audio_token = audio_token.expand(-1, 2 * self.config.aud_channel, -1).clone()
        token = torch.cat([text_token, audio_token], dim=1)
        return token