modeling_cheems.py

# coding=utf-8
# Copyright 2024 Jingze Shi and the HuggingFace Inc. team. All rights reserved.
#
# This code is based on the Wonderful Matrices paper implementation.
#
#     https://arxiv.org/abs/2412.11834
#
# 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 Cheems model."""

import math
from dataclasses import dataclass, field
from typing import Optional, Tuple, Union, Any, Dict, List

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

from transformers.activations import ACT2FN
from transformers.cache_utils import Cache, DynamicCache, StaticCache
from transformers.generation import GenerationMixin
from transformers.modeling_outputs import (
    BaseModelOutputWithPast,
    CausalLMOutputWithPast,
    SequenceClassifierOutputWithPast,
)
from transformers.modeling_rope_utils import ROPE_INIT_FUNCTIONS
from transformers.modeling_utils import PreTrainedModel
from transformers.utils import (
    add_start_docstrings,
    add_start_docstrings_to_model_forward,
    logging,
    replace_return_docstrings,
)
from transformers.utils.import_utils import is_mamba_2_ssm_available
from .configuration_cheems import CheemsConfig

try:
    from einx import add as einx_add
except ImportError:
    einx_add = None


logger = logging.get_logger(__name__)


if is_torch_greater_or_equal("2.5"):
    from torch.nn.attention.flex_attention import flex_attention

if is_mamba_2_ssm_available():
    from mamba_ssm.ops.triton.selective_state_update import selective_state_update
    from mamba_ssm.ops.triton.ssd_combined import mamba_chunk_scan_combined, mamba_split_conv1d_scan_combined
else:
    selective_state_update = None

is_fast_path_available = all((selective_state_update, mamba_chunk_scan_combined, mamba_split_conv1d_scan_combined))

_CONFIG_FOR_DOC = "CheemsConfig"


class RMSNorm(nn.Module):
    def __init__(self, hidden_size, eps=1e-6):
        """
        RMSNorm 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)

    def extra_repr(self):
        return f"{tuple(self.weight.shape)}, eps={self.variance_epsilon}"


class Residual(nn.Module):
    def __init__(self, hidden_size):
        super().__init__()
        self.weight = nn.Parameter(torch.ones(hidden_size))
    
    def forward(self, residual_states, hidden_states):
        return self.weight * residual_states + hidden_states

    def extra_repr(self):
        return f"{tuple(self.weight.shape)}"
    

class RotaryEmbedding(nn.Module):
    def __init__(self, config: Optional[DogeConfig] = None):
        super().__init__()
        self.rope_kwargs = {}

        if 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.base = config.rope_theta

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

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

    def _dynamic_frequency_update(self, position_ids, device):
        """
        dynamic RoPE layers should recompute `inv_freq` in the following situations:
        1 - growing beyond the cached sequence length (allow scaling)
        2 - the current sequence length is in the original scale (avoid losing precision with small sequences)
        """
        seq_len = torch.max(position_ids) + 1
        if seq_len > self.max_seq_len_cached:  # growth
            inv_freq, self.attention_scaling = self.rope_init_fn(
                self.config, device, seq_len=seq_len, **self.rope_kwargs
            )
            self.register_buffer("inv_freq", inv_freq, persistent=False)  # TODO joao: may break with compilation
            self.max_seq_len_cached = seq_len

        if seq_len < self.original_max_seq_len and self.max_seq_len_cached > self.original_max_seq_len:  # reset
            self.register_buffer("inv_freq", self.original_inv_freq, persistent=False)
            self.max_seq_len_cached = self.original_max_seq_len

    @torch.no_grad()
    def forward(self, x, position_ids):
        if "dynamic" in self.rope_type:
            self._dynamic_frequency_update(position_ids, device=x.device)

        # core RoPE block
        inv_freq_expanded = self.inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1)
        position_ids_expanded = position_ids[:, None, :].float()
        # Force float32 (see https://github.com/huggingface/transformers/pull/29285)
        device_type = x.device.type
        device_type = device_type if isinstance(device_type, str) and device_type != "mps" else "cpu"
        with torch.autocast(device_type=device_type, enabled=False):
            freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(1, 2)
            emb = torch.cat((freqs, freqs), dim=-1)
            cos = emb.cos()
            sin = emb.sin()

        # Advanced RoPE types (e.g. yarn) apply a post-processing scaling factor, equivalent to scaling attention
        cos = cos * self.attention_scaling
        sin = 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_CB_rotary_pos_emb(c, b, cos, sin, position_ids=None, unsqueeze_dim=2):
    cos = cos.unsqueeze(unsqueeze_dim)
    sin = sin.unsqueeze(unsqueeze_dim)
    c_embed = (c * cos) + (rotate_half(c) * sin)
    b_embed = (b * cos) + (rotate_half(b) * sin)
    return c_embed, b_embed


def apply_QK_rotary_pos_emb(q, k, cos, sin, position_ids=None, unsqueeze_dim=1):
    """Applies Rotary Position Embedding to the query and key tensors.

    Args:
        q (`torch.Tensor`): The query tensor.
        k (`torch.Tensor`): The key tensor.
        cos (`torch.Tensor`): The cosine part of the rotary embedding.
        sin (`torch.Tensor`): The sine part of the rotary embedding.
        position_ids (`torch.Tensor`, *optional*):
            Deprecated and unused.
        unsqueeze_dim (`int`, *optional*, defaults to 1):
            The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and
            sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. 
            For example, note that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. 
            Then, if q and k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. 
            Similarly, if q and k have the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2.
    Returns:
        `tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding.
    """
    cos = cos.unsqueeze(unsqueeze_dim)
    sin = sin.unsqueeze(unsqueeze_dim)
    q_embed = (q * cos) + (rotate_half(q) * sin)
    k_embed = (k * cos) + (rotate_half(k) * sin)
    return q_embed, k_embed


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 HybridSSDAttnDynamicCache(DynamicCache):
    """
    A dynamic cache that can handle both the attn cache (which has a seq_len dimension) and the ssd cache (which has a constant shape regardless of seq_len).

    This cache has two sets of lists of tensors: `key_cache` and `value_cache` for attn cache and `ssd_states` for ssd cache. 
    Each of these lists has `num_layers` tensors. The expected shape for each tensor.

    For attn layers, `key_cache` and `value_cache` have a shape of `(batch_size, num_attn_heads, seq_len, attn_head_dim)`,
    while `ssd_states` have a shape of `(batch_size, 0)` (empty tensors).

    For ssd layers, `key_cache` and `value_cache` have a shape of `(batch_size, 0)` (empty tensors),
    while `ssm_states` represents the ssm state and has a shape of `(batch_size, num_ssd_heads, ssd_head_dim, ssd_state_size)`.
    """
    def __init__(self, config: CheemsConfig, batch_size, dtype=torch.float16, device=None, layer_type=None):
        self.dtype = dtype
        self.layer_type = layer_type

        self.has_previous_state = False  # only used by ssd
        self.ssd_head_dim = config.hidden_size // config.num_attention_heads
        self.ssd_states = []
        for i in range(config.num_hidden_layers):
            if layer_type is None:
                has_ssd_state = True
            else:
                has_ssd_state = layer_type[i] == "ssd"
            
            if has_ssd_state:
                self.ssd_states += [
                    torch.zeros(
                        batch_size,
                        config.num_attention_heads,
                        self.ssd_head_dim,
                        self.ssd_head_dim,
                        device=device, dtype=dtype
                    )
                ]
            else:
                self.ssd_states += [torch.tensor([[]] * batch_size, device=device, dtype=dtype)]
        self.ssd_past_length = [0 for _ in range(config.num_hidden_layers)]
        
        self.key_cache = [torch.tensor([[]] * batch_size, device=device, dtype=dtype) for _ in range(config.num_hidden_layers)]
        self.value_cache = [torch.tensor([[]] * batch_size, device=device, dtype=dtype) for _ in range(config.num_hidden_layers)]
    
    def update(
        self,
        key_states: torch.Tensor,
        value_states: torch.Tensor,
        layer_idx: int,
        cache_kwargs: Optional[Dict[str, Any]] = None,
    ) -> Tuple[torch.Tensor, torch.Tensor]:
        # Update the cache
        if self.key_cache[layer_idx].shape[-1] == 0:
            self.key_cache[layer_idx] = key_states
            self.value_cache[layer_idx] = value_states
        else:
            self.key_cache[layer_idx] = torch.cat([self.key_cache[layer_idx], key_states], dim=2)
            self.value_cache[layer_idx] = torch.cat([self.value_cache[layer_idx], value_states], dim=2)

        return self.key_cache[layer_idx], self.value_cache[layer_idx]

    def reorder_cache(self, beam_idx: torch.LongTensor):
        """Reorders the cache for beam search, given the selected beam indices."""
        for layer_idx in range(len(self.key_cache)):
            device = self.key_cache[layer_idx].device
            self.key_cache[layer_idx] = self.key_cache[layer_idx].index_select(0, beam_idx.to(device))
            device = self.value_cache[layer_idx].device
            self.value_cache[layer_idx] = self.value_cache[layer_idx].index_select(0, beam_idx.to(device))
            device = self.ssd_states[layer_idx].device
            self.ssd_states[layer_idx] = self.ssd_states[layer_idx].index_select(0, beam_idx.to(device))
    
    def get_seq_length(self, layer_idx: Optional[int] = 0) -> int:
        """Returns the sequence length of the cached states. A layer index can be optionally passed."""
        # take any layer that contains cache and not empty tensor

        if self.layer_type[layer_idx] == 'ssd':
            return self.ssd_past_length[layer_idx]

        if self.key_cache[layer_idx].shape[-1] == 0:
            return 0

        return self.key_cache[layer_idx].shape[-2]
    
    def to_legacy_cache(self) -> Tuple[Tuple[torch.Tensor], Tuple[torch.Tensor]]:
        raise NotImplementedError("HybridSSDAttnDynamicCache does not have a legacy cache equivalent.")

    @classmethod
    def from_legacy_cache(cls, past_key_values: Optional[Tuple[Tuple[torch.FloatTensor]]] = None) -> "DynamicCache":
        raise NotImplementedError("HybridSSDAttnDynamicCache does not have a legacy cache equivalent.")


class CheemsStateSpaceDuality(nn.Module):
    """State Space Duality from 'Transformers are SSMs' paper."""

    def __init__(self, config: CheemsConfig, layer_idx: Optional[int] = None):
        super().__init__()

        self.config = config
        self.layer_idx = layer_idx

        self.hidden_dim = config.hidden_size
        self.num_heads = config.num_attention_heads
        self.head_dim = self.hidden_dim // self.num_heads
        self.chunk_len = config.ssd_chunk_size

        self.c_proj = nn.Linear(
            self.hidden_dim,
            self.num_heads * self.head_dim,
            bias=config.hidden_bias,
        )
        self.b_proj = nn.Linear(
            self.hidden_dim,
            self.num_heads * self.head_dim,
            bias=config.hidden_bias,
        )
        self.A = nn.Parameter(
            torch.ones(self.num_heads)
        )
        self.dt_proj = nn.Linear(
            self.hidden_dim,
            self.num_heads,
            bias=config.hidden_bias,
        )
        self.x_proj = nn.Linear(
            self.hidden_dim,
            self.num_heads * self.head_dim,
            bias=config.hidden_bias,
        )
        self.out_proj = nn.Linear(
            self.hidden_dim,
            self.hidden_dim,
            bias=config.hidden_bias,
        )
    
    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        cache_params: HybridSSDAttnDynamicCache = None,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
        **kwargs,
    ) -> torch.Tensor:
        bsz, c_len, _ = hidden_states.shape
        use_precomputed_states = (
            cache_params is not None
            and cache_params.has_previous_state
            and c_len == 1
            and cache_params.ssd_states[self.layer_idx].shape[0] == bsz
        )

        if attention_mask is not None:
            hidden_states = hidden_states * attention_mask[:, :, None]
        
        if use_precomputed_states:
            c_states = self.c_proj(hidden_states)
            b_states = self.b_proj(hidden_states)
            dt_states = self.dt_proj(hidden_states)
            x_states = self.x_proj(hidden_states)
            
            c_states = c_states.view(bsz, c_len, self.num_heads, self.head_dim)
            b_states = b_states.view(bsz, c_len, self.num_heads, self.head_dim)

            # apply rotary position embeddings
            cos, sin = position_embeddings
            c_states, b_states = apply_CB_rotary_pos_emb(c_states, b_states, cos, sin)

            # squeeze the c_len dimension
            c_states = c_states.squeeze(1)
            b_states = b_states.squeeze(1)
            dt_states = dt_states.squeeze(1)
            x_states = x_states.squeeze(1)

            # reshape the states for the selective state update
            c_states = c_states.view(bsz, self.num_heads, self.head_dim)
            b_states = b_states.view(bsz, self.num_heads, self.head_dim)
            dt_states = dt_states[:, :, None].expand(-1, -1, self.head_dim)
            a_states = self.A[:, None, ...][:, :, None].expand(-1, self.head_dim, self.head_dim)
            x_states = x_states.view(bsz, self.num_heads, self.head_dim)

            ssd_output = selective_state_update(
                cache_params.ssd_states[self.layer_idx],
                x_states,
                dt_states,
                a_states,
                b_states,
                c_states,
                z=None,
                dt_bias=None,
                dt_softplus=True,
            )
            ssd_output = ssd_output.view(bsz, self.num_heads * self.head_dim)[:, None, ...]
        else:
            c_states = self.c_proj(hidden_states)
            b_states = self.b_proj(hidden_states)
            dt_states = self.dt_proj(hidden_states)
            x_states = self.x_proj(hidden_states)

            c_states = c_states.view(bsz, c_len, self.num_heads, self.head_dim)
            b_states = b_states.view(bsz, c_len, self.num_heads, self.head_dim)
            x_states = x_states.view(bsz, c_len, self.num_heads, self.head_dim)

            # apply rotary position embeddings
            cos, sin = position_embeddings
            c_states, b_states = apply_CB_rotary_pos_emb(c_states, b_states, cos, sin)

            ssd_output, ssd_state = mamba_chunk_scan_combined(
                x_states,
                dt_states,
                self.A,
                b_states,
                c_states,
                chunk_size=self.chunk_len,
                z=None,
                seq_idx=None,
                return_final_states=True,
                dt_bias=None,
                dt_softplus=True,
            )
            if ssd_state is not None and cache_params is not None:
                cache_params.ssd_states[self.layer_idx].copy_(ssd_state)
            ssd_output = ssd_output.view(bsz, c_len, -1)
        ssd_output = self.out_proj(ssd_output)
        return ssd_output


class CheemsDynamicMaskAttention(nn.Module):
    """Dynamic Mask Attention from 'Wonderful Matrices' paper."""

    def __init__(self, config: CheemsConfig, 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.hidden_dim = config.hidden_size
        self.num_heads = config.num_attention_heads
        self.head_dim = self.hidden_dim // 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.attention_dropout = config.attention_dropout
        self.dynamic_mask_ratio = config.dynamic_mask_ratio

        # Q K V O projections
        self.q_proj = nn.Linear(self.hidden_dim, self.num_heads * self.head_dim, bias=config.hidden_bias)
        self.k_proj = nn.Linear(self.hidden_dim, self.num_key_value_heads * self.head_dim, bias=config.hidden_bias)
        self.v_proj = nn.Linear(self.hidden_dim, self.num_key_value_heads * self.head_dim, bias=config.hidden_bias)
        # dynamic mask for the QK^T attention score matrix
        self.A = nn.Parameter(torch.ones(self.num_heads))
        self.dt_proj = nn.Linear(self.num_key_value_heads * self.head_dim, self.num_heads, bias=config.hidden_bias)
        self.o_proj = nn.Linear(self.hidden_dim, self.hidden_dim, bias=config.hidden_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,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
        **kwargs,
    ) -> Tuple[torch.Tensor, Optional[Cache]]:
        bsz, q_len, _ = hidden_states.shape

        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, -1, self.head_dim).transpose(1, 2)
        key_states = key_states.view(bsz, q_len, -1, self.head_dim).transpose(1, 2)
        value_states = value_states.view(bsz, q_len, -1, self.head_dim).transpose(1, 2)

        cos, sin = position_embeddings
        query_states, key_states = apply_QK_rotary_pos_emb(query_states, key_states, cos, sin)

        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)

        # calculate dynamic mask from value_states
        dt_states = self.dt_proj(value_states.transpose(1, 2).reshape(bsz, value_states.shape[-2], -1))
        dynamic_mask = torch.exp(self.A * F.softplus(dt_states)).transpose(-1, -2)

        # repeat key and value states
        key_states = repeat_kv(key_states, self.num_key_value_groups)
        value_states = repeat_kv(value_states, self.num_key_value_groups)

        # compute attention scores matrix
        attn_weights = torch.matmul(query_states, key_states.transpose(-1, -2)) / math.sqrt(self.head_dim)

        # add mask to attention scores
        attn_mask = self.prepare_dynamic_mask(
            hidden_states=hidden_states,
            dynamic_mask=dynamic_mask,
            dynamic_mask_ratio=self.dynamic_mask_ratio,
            attention_mask=attention_mask,
        )
        attn_weights = attn_weights + attn_mask

        # upcast attention scores to fp32
        attn_weights = F.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)
        attn_weights = F.dropout(attn_weights, p=self.attention_dropout, training=self.training)

        # apply attention scores to value states
        attn_output = torch.matmul(attn_weights, value_states)

        attn_output = attn_output.transpose(1, 2).contiguous()
        attn_output = attn_output.reshape(bsz, q_len, -1)
        attn_output = self.o_proj(attn_output)

        return attn_output, past_key_value

    def prepare_dynamic_mask(
        self,
        hidden_states: torch.Tensor,
        dynamic_mask: torch.Tensor,
        dynamic_mask_ratio: float = 0.0,
        attention_mask: Optional[torch.Tensor] = None,
    ):
        """
        Combine `dynamic_mask` with `attention_mask` to generate the final `attn_mask`.

        Args:
            hidden_states (`torch.Tensor`): The input hidden_states, used to determine the minimum value of the current input precision.
            dynamic_mask (`torch.Tensor`): dynamic mask of shape `(batch_size, num_heads, key_sequence_length)`.
            dynamic_mask_ratio (`float`, *optional*): Ratio from 0.0 to 1.0 used to control the proportion of the dynamic mask filled with the minimum value.
            attention_mask (`torch.Tensor`, *optional*): attention mask of shape `(batch_size, 1, query_sequence_length, key_sequence_length)`.
        """
        min_type = torch.finfo(hidden_states.dtype).min
        attn_mask = dynamic_mask[:, :, None, :]
        if 0.0 < dynamic_mask_ratio < 1.0:
            num_dynamic_mask = int(attn_mask.shape[-1] * dynamic_mask_ratio)
            if num_dynamic_mask > 0:
                rate_value = torch.kthvalue(attn_mask, num_dynamic_mask, dim=-1, keepdim=True).values
                attn_mask = attn_mask.masked_fill(attn_mask < rate_value, min_type)
        if attention_mask is not None:
            attn_mask = attn_mask.masked_fill(attention_mask[:, :, :, : hidden_states.shape[-2]] == min_type, min_type)
        return attn_mask


class CheemsSdpaDynamicMaskAttention(CheemsDynamicMaskAttention):

    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,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
        **kwargs,
    ) -> Tuple[torch.Tensor, Optional[Cache]]:
        bsz, q_len, _ = hidden_states.shape

        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, -1, self.head_dim).transpose(1, 2)
        key_states = key_states.view(bsz, q_len, -1, self.head_dim).transpose(1, 2)
        value_states = value_states.view(bsz, q_len, -1, self.head_dim).transpose(1, 2)

        cos, sin = position_embeddings
        query_states, key_states = apply_QK_rotary_pos_emb(query_states, key_states, cos, sin)

        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)
        
        # calculate dynamic mask from value_states
        dt_states = self.dt_proj(value_states.transpose(1, 2).reshape(bsz, value_states.shape[-2], -1))
        dynamic_mask = torch.exp(self.A * F.softplus(dt_states)).transpose(-1, -2)

        attn_mask = self.prepare_dynamic_mask(
            hidden_states=hidden_states,
            dynamic_mask=dynamic_mask,
            dynamic_mask_ratio=self.dynamic_mask_ratio,
            attention_mask=attention_mask,
        )

        query_states = query_states.contiguous()
        key_states = key_states.contiguous()
        value_states = value_states.contiguous()

        # NOTE: As of pytorch 2.5.1, cuDNN's SDPA backward pass is still incorrect, so we disable cuDNN SDPA (see https://github.com/pytorch/pytorch/issues/138581)
        torch.backends.cuda.enable_cudnn_sdp(False)
        attn_output = F.scaled_dot_product_attention(
            query_states,
            key_states,
            value_states,
            attn_mask=attn_mask,
            dropout_p=self.attention_dropout if self.training else 0.0,
            enable_gqa=True,
        )

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

        return attn_output, past_key_value


class CheemsFlexDynamicMaskAttention(CheemsDynamicMaskAttention):

    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,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
        **kwargs,
    ) -> Tuple[torch.Tensor, Optional[Cache]]:
        bsz, q_len, _ = hidden_states.shape

        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, -1, self.head_dim).transpose(1, 2)
        key_states = key_states.view(bsz, q_len, -1, self.head_dim).transpose(1, 2)
        value_states = value_states.view(bsz, q_len, -1, self.head_dim).transpose(1, 2)

        cos, sin = position_embeddings
        query_states, key_states = apply_QK_rotary_pos_emb(query_states, key_states, cos, sin)

        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)
        
        dt_states = self.dt_proj(value_states.transpose(1, 2).reshape(bsz, value_states.shape[-2], -1))
        dynamic_mask = torch.exp(self.A * F.softplus(dt_states)).transpose(-1, -2)

        attn_mask = self.prepare_dynamic_mask(
            hidden_states=hidden_states,
            dynamic_mask=dynamic_mask,
            dynamic_mask_ratio=self.dynamic_mask_ratio,
            attention_mask=attention_mask,
        )
        # TODO: flex_attention: Captured buffers that require grad are not yet supported.
        # NOTE: So we only use flex_attention in inference mode.
        def dynamic_mask_mod(score, batch, head, q_idx, kv_idx):
            score = score + attn_mask[batch][head][q_idx][kv_idx]
            return score

        attn_output = flex_attention(
            query_states,
            key_states,
            value_states,
            score_mod=dynamic_mask_mod,
            enable_gqa=True,
        )

        attn_output = attn_output.transpose(1, 2).contiguous()
        attn_output = attn_output.view(bsz, q_len, -1)
        attn_output = self.o_proj(attn_output)
        
        return attn_output, past_key_value


CHEEMS_ATTENTION_CLASSES = {
    "flex_attention": CheemsFlexDynamicMaskAttention,
    "eager": CheemsDynamicMaskAttention,
    "sdpa": CheemsSdpaDynamicMaskAttention,
}


class CheemsMLP(nn.Module):

    def __init__(self, config: CheemsConfig):
        super().__init__()
        self.hidden_dim = config.hidden_size
        self.intermediate_dim = config.intermediate_size
        self.act_fn = ACT2FN[config.hidden_act]

        self.gate_proj = nn.Linear(self.hidden_dim, self.intermediate_dim, bias=config.hidden_bias)
        self.up_proj = nn.Linear(self.hidden_dim, self.intermediate_dim, bias=config.hidden_bias)
        self.down_proj = nn.Linear(self.intermediate_dim, self.hidden_dim, bias=config.hidden_bias)

    def forward(
        self,
        hidden_states: torch.Tensor,
        **kwargs,
    ) -> torch.Tensor:
        hidden_states = self.down_proj(self.act_fn(self.gate_proj(hidden_states)) * self.up_proj(hidden_states))
        return hidden_states


class CheemsCDMoE(CheemsMLP):
    """Cross Domain Mixture of Experts from 'Wonderful Matrices' paper."""

    def __init__(self, config: CheemsConfig):
        super().__init__(config)
        self.hidden_dim = config.hidden_size
        self.act_fn = ACT2FN[config.hidden_act]

        self.expert_retrieval_dim = config.expert_retrieval_size
        self.num_cdmoe_experts = config.num_cdmoe_experts
        self.num_cdmoe_heads = config.num_cdmoe_heads
        self.num_cdmoe_experts_per_head = config.num_cdmoe_experts_per_head
        self.num_keys = int(math.sqrt(self.num_cdmoe_experts))

        # queries and keys for retrieval experts
        self.queries = nn.Linear(self.hidden_dim, self.num_cdmoe_heads * self.expert_retrieval_dim, bias=False)
        self.keys = nn.Parameter(torch.zeros(self.num_cdmoe_heads, self.num_keys, 2, self.expert_retrieval_dim // 2))

        # experts
        self.down_embed  = nn.Embedding(self.num_cdmoe_experts, self.hidden_dim)
        self.up_embed = nn.Embedding(self.num_cdmoe_experts, self.hidden_dim)

    def forward(
        self,
        hidden_states: torch.Tensor,
        **kwargs,
    ) -> torch.Tensor:
        bsz, seq_len, _ = hidden_states.shape

        # get similarity with queries and keys
        queries = self.queries(hidden_states)
        queries = queries.view(bsz, seq_len, 2, self.num_cdmoe_heads, -1).permute(2, 0, 1, 3, 4)
        sim = torch.einsum("p b t h n, h k p n -> p b t h k", queries, self.keys)

        # get experts with the highest similarity
        (scores_x, scores_y), (indices_x, indices_y) = sim.topk(self.num_cdmoe_experts_per_head, dim=-1)
        if einx_add is not None:
            all_scores = einx_add("... i, ... j -> ... (i j)", scores_x, scores_y)
            all_indices = einx_add("... i, ... j -> ... (i j)", indices_x * self.num_keys, indices_y)
        else:
            all_scores = scores_x.unsqueeze(-1) + scores_y.unsqueeze(-2)
            all_scores = all_scores.view(*scores_x.shape[:-1], -1)
            all_indices = (indices_x.unsqueeze(-1) * self.num_keys) + indices_y.unsqueeze(-2)
            all_indices = all_indices.view(*indices_x.shape[:-1], -1)
        scores, pk_indices = all_scores.topk(self.num_cdmoe_experts_per_head, dim=-1)
        indices = all_indices.gather(-1, pk_indices)
        down_embed = self.down_embed(indices)
        up_embed = self.up_embed(indices)

        # mix experts states with cross domain states
        experts_weights = torch.einsum("b t d, b t h k d -> b t h k", hidden_states, down_embed)
        experts_weights = self.act_fn(experts_weights) * scores.softmax(dim=-1)
        experts_states = torch.einsum("b t h k, b t h k d -> b t d", experts_weights, up_embed)
        hidden_states = self.down_proj(self.act_fn(self.gate_proj(hidden_states)) * self.up_proj(hidden_states))
        hidden_states = hidden_states + experts_states
        return hidden_states


class CheemsSSDDecoderLayer(nn.Module):
    def __init__(self, config: CheemsConfig, layer_idx: Optional[int] = None):
        super().__init__()
        self.hidden_dropout = config.hidden_dropout

        self.pre_sequence_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.ssd = CheemsStateSpaceDuality(config, layer_idx=layer_idx)
        self.post_sequence_residual = Residual(config.hidden_size)

        self.pre_state_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.feed_forward = CheemsMLP(config) if config.is_moe == False else CheemsCDMoE(config)
        self.post_state_residual = Residual(config.hidden_size)

    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: Optional[bool] = False,
        use_cache: Optional[bool] = False,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
        **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
            cache_position (`torch.LongTensor` of shape `(sequence_length)`, *optional*):
                Indices depicting the position of the input sequence tokens in the sequence
            position_embeddings (`Tuple[torch.FloatTensor, torch.FloatTensor]`, *optional*):
                Tuple containing the cosine and sine positional embeddings of shape `(batch_size, seq_len, head_dim)`,
                with `head_dim` being the embedding dimension of each attention head.
            kwargs (`dict`, *optional*):
                Arbitrary kwargs to be ignored, used for FSDP and other methods that injects code
                into the model
        """

        # sequence transformation
        residual = hidden_states
        hidden_states = self.pre_sequence_layernorm(hidden_states)
        hidden_states = self.ssd(
            hidden_states,
            attention_mask,
            position_ids,
            past_key_value,
            cache_position,
            position_embeddings,
            **kwargs,
        )
        self_attn_weights = None
        hidden_states = F.dropout(hidden_states, p=self.hidden_dropout, training=self.training)
        hidden_states = self.post_sequence_residual(residual, hidden_states)

        # state transformation
        residual = hidden_states
        hidden_states = self.pre_state_layernorm(hidden_states)
        hidden_states = self.feed_forward(hidden_states)
        hidden_states = F.dropout(hidden_states, p=self.hidden_dropout, training=self.training)
        hidden_states = self.post_state_residual(residual, hidden_states)

        outputs = (hidden_states,)

        if output_attentions:
            outputs += (self_attn_weights,)

        if use_cache:
            outputs += (past_key_value,)

        return outputs


class CheemsAttnDecoderLayer(nn.Module):
    def __init__(self, config: CheemsConfig, layer_idx: Optional[int] = None):
        super().__init__()
        self.hidden_dropout = config.hidden_dropout

        self.pre_sequence_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.attn = CHEEMS_ATTENTION_CLASSES[config._attn_implementation](config=config, layer_idx=layer_idx)
        self.post_sequence_residual = Residual(config.hidden_size)

        self.pre_state_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.feed_forward = CheemsMLP(config) if config.is_moe == False else CheemsCDMoE(config)
        self.post_state_residual = Residual(config.hidden_size)

    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: Optional[bool] = False,
        use_cache: Optional[bool] = False,
        cache_position: Optional[torch.LongTensor] = None,
        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
        **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
            cache_position (`torch.LongTensor` of shape `(sequence_length)`, *optional*):
                Indices depicting the position of the input sequence tokens in the sequence
            position_embeddings (`Tuple[torch.FloatTensor, torch.FloatTensor]`, *optional*):
                Tuple containing the cosine and sine positional embeddings of shape `(batch_size, seq_len, head_dim)`,
                with `head_dim` being the embedding dimension of each attention head.
            kwargs (`dict`, *optional*):
                Arbitrary kwargs to be ignored, used for FSDP and other methods that injects code
                into the model
        """

        # sequence transformation
        residual = hidden_states
        hidden_states = self.pre_sequence_layernorm(hidden_states)
        hidden_states, present_key_value = self.attn(
            hidden_states,
            attention_mask,
            position_ids,
            past_key_value,
            cache_position,
            position_embeddings,
            **kwargs,
        )
        self_attn_weights = None
        hidden_states = F.dropout(hidden_states, p=self.hidden_dropout, training=self.training)
        hidden_states = self.post_sequence_residual(residual, hidden_states)

        # state transformation
        residual = hidden_states
        hidden_states = self.pre_state_layernorm(hidden_states)
        hidden_states = self.feed_forward(hidden_states)
        hidden_states = F.dropout(hidden_states, p=self.hidden_dropout, training=self.training)
        hidden_states = self.post_state_residual(residual, hidden_states)

        outputs = (hidden_states,)

        if output_attentions:
            outputs += (self_attn_weights,)

        if use_cache:
            outputs += (present_key_value,)

        return outputs


@add_start_docstrings("The bare Cheems Model outputting raw hidden-states without any specific head on top.")
class CheemsPreTrainedModel(PreTrainedModel):
    config_class = CheemsConfig
    base_model_prefix = "model"
    supports_gradient_checkpointing = True
    _no_split_modules = ["CheemsSSDDecoderLayer", "CheemsAttnDecoderLayer"]
    _skip_keys_device_placement = ["past_key_values"]
    _supports_flex_attn = True
    _supports_sdpa = True
    _supports_cache_class = True
    _supports_quantized_cache = True
    _supports_static_cache = 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_()


CHEEMS_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, see our
            [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache);
            - 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.
"""


ALL_DECODER_LAYERS_TYPE = {
    "ssd": CheemsSSDDecoderLayer,
    "attn": CheemsAttnDecoderLayer,
}


@add_start_docstrings("The bare Cheems Model outputting raw hidden-states without any specific head on top.")
class CheemsModel(CheemsPreTrainedModel):
    def __init__(self, config: CheemsConfig):
        super().__init__(config)
        self.config = config
        self.padding_idx = config.pad_token_id
        self.vocab_size = config.vocab_size

        self.word_embed = nn.Embedding(config.vocab_size, config.hidden_size, padding_idx=config.pad_token_id)
        self.rotary_emb = RotaryEmbedding(config)

        decoder_layers = []
        for i in range(config.num_hidden_layers):
            layer_class = ALL_DECODER_LAYERS_TYPE[config.layers_type[i]]
            decoder_layers.append(layer_class(config, layer_idx=i))
        self.layers = nn.ModuleList(decoder_layers)
        
        self.final_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.gradient_checkpointing = False

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

    def get_input_embeddings(self):
        return self.word_embed

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

    @add_start_docstrings_to_model_forward(CHEEMS_INPUTS_DOCSTRING)
    def forward(
        self,
        input_ids: torch.LongTensor = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[HybridSSDAttnDynamicCache] = 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,
        cache_position: Optional[torch.LongTensor] = None,
    ) -> 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 (input_ids is None) ^ (inputs_embeds is not None):
            raise ValueError("You cannot specify both input_ids and inputs_embeds")

        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.word_embed(input_ids)
        hidden_states = inputs_embeds

        if use_cache and past_key_values is None:
            logger.warning_once(
                "Cheems requires an initialized `HybridSSDAttnDynamicCache` to return a cache. None was provided, so no cache will be returned."
            )

        if cache_position is None:
            cache_position = torch.arange(
                hidden_states.shape[1],
                device=hidden_states.device,
            )
        if position_ids is None:
            position_ids = cache_position.unsqueeze(0)

        attn_mask = self._update_attn_mask(attention_mask, hidden_states, cache_position)
        ssd_mask = self._update_ssd_mask(attention_mask, cache_position)
    
        # create position embeddings to be shared across the decoder layers
        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

        for decoder_layer in self.layers[: self.config.num_hidden_layers]:
            layer_mask = ssd_mask if isinstance(decoder_layer, CheemsSSDDecoderLayer) else attn_mask

            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,
                    layer_mask,
                    position_ids,
                    past_key_values,
                    output_attentions,
                    use_cache,
                    cache_position,
                    position_embeddings,
                )
            else:
                layer_outputs = decoder_layer(
                    hidden_states,
                    attention_mask=layer_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 output_attentions:
                all_self_attns += (layer_outputs[1],)

        hidden_states = self.final_layernorm(hidden_states)

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

        if past_key_values and not past_key_values.has_previous_state:
            past_key_values.has_previous_state = True

        next_cache = None if not use_cache else past_key_values

        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,
        )

    def _update_attn_mask(
        self,
        attention_mask: torch.Tensor = None,
        input_tensor: torch.Tensor = None,
        cache_position: torch.Tensor = None,
    ):

        dtype, device = input_tensor.dtype, input_tensor.device
        min_dtype = torch.finfo(dtype).min
        sequence_length = input_tensor.shape[1]
        target_length = 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()
            mask_length = attention_mask.shape[-1]
            padding_mask = causal_mask[:, :, :, :mask_length] + attention_mask[:, None, None, :]
            padding_mask = padding_mask == 0
            causal_mask[:, :, :, :mask_length] = causal_mask[:, :, :, :mask_length].masked_fill(
                padding_mask, min_dtype
            )

        return causal_mask

    def _update_ssd_mask(
        self,
        attention_mask: torch.Tensor,
        cache_position: torch.Tensor,
    ):
        """
        No need for zeroing states when
            1. Cached forward
            2. Attending to all inputs
        """
        ssd_mask = attention_mask
        if cache_position[0] > 0 or (attention_mask is not None and torch.all(attention_mask == 1)):
            ssd_mask = None
        return ssd_mask


class CheemsForCausalLM(CheemsPreTrainedModel, GenerationMixin):
    _tied_weights_keys = ["lm_head.weight"]

    def __init__(self, config: CheemsConfig):
        super().__init__(config)
        self.config = config
        self.model = CheemsModel(config)
        self.vocab_size = config.vocab_size
        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)

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

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

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

    @add_start_docstrings_to_model_forward(CHEEMS_INPUTS_DOCSTRING)
    @replace_return_docstrings(output_type=CausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC)
    def forward(
        self,
        input_ids: torch.LongTensor = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[HybridSSDAttnDynamicCache] = 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,
        cache_position: Optional[torch.LongTensor] = None,
        num_logits_to_keep: int = 0,
        **loss_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]`.

            num_logits_to_keep (`int`, *optional*):
                Calculate logits for the last `num_logits_to_keep` tokens. If `0`, calculate logits for all
                `input_ids` (special case). Only last token logits are needed for generation, and calculating them only for that
                token can save memory, which becomes pretty significant for long sequences or large vocabulary size.

        Returns:
        """
        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 output consists of (dec_features, layer_state, dec_hidden, dec_attn)
        outputs = self.model(
            input_ids=input_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,
        )

        hidden_states = outputs[0]

        # only compute necessary logits, and do not upcast them to float if we are not computing the loss
        logits = self.lm_head(hidden_states[:, -num_logits_to_keep:, :])

        loss = None
        if labels is not None:
            loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.vocab_size, **loss_kwargs)

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

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


@add_start_docstrings(
    """
    The Doge Model transformer with a sequence classification head on top (linear layer).

    [`DogeForSequenceClassification`] uses the last token in order to do the classification, as other causal models
    (e.g. GPT-2) do.

    Since it does classification on the last token, it requires to know the position of the last token. If a
    `pad_token_id` is defined in the configuration, it finds the last token that is not a padding token in each row. If
    no `pad_token_id` is defined, it simply takes the last value in each row of the batch. Since it cannot guess the
    padding tokens when `inputs_embeds` are passed instead of `input_ids`, it does the same (take the last value in
    each row of the batch).
    """
)
class CheemsForSequenceClassification(CheemsPreTrainedModel):
    def __init__(self, config: CheemsConfig):
        super().__init__(config)
        self.config = config
        self.num_labels = config.num_labels

        self.model = CheemsModel(config)
        self.classifier = nn.Linear(config.hidden_size, self.num_labels, bias=False)

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

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

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

    @add_start_docstrings_to_model_forward(CHEEMS_INPUTS_DOCSTRING)
    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[Union[Cache, 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,
    ) -> Union[Tuple, SequenceClassifierOutputWithPast]:
        r"""
        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
        """
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict

        outputs = self.model(
            input_ids=input_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,
        )
        hidden_states = outputs[0]
        logits = self.classifier(hidden_states)

        if input_ids is not None:
            batch_size = input_ids.shape[0]
        else:
            batch_size = inputs_embeds.shape[0]

        if self.config.pad_token_id is None and batch_size != 1:
            raise ValueError("Cannot handle batch sizes > 1 if no padding token is defined.")
        if self.config.pad_token_id is None:
            sequence_lengths = -1
        else:
            if input_ids is not None:
                # if no pad token found, use modulo instead of reverse indexing for ONNX compatibility
                sequence_lengths = torch.eq(input_ids, self.config.pad_token_id).int().argmax(-1) - 1
                sequence_lengths = sequence_lengths % input_ids.shape[-1]
                sequence_lengths = sequence_lengths.to(logits.device)
            else:
                sequence_lengths = -1

        pooled_logits = logits[torch.arange(batch_size, device=logits.device), sequence_lengths]

        loss = None
        if labels is not None:
            loss = self.loss_function(
                logits=logits,
                labels=labels,
                pooled_logits=pooled_logits,
                config=self.config,
            )

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

        return SequenceClassifierOutputWithPast(
            loss=loss,
            logits=pooled_logits,
            past_key_values=outputs.past_key_values,
            hidden_states=outputs.hidden_states,
            attentions=outputs.attentions,
        )