modelling_mistral_vib.py

# coding=utf-8
# Copyright 2023 Mistral AI and the HuggingFace Inc. team. All rights reserved.
#
# This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
# and OPT implementations in this library. It has been modified from its
# original forms to accommodate minor architectural differences compared
# to GPT-NeoX and OPT used by the Meta AI team that trained the model.
#
# 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 Mistral model.
	Modified in TVA_prune to incorporate VIB-based masks to prune weights
"""
import inspect
import math
import time
import warnings
from typing import List, Optional, Tuple, Union
import gc
import torch
import torch.nn.functional as F
import torch.utils.checkpoint
from torch import nn
from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss

from transformers.activations import ACT2FN
from utils.cache_utils import Cache, DynamicCache
from utils.modeling_attn_mask_utils import _prepare_4d_causal_attention_mask, _prepare_4d_causal_attention_mask_for_sdpa
from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast, SequenceClassifierOutputWithPast
from transformers.modeling_utils import PreTrainedModel,prune_linear_layer
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 transformers.models.mistral.configuration_mistral import MistralConfig
from vib_layer_lay import InformationBottleneck

# 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

# 	_flash_supports_window_size = "window_size" in list(inspect.signature(flash_attn_func).parameters)
logger = logging.get_logger(__name__)
_CONFIG_FOR_DOC = "MistralConfig"

# Copied from transformers.models.llama.modeling_llama._get_unpad_data
# 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,
# 	)


# Copied from transformers.models.llama.modeling_llama.LlamaRMSNorm with Llama->Mistral
class MistralRMSNorm(nn.Module):
	def __init__(self, hidden_size, eps=1e-6):
		"""
		MistralRMSNorm is equivalent to T5LayerNorm
		"""
		super().__init__()
		self.weight = nn.Parameter(torch.ones(hidden_size))
		self.variance_epsilon = eps

	def forward(self, hidden_states,hidden_z=None): #adapted from llama-vib
		input_dtype = hidden_states.dtype
		if hidden_z is not None:
			remaining_index = torch.where(~hidden_z.eq(0))[0]
			compressed_hidden_states = torch.index_select( hidden_states, dim=-1, index=remaining_index)
			compressed_weight = self.weight[remaining_index]
			normalized_shape = len(remaining_index)
			variance = compressed_hidden_states.to(torch.float32).pow(2).mean(-1, keepdim=True)
			compressed_hidden_states = compressed_hidden_states * torch.rsqrt(variance + self.variance_epsilon)
			if compressed_weight.dtype in [torch.float16, torch.bfloat16]:
				compressed_hidden_states = compressed_hidden_states.to(compressed_weight.dtype)
			normed_states= compressed_weight * compressed_hidden_states
			output = hidden_states.clone()
			output[:, :, remaining_index] = normed_states
			return output
		else:
			variance = hidden_states.to(torch.float32).pow(2).mean(-1, keepdim=True)
			hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
			# convert into half-precision if necessary
			if self.weight.dtype in [torch.float16, torch.bfloat16]:
				hidden_states = hidden_states.to(self.weight.dtype)			
		return self.weight * hidden_states.to(input_dtype) 

# copied from transformers.models.llama.modeling_llama.LlamaRotaryEmbedding with Llama->Mistral
# TODO @Arthur no longer copied from LLama after static cache
class MistralRotaryEmbedding(nn.Module):
	def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):
		super().__init__()
		self.dim = dim
		self.max_position_embeddings = max_position_embeddings
		self.base = base
		inv_freq = 1.0 / (self.base ** (torch.arange(0, self.dim, 2, dtype=torch.int64).float().to(device) / self.dim))
		self.register_buffer("inv_freq", inv_freq, persistent=False)

		# Build here to make `torch.jit.trace` work.
		self._set_cos_sin_cache(
			seq_len=max_position_embeddings, device=self.inv_freq.device, dtype=torch.get_default_dtype()
		)

	def _set_cos_sin_cache(self, seq_len, device, dtype):
		self.max_seq_len_cached = seq_len
		t = torch.arange(self.max_seq_len_cached, device=device, dtype=torch.int64).type_as(self.inv_freq)

		freqs = torch.einsum("i,j->ij", t, self.inv_freq) #freqs = torch.outer(t, self.inv_freq)
		# Different from paper, but it uses a different permutation in order to obtain the same calculation
		emb = torch.cat((freqs, freqs), dim=-1)
		self.register_buffer("cos_cached", emb.cos()[None, None, :, :], persistent=False)
		self.register_buffer("sin_cached", emb.sin()[None, None, :, :], persistent=False)
		
	def forward(self, x, seq_len=None):
		if seq_len > self.max_seq_len_cached:
			self.max_seq_len_cached = seq_len
			t = torch.arange(self.max_seq_len_cached, device=x.device, dtype=self.inv_freq.dtype)
			freqs = torch.einsum("i,j->ij", t, self.inv_freq)
			# Different from paper, but it uses a different permutation in order to obtain the same calculation
			emb = torch.cat((freqs, freqs), dim=-1).to(x.device)
			self.register_buffer("cos_cached", emb.cos()[None, None, :, :], persistent=False)
			self.register_buffer("sin_cached", emb.sin()[None, None, :, :], persistent=False)
		
		return (
			self.cos_cached[:, :, :seq_len, ...].to(dtype=x.dtype),
			self.sin_cached[:, :, :seq_len, ...].to(dtype=x.dtype),
		)


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


def apply_rotary_pos_emb(q, k, cos, sin, position_ids):
	gather_indices = position_ids[:, None, :, None]  # [bs, 1, seq_len, 1]
	gather_indices = gather_indices.repeat(1, cos.shape[1], 1, cos.shape[3])
	cos = torch.gather(cos.repeat(gather_indices.shape[0], 1, 1, 1), 2, gather_indices)
	sin = torch.gather(sin.repeat(gather_indices.shape[0], 1, 1, 1), 2, gather_indices)
	q_embed = (q * cos) + (rotate_half(q) * sin)
	k_embed = (k * cos) + (rotate_half(k) * sin)
	#del q,k, cos, sin, position_ids
	return q_embed, k_embed
class MistralMLP(nn.Module):
	def __init__(self, config):
		super().__init__()
		self.config = config
		self.hidden_size = config.hidden_size
		self.intermediate_size = config.intermediate_size
		kl_mult= self.intermediate_size//self.hidden_size
		if config.vib_layers==True:
			self.gate_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.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
			self.ib_1= InformationBottleneck(self.intermediate_size,kl_mult=config.inter_mul)
			
		else:
			self.gate_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.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
		self.act_fn = ACT2FN[config.hidden_act]

	def get_num_rem_weights(self,hidden_mask_size):
		mask_1 = self.ib_1.get_mask_hard()
		num_states_rem_1 = torch.sum((mask_1 == 1).int())
		rem= ((2*num_states_rem_1 * hidden_mask_size) + (hidden_mask_size * num_states_rem_1)) #no bias
		return rem	

	def get_sparsity(self,hidden_mask_sparse):
		att_sp= (3*(self.ib_1.sparse/1e3) * (hidden_mask_sparse/1e3))         
		return att_sp

	def get_kld_loss(self,hidden_mask,kl_fac):
		kl=((self.ib_1.kld*kl_fac) + (hidden_mask.kld*kl_fac))        
		return kl

	def forward(self, x,hidden_mask):
		intermed_result = self.act_fn(self.gate_proj(x)) * self.up_proj(x)     
		bsz,seq,dim = intermed_result.size()   
		if self.config.vib_layers==True: 
			intermed_result= intermed_result.reshape(bsz*seq,dim)
			intermed_result = self.ib_1(intermed_result).to(intermed_result.dtype)
			intermed_result= intermed_result.reshape(bsz,seq,dim)			
		
		intermed_result=self.down_proj(intermed_result)
		dim =intermed_result.size(2)
		if self.config.vib_layers==True: #only during finetuning
			intermed_result= intermed_result.reshape(bsz*seq,dim)
			intermed_result = hidden_mask(intermed_result).to(intermed_result.dtype)
			intermed_result= intermed_result.reshape(bsz,seq,dim)			
		return intermed_result


# Copied from transformers.models.llama.modeling_llama.repeat_kv
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 MistralAttention(nn.Module):
	"""
	Multi-headed attention from 'Attention Is All You Need' paper. Modified to use sliding window attention: Longformer
	and "Generating Long Sequences with Sparse Transformers".
	"""

	def __init__(self, config: MistralConfig, 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_size = config.hidden_size
		self.num_heads = config.num_attention_heads
		self.head_dim = self.hidden_size // self.num_heads if self.config.vib_layers == True else 128 #for 7b model
		self.num_key_value_heads = config.num_key_value_heads
		self.num_key_value_groups = self.num_heads // self.num_key_value_heads 
		self.max_position_embeddings = config.max_position_embeddings
		self.rope_theta = config.rope_theta
		self.is_causal = True
		self.attention_dropout = config.attention_dropout

		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})."
			)
		if self.config.vib_layers == True:
			self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=False)
			self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
			self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
			self.o_proj = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=False)
			#self.ib_1= InformationBottleneck(self.num_heads,kl_mult=self.head_dim)
			self.ib_1= InformationBottleneck(self.num_key_value_heads,kl_mult=config.att_mul) #(self.head_dim* self.num_key_value_groups)//config.att_mul)
			
		else:
			self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=False)
			self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
			self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
			self.o_proj = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=False)
		
		self.rotary_emb = MistralRotaryEmbedding(
			self.head_dim,
			max_position_embeddings=self.max_position_embeddings,
			base=self.rope_theta,
		)

	def _shape(self, tensor: torch.Tensor, seq_len: int, bsz: int):
		return tensor.view(bsz, seq_len, self.num_heads, self.head_dim).transpose(1, 2).contiguous()

	def get_num_rem_weights(self,hidden_mask_size):
		#mask_1= self.ib_1.get_mask_hard()
		mask_2 = self.ib_1.get_mask_hard().int()
		#num_states_rem_1 =torch.sum((mask_1 == 1).int()) #if query, key and value are all present only then value exists
		num_states_rem_2 =torch.sum((mask_2 == 1))
		return ((self.head_dim*num_states_rem_2*2 * hidden_mask_size) *  (1+ self.num_key_value_groups)) #no bias, (key,value +  q,out)

	def get_sparsity(self, hidden_mask_sparse):
		att_sp= (self.head_dim * (self.ib_1.sparse/1e3) * 2 * (hidden_mask_sparse/1e3)) * (1+ self.num_key_value_groups)#no bias, 
		return att_sp #no bias

	def get_kld_loss(self,hidden_mask,kl_fac):
		#print("\n losses=",self.ib_1.kld.item() , hidden_mask.kld.item() )
		kl= (self.ib_1.kld*kl_fac) + (hidden_mask.kld*kl_fac)
		return kl
	
	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,
		hidden_mask=None,
		**kwargs,
	) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
		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.`"
			)
		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) #bsz, self.num_key_value_heads,q_len, self.head_dim
		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)

		kv_seq_len = key_states.shape[-2]
		if past_key_value is not None:
			if self.layer_idx is None:
				raise ValueError(
					f"The cache structure has changed since version v4.36. If you are using {self.__class__.__name__} "
					"for auto-regressive decoding with k/v caching, please make sure to initialize the attention class "
					"with a layer index."
				)
			kv_seq_len += past_key_value.get_usable_length(kv_seq_len, self.layer_idx)

		cos, sin = self.rotary_emb(value_states, seq_len=kv_seq_len)
		query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)

		if past_key_value is not None:
			key_states = torch.cat([past_key_value[0], key_states], dim=2)
			value_states = torch.cat([past_key_value[1], value_states], dim=2)

		# repeat k/v heads if n_kv_heads < n_heads
		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) #bsz, self.num_heads, q_len, kv_seq_len
		
		if attn_weights.size() != (bsz, self.num_heads, q_len, kv_seq_len):
			raise ValueError(
				f"Attention weights should be of size {(bsz, self.num_heads, q_len, kv_seq_len)}, but is"
				f" {attn_weights.size()}"
			)

		if attention_mask is not None:
			if attention_mask.size() != (bsz, 1, q_len, kv_seq_len):
				raise ValueError(
					f"Attention mask should be of size {(bsz, 1, q_len, kv_seq_len)}, but is {attention_mask.size()}"
				)

			attn_weights = attn_weights + attention_mask
			#attn_weights = torch.max(attn_weights, torch.tensor(torch.finfo(attn_weights.dtype).min))

		# 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) #bsz, self.num_heads, q_len, self.head_dim
		if self.config.vib_layers==True:
			attn_output = attn_output.permute(0,2,3,1).reshape(bsz* q_len* self.head_dim, self.num_key_value_groups,self.num_key_value_heads).reshape(bsz* q_len* self.head_dim*self.num_key_value_groups,-1) #bsz,q_len ,self.num_key_value_heads, self.num_key_value_groups,self.head_dim
			attn_output= self.ib_1(attn_output).to(attn_output.dtype)
			attn_output = attn_output.reshape(bsz,q_len,self.head_dim,self.num_key_value_heads* self.num_key_value_groups).permute(0,3,1,2) #bsz,self.num_heads,q_len, self.head_dim
		
		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)
		attn_output = attn_output.reshape(bsz, q_len, self.num_heads*self.head_dim)
		attn_output = self.o_proj(attn_output)
		
		if self.config.vib_layers==True: 
			attn_output= attn_output.reshape(bsz* q_len,self.hidden_size)
			attn_output = hidden_mask(attn_output).to(attn_output.dtype)
			attn_output= attn_output.reshape(bsz, q_len, self.hidden_size)
						
		if not output_attentions:
			attn_weights = None
		
		return attn_output, attn_weights, past_key_value


# class MistralFlashAttention2(MistralAttention):
# 	"""
# 	Mistral flash attention module. This module inherits from `MistralAttention` 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.
# 	"""

# 	# Copied from transformers.models.llama.modeling_llama.LlamaFlashAttention2.__init__
# 	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.Tensor] = None,
# 		position_ids: Optional[torch.LongTensor] = None,
# 		past_key_value: Optional[Cache] = None,
# 		output_attentions: bool = False,
# 		use_cache: bool = False,
# 		**kwargs,
# 	):
# 		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.`"
# 			)

# 			# overwrite attention_mask with padding_mask
# 			attention_mask = kwargs.pop("padding_mask")
# 		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)

# 		kv_seq_len = key_states.shape[-2]
# 		if past_key_value is not None:
# 			if self.layer_idx is None:
# 				raise ValueError(
# 					f"The cache structure has changed since version v4.36. If you are using {self.__class__.__name__} "
# 					"for auto-regressive decoding with k/v caching, please make sure to initialize the attention class "
# 					"with a layer index."
# 				)
# 			kv_seq_len += past_key_value.get_usable_length(kv_seq_len, self.layer_idx)

# 		# Because the input can be padded, the absolute sequence length depends on the max position id.
# 		rotary_seq_len = max(kv_seq_len, position_ids[:, -1].max().item()) + 1
# 		cos, sin = self.rotary_emb(value_states, seq_len=rotary_seq_len)

# 		query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)

# 		use_sliding_windows = (
# 			_flash_supports_window_size
# 			and getattr(self.config, "sliding_window", None) is not None
# 			and kv_seq_len > self.config.sliding_window
# 		)

# 		if not _flash_supports_window_size:
# 			logger.warning_once(
# 				"The current flash attention version does not support sliding window attention, for a more memory efficient implementation"
# 				" make sure to upgrade flash-attn library."
# 			)

# 		if past_key_value is not None:
# 			# Activate slicing cache only if the config has a value `sliding_windows` attribute
# 			cache_has_contents = past_key_value.get_seq_length(self.layer_idx) > 0
# 			if (
# 				getattr(self.config, "sliding_window", None) is not None
# 				and kv_seq_len > self.config.sliding_window
# 				and cache_has_contents
# 			):
# 				slicing_tokens = 1 - self.config.sliding_window

# 				past_key = past_key_value[self.layer_idx][0]
# 				past_value = past_key_value[self.layer_idx][1]

# 				past_key = past_key[:, :, slicing_tokens:, :].contiguous()
# 				past_value = past_value[:, :, slicing_tokens:, :].contiguous()

# 				if past_key.shape[-2] != self.config.sliding_window - 1:
# 					raise ValueError(
# 						f"past key must have a shape of (`batch_size, num_heads, self.config.sliding_window-1, head_dim`), got"
# 						f" {past_key.shape}"
# 					)

# 				if attention_mask is not None:
# 					attention_mask = attention_mask[:, slicing_tokens:]
# 					attention_mask = torch.cat([attention_mask, torch.ones_like(attention_mask[:, -1:])], dim=-1)

# 			cache_kwargs = {"sin": sin, "cos": cos}  # Specific to RoPE models
# 			key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)

# 		# repeat k/v heads if n_kv_heads < n_heads
# 		key_states = repeat_kv(key_states, self.num_key_value_groups)
# 		value_states = repeat_kv(value_states, self.num_key_value_groups)
# 		dropout_rate = 0.0 if not self.training else self.attention_dropout

# 		# 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 float16 just to be sure everything works as expected.
# 		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)

# 		# Reashape to the expected shape for Flash Attention
# 		query_states = query_states.transpose(1, 2)
# 		key_states = key_states.transpose(1, 2)
# 		value_states = value_states.transpose(1, 2)

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

# 		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,
# 		use_sliding_windows=False,
# 	):
# 		"""
# 		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)
# 			use_sliding_windows (`bool`, *optional*):
# 				Whether to activate sliding window attention.
# 		"""
# 		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 LlamaFlashAttention2 __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

# 			if not use_sliding_windows:
# 				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,
# 				)
# 			else:
# 				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,
# 					window_size=(self.config.sliding_window, self.config.sliding_window),
# 				)

# 			attn_output = pad_input(attn_output_unpad, indices_q, batch_size, query_length)
# 		else:
# 			if not use_sliding_windows:
# 				attn_output = flash_attn_func(
# 					query_states,
# 					key_states,
# 					value_states,
# 					dropout,
# 					softmax_scale=softmax_scale,
# 					causal=causal,
# 				)
# 			else:
# 				attn_output = flash_attn_func(
# 					query_states,
# 					key_states,
# 					value_states,
# 					dropout,
# 					softmax_scale=softmax_scale,
# 					causal=causal,
# 					window_size=(self.config.sliding_window, self.config.sliding_window),
# 				)

# 		return attn_output

# 	def _upad_input(self, query_layer, key_layer, value_layer, attention_mask, query_length):
# 		batch_size, kv_seq_len, num_heads, head_dim = key_layer.shape

# 		# On the first iteration we need to properly re-create the padding mask
# 		# by slicing it on the proper place
# 		if kv_seq_len != attention_mask.shape[-1]:
# 			attention_mask_num_tokens = attention_mask.shape[-1]
# 			attention_mask = attention_mask[:, attention_mask_num_tokens - kv_seq_len :]

# 		indices_k, cu_seqlens_k, max_seqlen_in_batch_k = _get_unpad_data(attention_mask)

# 		key_layer = index_first_axis(key_layer.reshape(batch_size * kv_seq_len, num_heads, head_dim), indices_k)
# 		value_layer = index_first_axis(value_layer.reshape(batch_size * kv_seq_len, num_heads, head_dim), indices_k)

# 		if query_length == kv_seq_len:
# 			query_layer = index_first_axis(
# 				query_layer.reshape(batch_size * kv_seq_len, 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),
# 		)


# # copied from transformers.models.llama.modeling_llama.LlamaSdpaAttention with Llama->Mistral
# # TODO @Arthur no longer copied from LLama after static cache
class MistralSdpaAttention(MistralAttention):
	"""
	Mistral attention module using torch.nn.functional.scaled_dot_product_attention. This module inherits from
	`MistralAttention` as the weights of the module stays untouched. The only changes are on the forward pass to adapt to
	SDPA API.
	"""

	# Adapted from MistralAttention.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,
		hidden_mask=None,
	) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
		if output_attentions:
			logger.warning_once(
				"MistralModel is using MistralSdpaAttention, 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,
				hidden_mask=hidden_mask,
			)

		bsz, q_len, last_dim = 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)

		kv_seq_len = key_states.shape[-2]
		if past_key_value is not None:
			kv_seq_len += past_key_value.get_usable_length(kv_seq_len, self.layer_idx)
		cos, sin = self.rotary_emb(value_states, seq_len=kv_seq_len)

		query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)

		if past_key_value is not None:
			cache_kwargs = {"sin": sin, "cos": cos}  # Specific to RoPE models
			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)

		if attention_mask is not None:
			if attention_mask.size() != (bsz, 1, q_len, kv_seq_len):
				raise ValueError(
					f"Attention mask should be of size {(bsz, 1, q_len, kv_seq_len)}, but is {attention_mask.size()}"
				)
		if query_states.device.type == "cuda" and attention_mask is not None:
			query_states = query_states.contiguous()
			key_states = key_states.contiguous()
			value_states = value_states.contiguous()

		attn_output = torch.nn.functional.scaled_dot_product_attention(
			query_states,
			key_states,
			value_states,
			attn_mask=attention_mask,
			dropout_p=self.attention_dropout if self.training else 0.0,
			# The q_len > 1 is necessary to match with AttentionMaskConverter.to_causal_4d that does not create a causal mask in case q_len == 1.
			is_causal=self.is_causal and attention_mask is None and q_len > 1,
			) # bsz, self.num_heads, q_len, self.head_dim
		
		if self.config.vib_layers==True:
			attn_output = attn_output.permute(0,2,3,1).reshape(bsz* q_len* self.head_dim, self.num_key_value_groups,self.num_key_value_heads).reshape(bsz* q_len* self.head_dim*self.num_key_value_groups,-1) #bsz,q_len ,self.num_key_value_heads, self.num_key_value_groups,self.head_dim
			attn_output= self.ib_1(attn_output).to(attn_output.dtype)
			attn_output = attn_output.reshape(bsz,q_len,self.head_dim,self.num_key_value_heads* self.num_key_value_groups).permute(0,3,1,2)  #bsz,self.num_heads,q_len, self.head_dim
		
		attn_output = attn_output.transpose(1, 2).contiguous()
		attn_output = attn_output.view(bsz, q_len, self.num_heads*self.head_dim)

		attn_output = self.o_proj(attn_output)
		if self.config.vib_layers==True: #only during finetuning
			attn_output= attn_output.reshape(bsz* q_len,self.hidden_size)
			attn_output = hidden_mask(attn_output).to(attn_output.dtype)
			attn_output= attn_output.reshape(bsz, q_len, self.hidden_size)			
		return attn_output, None, past_key_value


MISTRAL_ATTENTION_CLASSES = {
	"eager": MistralAttention,
	#"flash_attention_2": MistralFlashAttention2,
	"sdpa": MistralSdpaAttention,
}


class MistralDecoderLayer(nn.Module):
	def __init__(self, config: MistralConfig, layer_idx: int):
		super().__init__()
		self.hidden_size = config.hidden_size
		self.config=config
		self.intermediate_size = config.intermediate_size
		self.self_attn = MISTRAL_ATTENTION_CLASSES[config._attn_implementation](config, layer_idx)
		self.mlp = MistralMLP(config)
		self.input_layernorm = MistralRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
		self.post_attention_layernorm = MistralRMSNorm(config.hidden_size, eps=config.rms_norm_eps)

	def get_kld_loss(self,hidden_mask,kl_fac):
		att_kl= self.self_attn.get_kld_loss(hidden_mask,kl_fac)
		mlp_kl=self.mlp.get_kld_loss(hidden_mask,kl_fac)
		return att_kl + mlp_kl

	def get_pars(self):
		inter_pars = 3*(self.intermediate_size * self.hidden_size) 
		atten_pars= ((2* self.hidden_size * self.hidden_size) + (self.hidden_size * 2 * self.self_attn.num_key_value_heads * self.self_attn.head_dim))        
		layer_norm=  2* self.hidden_size  
		return (inter_pars+atten_pars+layer_norm)		
	
	def get_num_rem_weights(self,hidden_mask_size): #head_list, mlp_list
		if self.mlp is not None:
			mlp_out_rem = self.mlp.get_num_rem_weights(hidden_mask_size)
		else:
			mlp_out_rem=0
		
		if self.self_attn is not None:
			att_out= self.self_attn.get_num_rem_weights(hidden_mask_size)
		else:
			att_out= 0
		layer_norm= 2 * hidden_mask_size
		return att_out, (mlp_out_rem + layer_norm) 

	def get_sparsity(self,hidden_mask): 
		if self.mlp is not None:
			hidden_mask_sparse= hidden_mask.sparse 
			out_sparse= self.mlp.get_sparsity(hidden_mask_sparse) 
		else:
			out_sparse = 0.0
		if self.self_attn is not None:
			hidden_mask_sparse= hidden_mask.sparse
			attn_sparse=  self.self_attn.get_sparsity(hidden_mask_sparse)
		else:
			attn_sparse=0.0
		layer_norm= 2 * hidden_mask_sparse.detach()/1e6
		return attn_sparse,(out_sparse +layer_norm) 

	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,
		hidden_mask= None,
		**kwargs,
	) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
		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.`"
			)
		"""
		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, sequence_length)` where padding elements are indicated by 0.
			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
		"""
		
		hidden_z= hidden_mask.get_mask_hard().int() if self.config.vib_layers==True else None
		residual = hidden_states
		if self.config.vib_layers==True: 
			hidden_states = self.input_layernorm(hidden_states,hidden_z)			
		else:
			#start_nor= time.time()
			hidden_states = self.input_layernorm(hidden_states)
			#lay_time= time.time() - start_nor
			
		# Self Attention
		#start_t= time.time()
		if self.self_attn is not None:
			st_att= time.time()
			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,
				hidden_mask=hidden_mask,
			)
			
			#att_time= time.time() - st_att 
			hidden_states = residual + hidden_states

			# Fully Connected
			residual = hidden_states
			if self.config.vib_layers==True:
				hidden_z= hidden_mask.get_mask_hard()
				hidden_states = self.post_attention_layernorm(hidden_states,hidden_z)				
			else:
				#start_norm= time.time()
				hidden_states = self.post_attention_layernorm(hidden_states)
				#lay_time += (time.time() - start_norm)
		#start_ti= time.time()
		if self.mlp is not None:
			#st_mlp= time.time()
			hidden_states = self.mlp(hidden_states,hidden_mask=hidden_mask)
			#mlp_time= time.time() - st_mlp
			hidden_states = residual + hidden_states

		outputs = (hidden_states,)

		if output_attentions:
			outputs += (self_attn_weights,)

		if use_cache:
			outputs += (present_key_value,)

		return outputs #, att_time, mlp_time,lay_time


MISTRAL_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 ([`MistralConfig`]):
			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 Mistral Model outputting raw hidden-states without any specific head on top.",
	MISTRAL_START_DOCSTRING,
)
class MistralPreTrainedModel(PreTrainedModel):
	config_class = MistralConfig
	base_model_prefix = "model"
	supports_gradient_checkpointing = True
	_no_split_modules = ["MistralDecoderLayer"]
	_skip_keys_device_placement = "past_key_values"
	_supports_flash_attn_2 = True
	_supports_sdpa = 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)
			module.weight.data.half()
			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_()
		elif isinstance(module, InformationBottleneck):
			module.post_z_mu.data.normal_(mean=1, std=0.01)
			module.post_z_logD.data.normal_(mean=-9, std=0.01)
			module.post_z_mu.data= module.post_z_mu.data.half()
			module.post_z_logD.data= module.post_z_logD.data.half()



MISTRAL_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 `decoder_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.
"""


@add_start_docstrings(
	"The bare Mistral Model outputting raw hidden-states without any specific head on top.",
	MISTRAL_START_DOCSTRING,
)
class MistralModel(MistralPreTrainedModel):
	"""
	Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`MistralDecoderLayer`]

	Args:
		config: MistralConfig
	"""

	def __init__(self, config: MistralConfig):
		super().__init__(config)
		self.padding_idx = config.pad_token_id
		self.vocab_size = config.vocab_size
		if config.vib_layers == True:
			self.hidden_mask= InformationBottleneck(config.hidden_size)
		else:
			self.hidden_mask=None
		self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
		self.layers = nn.ModuleList(
			[MistralDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
		)
		
		self._attn_implementation = config._attn_implementation
		self.norm = MistralRMSNorm(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.embed_tokens

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

	def get_num_rem_weights(self,mlp_list):
		mask = self.hidden_mask.get_mask_hard().int()
		num_states_rem = torch.sum((mask == 1).int())
		em_rem= (num_states_rem * self.config.vocab_size)+ num_states_rem #taking laynorm		
		return em_rem 

	def get_pars(self): 
		emb_pars= (self.config.vocab_size* self.config.hidden_size) 
		emb_pars+= self.norm.weight.size(0)
		return emb_pars

	def get_sparsity(self,mlp_list):
		em_sp=((self.hidden_mask.sparse/1e6) * self.config.vocab_size)
		em_sp += self.hidden_mask.sparse/1e6	
		return em_sp 

	@add_start_docstrings_to_model_forward(MISTRAL_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[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,

	) -> 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

		# retrieve input_ids and inputs_embeds
		if input_ids is not None and inputs_embeds is not None:
			raise ValueError("You cannot specify both decoder_input_ids and decoder_inputs_embeds at the same time")
		elif input_ids is not None:
			batch_size, seq_length = input_ids.shape
		elif inputs_embeds is not None:
			batch_size, seq_length, _ = inputs_embeds.shape
		else:
			raise ValueError("You have to specify either decoder_input_ids or decoder_inputs_embeds")

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

		past_key_values_length = 0

		if use_cache:
			use_legacy_cache = not isinstance(past_key_values, Cache)
			if use_legacy_cache:
				past_key_values = DynamicCache.from_legacy_cache(past_key_values)
			past_key_values_length = past_key_values.get_usable_length(seq_length)

		if position_ids is None:
			device = input_ids.device if input_ids is not None else inputs_embeds.device
			position_ids = torch.arange(
				past_key_values_length, seq_length + past_key_values_length, dtype=torch.long, device=device
			)
			position_ids = position_ids.unsqueeze(0).view(-1, seq_length)
		else:
			position_ids = position_ids.view(-1, seq_length).long()
		#st_em= time.time()
		if inputs_embeds is None:
			inputs_embeds = self.embed_tokens(input_ids)
			#emb_time= time.time() - st_em

		if attention_mask is not None and self._attn_implementation == "flash_attention_2" and use_cache:
			is_padding_right = attention_mask[:, -1].sum().item() != batch_size
			if is_padding_right:
				raise ValueError(
					"You are attempting to perform batched generation with padding_side='right'"
					" this may lead to unexpected behaviour for Flash Attention version of Mistral. Make sure to "
					" call `tokenizer.padding_side  = 'left'` before tokenizing the input. "
				)
		if self.config.vib_layers is True: #only finetuning 
			bsz,seq,dim= inputs_embeds.size()
			inputs_embeds =inputs_embeds.reshape(bsz*seq,dim)
			inputs_embeds=self.hidden_mask(inputs_embeds).to(inputs_embeds.dtype)
			inputs_embeds =inputs_embeds.reshape(bsz,seq,dim)			

		if self._attn_implementation == "flash_attention_2":
			# 2d mask is passed through the layers
			attention_mask = attention_mask if (attention_mask is not None and 0 in attention_mask) else None
		elif self._attn_implementation == "sdpa" and not output_attentions:
			# output_attentions=True can not be supported when using SDPA, and we fall back on
			# the manual implementation that requires a 4D causal mask in all cases.
			attention_mask = _prepare_4d_causal_attention_mask_for_sdpa(
				attention_mask,
				(batch_size, seq_length),
				inputs_embeds,
				past_key_values_length,
				sliding_window=self.config.sliding_window,
			)
		else:
			# 4d mask is passed through the layers
			attention_mask = _prepare_4d_causal_attention_mask(
				attention_mask,
				(batch_size, seq_length),
				inputs_embeds,
				past_key_values_length,
				sliding_window=self.config.sliding_window,
			)

		if self.config.finetune is True and self.config.vib_layers is True: #only finetuning 
			inputs_embeds=self.hidden_mask(inputs_embeds)
			hidden_z= self.hidden_mask.get_mask_hard().int()
			hidden_states = hidden_states.mul(hidden_z)
		elif self.config.vib_layers is True: #during pruning
			bsz,seq,dim= inputs_embeds.size()
			inputs_embeds=self.hidden_mask(inputs_embeds.reshape(bsz*seq,dim)).reshape(bsz,seq,dim)

		hidden_states = inputs_embeds
		# decoder layers
		all_hidden_states = () if output_hidden_states else None
		all_self_attns = () if output_attentions else None
		next_decoder_cache = None
		att_time=0
		mlp_time=0
		norm_time=0
		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,
					attention_mask,
					position_ids,
					past_key_values,
					output_attentions,
					use_cache,
				)
			else:
				layer_outputs = decoder_layer(
					hidden_states,
					attention_mask=attention_mask,
					position_ids=position_ids,
					past_key_value=past_key_values,
					output_attentions=output_attentions,
					use_cache=use_cache,
					hidden_mask=self.hidden_mask 
				)
				# att_time += att_t
				# mlp_time += mlp_t 
				# norm_time += lay_t
			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],)			
			
		if self.config.vib_layers == True:
			hidden_z=self.hidden_mask.get_mask_hard().int()
			hidden_states = self.norm(hidden_states,hidden_z)			
		else:
			#st_norm =time.time()
			hidden_states = self.norm(hidden_states)
			#norm_time += (time.time() - st_norm)
		# 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 use_legacy_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,
		)


class MistralForCausalLM(MistralPreTrainedModel):
	_tied_weights_keys = ["lm_head.weight"]

	def __init__(self, config):
		super().__init__(config)
		self.model = MistralModel(config)
		self.vocab_size = config.vocab_size
		self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
		if config.vib_layers is True:
			self.lambda_1 = torch.nn.Parameter(torch.zeros(1)) 
			self.lambda_2 = torch.nn.Parameter(torch.zeros(1))
		else:
			self.lambda_1=None 
			self.lambda_2=None

		self.config = config
		self.sparsity=0.0
		self.start_sparsity=0.0
		self.target_sparsity=0.0
		self.spar_loss_1=0.0
		self.spar_loss_2=0.0
		self.lagrangian_warmup=20
		self.emb_mult=1
		self.prunable_parameters=self.get_prunable_pars()

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

	def set_lagpars(self):
		torch.nn.init.zeros_(self.lambda_1)
		torch.nn.init.zeros_(self.lambda_2)

	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 prune_heads(self, heads):
		len_heads = len(heads)
		if len_heads == 0:
			return

		_, index = find_pruneable_heads_and_indices(heads, config.num_attention_heads, 128, set())
		
		# Prune linear layers
		if len(index) == 0:
			self.self.query = None
			self.self.key = None
			self.self.value = None
			self.output = None
		else:
			self.self.query = prune_linear_layer(self.self.query, index)
			self.self.key = prune_linear_layer(self.self.key, index)
			self.self.value = prune_linear_layer(self.self.value, index)
			self.output.dense = prune_linear_layer(
				self.output.dense, index, dim=1)

		# Update hyper params and store pruned heads
		self.num_heads = self.num_heads -  len(index)
		self.num_key_value_heads=self.num_key_value_heads - len(index)

	
	def get_prunable_pars(self):
		pars=0
		for layer in self.model.layers:
			pars += int(layer.get_pars())
			
		# if self.emb_mult!=0:
		# 	pars+= int(self.model.get_pars())
			
		return pars


	def use_masking(self,mask,emb=0.0): #changed
		if self.emb_mult ==1: #first time
			self.emb_mult=emb
			self.prunable_parameters= self.get_prunable_pars()
			
		if self.config.vib_layers==True:
			self.model.hidden_mask.masking= mask
			self.model.hidden_mask.sample_in_training= False if self.config.finetune == True else True
			for layer in self.model.layers:
				if layer.self_attn is not None: 
					layer.self_attn.ib_1.masking = mask
					layer.self_attn.ib_1.sample_in_training= False if self.config.finetune == True else True
					
				if layer.mlp is not None:
					layer.mlp.ib_1.masking =mask
					layer.mlp.ib_1.sample_in_training= False if self.config.finetune == True else True
					

	def get_kld_loss(self,emb_mult=25,layer_mult=1000,kl_fac=1e-5):
		loss = 0
		self.emb_mult=emb_mult
		for layer in self.model.layers:
			loss += layer.get_kld_loss(self.model.hidden_mask,kl_fac)
			
		if self.emb_mult != 0:
			loss+= (self.model.hidden_mask.kld*emb_mult*kl_fac) #embedding loss weighted more #commented it to have only 1 loss for hiddenmask
			
		return loss
	
	def get_num_rem_weights(self):
		num_states_rem = 0
		hidden_mask_size= torch.sum((self.model.hidden_mask.get_mask_hard().int() ==1))#.item()
		for (i, layer) in enumerate(self.model.layers):
			att_sparsity, ffn_sparsity = layer.get_num_rem_weights(hidden_mask_size) #mlp_scores[i]) #head_scores[i],
			num_states_rem += (int(att_sparsity  + ffn_sparsity)) 
			
		
		# if self.emb_mult != 0:
		# 	num_states_rem += self.model.get_num_rem_weights(hidden_mask_size) #mlp_scores)
		return num_states_rem

	def get_sparsity(self):
		rem_sum = 0.0
		hidden_mask_sparse =self.model.hidden_mask.sparse
		for (i, layer) in enumerate(self.model.layers):
			att_sparsity, ffn_sparsity = layer.get_sparsity(self.model.hidden_mask) 
			rem_sum += (att_sparsity + ffn_sparsity )     
				
		# if self.emb_mult != 0:
		# 	rem_sum+= self.model.get_sparsity(hidden_mask_sparse) #mlp_scores) 
			
		remaining=rem_sum / (self.prunable_parameters/1e6)
		
		
		self.sparsity=1-remaining 
		return self.sparsity

	def set_lagrangian_warmup_steps(self, lagrangian_warmup):
		self.lagrangian_warmup = lagrangian_warmup

	def set_start_and_target_sparsity(self, target_sparsity):
		rem_pars= self.get_num_rem_weights()
		self.start_sparsity=(self.prunable_parameters-rem_pars)/self.prunable_parameters
		self.target_sparsity=target_sparsity
		print("\n start sp {} target sp {}".format(self.start_sparsity,self.target_sparsity))
		


	def get_target_sparsity(self,pruned_steps):
		target_sparsity = (self.target_sparsity - self.start_sparsity) * min(1, pruned_steps/self.lagrangian_warmup) + self.start_sparsity 
		return target_sparsity


	def lagrangian_regularization(self, pruned_steps): 
		expected_sparsity = self.get_sparsity()
		if self.lagrangian_warmup > 0:
			target_sparsity = self.get_target_sparsity(pruned_steps)
		else:
			target_sparsity=target_sparsity

		lagrangian_loss = (  (self.lambda_1[0] * (expected_sparsity - target_sparsity)) + (self.lambda_2[0] * ((expected_sparsity - target_sparsity) ** 2))) 
		spar_loss_1=self.lambda_1.detach().item() * (expected_sparsity.detach().cpu() - target_sparsity)
		spar_loss_2=self.lambda_2.detach().item() * ((expected_sparsity.detach().cpu() - target_sparsity) ** 2)
		return lagrangian_loss, expected_sparsity, target_sparsity,spar_loss_1,spar_loss_2

	def calculate_model_size(self): 
		remaining_model_size = self.get_num_rem_weights()
		
		pruned_size = (self.prunable_parameters - remaining_model_size)
		hidden_mask_size= self.model.hidden_mask.get_mask_hard().int()
		embedded_dims=torch.sum((hidden_mask_size ==1).int()).item()  #number of embedding hidden_states left
		embed_rem= (self.config.vocab_size * embedded_dims) 
		intermediate_dims=[]
		attention_head_dims=[]		
		att_head_rem=0
		atten_out_rem=0
		inter_rem=0
		out_rem=0
		tot_att_head_rem=0
		tot_atten_out_rem=0
		tot_inter_rem=0
		tot_out_rem=0
		for (i, layer) in enumerate(self.model.layers):
			if layer.mlp is not None:
				mask_1 = torch.sum((layer.mlp.ib_1.get_mask_hard()==1).int()).item()
				intermediate_dims.append((mask_1))
				inter_rem+= (2*embedded_dims*mask_1) #no bias
				out_rem+= (embedded_dims* mask_1)#no bias
				tot_inter_rem+= (2* self.config.hidden_size * self.config.intermediate_size)
				tot_out_rem+= (self.config.hidden_size * self.config.intermediate_size)
			
			if layer.self_attn is not None:
				mask_1= torch.sum((layer.self_attn.ib_1.get_mask_hard() == 1).int()).item()
				num_states_rem_self= mask_1#* mlp_z[i]
				attention_head_dims.append(num_states_rem_self)				
				att_head_rem += (embedded_dims* num_states_rem_self*layer.self_attn.head_dim*2) + (embedded_dims* num_states_rem_self*layer.self_attn.num_key_value_groups* layer.self_attn.head_dim) #multiplied by attention head size- which is 128 for llama model used
				tot_att_head_rem += ((self.config.hidden_size * self.config.hidden_size) + (self.config.hidden_size * self.config.num_key_value_heads * layer.self_attn.head_dim * 2))
				atten_out_rem+= (layer.self_attn.num_key_value_groups * num_states_rem_self*layer.self_attn.head_dim*embedded_dims)#*mlp_z[i])
				tot_atten_out_rem += (self.config.hidden_size * self.config.hidden_size)
			

		results = {}		
		results["embedded_dims"] = embedded_dims
		results["attention_head_dim1"] = embedded_dims 
		results["attention_head_dims"] = attention_head_dims
		results["attention_output_dims"] = embedded_dims 
		results["intermediate_dim1"] = embedded_dims 
		results["intermediate_dims"] = intermediate_dims
		results["output_layer_dims"] = embedded_dims
		results["pruned_params (Mil)"] = pruned_size/1e6
		results["remaining_params (Mil)"] = remaining_model_size/1e6
		results["embedding pars "] = embed_rem / ((self.config.vocab_size * self.config.hidden_size))
		results["attention head pars "] = att_head_rem / tot_att_head_rem
		results["atten_out pars "] = atten_out_rem /tot_atten_out_rem
		results["intermediate pars "] = inter_rem /tot_inter_rem
		results["out pars "] = out_rem /tot_out_rem       
		results["actual_model_sparsity"] = round((pruned_size / self.prunable_parameters),5)
		return results

	@add_start_docstrings_to_model_forward(MISTRAL_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[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, 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, MistralForCausalLM

		>>> model = MistralForCausalLM.from_pretrained("mistralai/Mistral-7B-v0.1")
		>>> tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")

		>>> 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)
		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.lm_head(hidden_states)
		logits = logits.float()

		loss = None
		if labels is not None:
			# Shift so that tokens < n predict n
			shift_logits = logits[..., :-1, :].contiguous()
			shift_labels = labels[..., 1:].contiguous()
			# Flatten the tokens
			shift_logits = shift_logits.view(-1, self.config.vocab_size)
			shift_labels = shift_labels.view(-1)
			# Ensure tensors are on the same device
			shift_labels = shift_labels.to(shift_logits.device)
			loss_fct = CrossEntropyLoss()
			loss = loss_fct(shift_logits, shift_labels)

		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,
		)

	def prepare_inputs_for_generation(
		self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs
	):
		# Omit tokens covered by past_key_values
		if past_key_values is not None:
			if isinstance(past_key_values, Cache):
				cache_length = past_key_values.get_seq_length()
				past_length = past_key_values.seen_tokens
				max_cache_length = past_key_values.get_max_length()
			else:
				cache_length = past_length = past_key_values[0][0].shape[2]
				max_cache_length = None

			# Keep only the unprocessed tokens:
			# 1 - If the length of the attention_mask exceeds the length of input_ids, then we are in a setting where
			# some of the inputs are exclusively passed as part of the cache (e.g. when passing input_embeds as
			# input)
			if attention_mask is not None and attention_mask.shape[1] > input_ids.shape[1]:
				input_ids = input_ids[:, -(attention_mask.shape[1] - past_length) :]
			# 2 - If the past_length is smaller than input_ids', then input_ids holds all input tokens. We can discard
			# input_ids based on the past_length.
			elif past_length < input_ids.shape[1]:
				input_ids = input_ids[:, past_length:]
			# 3 - Otherwise (past_length >= input_ids.shape[1]), let's assume input_ids only has unprocessed tokens.

			# If we are about to go beyond the maximum cache length, we need to crop the input attention mask.
			if (
				max_cache_length is not None
				and attention_mask is not None
				and cache_length + input_ids.shape[1] > max_cache_length
			):
				attention_mask = attention_mask[:, -max_cache_length:]

		position_ids = kwargs.get("position_ids", None)
		if attention_mask is not None and position_ids is None:
			# create position_ids on the fly for batch generation
			position_ids = attention_mask.long().cumsum(-1) - 1
			position_ids.masked_fill_(attention_mask == 0, 1)
			if past_key_values:
				position_ids = position_ids[:, -input_ids.shape[1] :]

		# if `inputs_embeds` are passed, we only want to use them in the 1st generation step
		if inputs_embeds is not None and past_key_values is None:
			model_inputs = {"inputs_embeds": inputs_embeds}
		else:
			model_inputs = {"input_ids": input_ids}

		model_inputs.update(
			{
				"position_ids": position_ids,
				"past_key_values": past_key_values,
				"use_cache": kwargs.get("use_cache"),
				"attention_mask": attention_mask,
			}
		)
		return model_inputs

	@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


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

	[`MistralForSequenceClassification`] 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).
	""",
	MISTRAL_START_DOCSTRING,
)
# Copied from transformers.models.llama.modeling_llama.LlamaForSequenceClassification with Llama->Mistral, LLAMA->MISTRAL
class MistralForSequenceClassification(MistralPreTrainedModel):
	def __init__(self, config):
		super().__init__(config)
		self.num_labels = config.num_labels
		self.model = MistralModel(config)
		self.score = nn.Linear(config.hidden_size, self.num_labels, bias=False)

		# 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

	@add_start_docstrings_to_model_forward(MISTRAL_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[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

		transformer_outputs = self.model(
			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 = transformer_outputs[0]
		logits = self.score(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:
			labels = labels.to(logits.device)
			if self.config.problem_type is None:
				if self.num_labels == 1:
					self.config.problem_type = "regression"
				elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int):
					self.config.problem_type = "single_label_classification"
				else:
					self.config.problem_type = "multi_label_classification"

			if self.config.problem_type == "regression":
				loss_fct = MSELoss()
				if self.num_labels == 1:
					loss = loss_fct(pooled_logits.squeeze(), labels.squeeze())
				else:
					loss = loss_fct(pooled_logits, labels)
			elif self.config.problem_type == "single_label_classification":
				loss_fct = CrossEntropyLoss()
				loss = loss_fct(pooled_logits.view(-1, self.num_labels), labels.view(-1))
			elif self.config.problem_type == "multi_label_classification":
				loss_fct = BCEWithLogitsLoss()
				loss = loss_fct(pooled_logits, labels)
		if not return_dict:
			output = (pooled_logits,) + transformer_outputs[1:]
			return ((loss,) + output) if loss is not None else output

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