Upload modeling_phi3.py with huggingface_hub

modeling_phi3.py

@@ -0,0 +1,1606 @@
+# coding=utf-8
+# Copyright 2024 Microsoft and the HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+""" PyTorch Phi-3 model."""
+
+import inspect
+import math
+import warnings
+from typing import List, Optional, Tuple, Union
+
+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 transformers.cache_utils import Cache, DynamicCache
+from transformers.modeling_attn_mask_utils import _prepare_4d_causal_attention_mask
+from transformers.modeling_outputs import (
+    BaseModelOutputWithPast,
+    CausalLMOutputWithPast,
+    SequenceClassifierOutputWithPast,
+    TokenClassifierOutput,
+)
+from transformers.modeling_utils import PreTrainedModel
+from transformers.utils import (
+    add_code_sample_docstrings,
+    add_start_docstrings,
+    add_start_docstrings_to_model_forward,
+    is_flash_attn_2_available,
+    is_flash_attn_greater_or_equal_2_10,
+    logging,
+    replace_return_docstrings,
+)
+from .configuration_phi3 import Phi3Config
+
+
+logger = logging.get_logger(__name__)
+
+# Transformers scans dependencies in the modeling file, causing issues on conditional loading. The regex only ignores try/catch blocks, but not if statements
+# if is_flash_attn_2_available():
+_flash_supports_window_size = False
+try:
+    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)
+except ImportError as error:
+    logger.warning(
+        f"`flash-attention` package not found, consider installing for better performance: {error}."
+    )
+    if not _flash_supports_window_size:
+        logger.warning(
+            "Current `flash-attenton` does not support `window_size`. Either upgrade or use `attn_implementation='eager'`."
+        )
+
+_CHECKPOINT_FOR_DOC = "microsoft/Phi-3-mini-4k-instruct"
+_CONFIG_FOR_DOC = "Phi3Config"
+
+PHI3_PRETRAINED_MODEL_ARCHIVE_LIST = [
+    "microsoft/Phi-3-mini-4k-instruct",
+    "microsoft/Phi-3-mini-128k-instruct",
+    # See all Phi-3 models at https://huggingface.co/models?filter=Phi-3
+]
+
+
+# Copied from transformers.models.llama.modeling_llama.LlamaRMSNorm with Llama->Phi3
+class Phi3RMSNorm(nn.Module):
+    def __init__(self, hidden_size, eps=1e-6):
+        """
+        Phi3RMSNorm 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)
+
+
+# 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.gemma.modeling_gemma.GemmaRotaryEmbedding with gemma->phi3, Gemma->Phi3
+class Phi3RotaryEmbedding(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
+        self.register_buffer("inv_freq", None, persistent=False)
+
+    @torch.no_grad()
+    def forward(self, x, position_ids, seq_len=None):
+        # x: [bs, num_attention_heads, seq_len, head_size]
+        if self.inv_freq is None:
+            self.inv_freq = 1.0 / (
+                self.base ** (torch.arange(0, self.dim, 2, dtype=torch.int64, device=x.device).float() / self.dim)
+            )
+        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 since bfloat16 loses precision on long contexts
+        # 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()
+        return cos.to(dtype=x.dtype), sin.to(dtype=x.dtype)
+
+
+class Phi3SuScaledRotaryEmbedding(Phi3RotaryEmbedding):
+    def __init__(self, dim, config, device=None):
+        super().__init__(dim, config.max_position_embeddings, config.rope_theta, device)
+
+        self.short_factor = config.rope_scaling["short_factor"]
+        self.long_factor = config.rope_scaling["long_factor"]
+        self.original_max_position_embeddings = config.original_max_position_embeddings
+
+    @torch.no_grad()
+    def forward(self, x, position_ids, seq_len=None):
+        seq_len = torch.max(position_ids) + 1
+        if seq_len > self.original_max_position_embeddings:
+            ext_factors = torch.tensor(self.long_factor, dtype=torch.float32, device=x.device)
+        else:
+            ext_factors = torch.tensor(self.short_factor, dtype=torch.float32, device=x.device)
+
+        inv_freq_shape = torch.arange(0, self.dim, 2, dtype=torch.int64, device=x.device).float() / self.dim
+        self.inv_freq = 1.0 / (ext_factors * self.base**inv_freq_shape)
+
+        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 since bfloat16 loses precision on long contexts
+        # 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)
+
+            scale = self.max_position_embeddings / self.original_max_position_embeddings
+            if scale <= 1.0:
+                scaling_factor = 1.0
+            else:
+                scaling_factor = math.sqrt(1 + math.log(scale) / math.log(self.original_max_position_embeddings))
+
+            cos = emb.cos() * scaling_factor
+            sin = emb.sin() * scaling_factor
+        return cos.to(dtype=x.dtype), sin.to(dtype=x.dtype)
+
+
+class Phi3YarnScaledRotaryEmbedding(Phi3RotaryEmbedding):
+    def __init__(self, dim, config, device=None):
+        super().__init__(dim, config.max_position_embeddings, config.rope_theta, device)
+
+        self.short_factor = config.rope_scaling["short_factor"]
+        self.long_factor = config.rope_scaling["long_factor"]
+        self.original_max_position_embeddings = config.original_max_position_embeddings
+
+    @torch.no_grad()
+    def forward(self, x, position_ids, seq_len=None):
+        seq_len = torch.max(position_ids) + 1
+        if seq_len > self.original_max_position_embeddings:
+            ext_factors = torch.tensor(self.long_factor, dtype=torch.float32, device=x.device)
+        else:
+            ext_factors = torch.tensor(self.short_factor, dtype=torch.float32, device=x.device)
+
+        inv_freq_shape = torch.arange(0, self.dim, 2, dtype=torch.int64, device=x.device).float() / self.dim
+        self.inv_freq = 1.0 / (ext_factors * self.base**inv_freq_shape)
+
+        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 since bfloat16 loses precision on long contexts
+        # 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)
+
+            scale = self.max_position_embeddings / self.original_max_position_embeddings
+            if scale <= 1.0:
+                scaling_factor = 1.0
+            else:
+                scaling_factor = 0.1 * math.log(scale) + 1.0
+
+            cos = emb.cos() * scaling_factor
+            sin = emb.sin() * scaling_factor
+        return cos.to(dtype=x.dtype), sin.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)
+
+
+# Copied from transformers.models.llama.modeling_llama.apply_rotary_pos_emb
+def apply_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
+
+
+class Phi3MLP(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+
+        self.config = config
+        self.gate_up_proj = nn.Linear(config.hidden_size, 2 * config.intermediate_size, bias=False)
+        self.down_proj = nn.Linear(config.intermediate_size, config.hidden_size, bias=False)
+
+        self.activation_fn = ACT2FN[config.hidden_act]
+
+    def forward(self, hidden_states: torch.FloatTensor) -> torch.FloatTensor:
+        up_states = self.gate_up_proj(hidden_states)
+
+        gate, up_states = up_states.chunk(2, dim=-1)
+        up_states = up_states * self.activation_fn(gate)
+
+        return self.down_proj(up_states)
+
+
+# Copied from transformers.models.llama.modeling_llama.repeat_kv with llama->phi
+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 Phi3Attention(nn.Module):
+    """Multi-headed attention from 'Attention Is All You Need' paper"""
+
+    def __init__(self, config: Phi3Config, layer_idx: Optional[int] = None):
+        super().__init__()
+        self.config = config
+        self.layer_idx = layer_idx
+        if layer_idx is None:
+            logger.warning_once(
+                f"Instantiating {self.__class__.__name__} without passing a `layer_idx` is not recommended and will "
+                "lead to errors during the forward call if caching is used. Please make sure to provide a `layer_idx` "
+                "when creating this class."
+            )
+
+        self.attention_dropout = config.attention_dropout
+        self.hidden_size = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = self.hidden_size // self.num_heads
+        self.num_key_value_heads = config.num_key_value_heads
+        self.num_key_value_groups = self.num_heads // self.num_key_value_heads
+        self.max_position_embeddings = config.max_position_embeddings
+        self.original_max_position_embeddings = config.original_max_position_embeddings
+        self.rope_theta = config.rope_theta
+        self.rope_scaling = config.rope_scaling
+        self.is_causal = True
+
+        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})."
+            )
+
+        op_size = self.num_heads * self.head_dim + 2 * (self.num_key_value_heads * self.head_dim)
+        self.o_proj = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=False)
+        self.qkv_proj = nn.Linear(self.hidden_size, op_size, bias=False)
+        self._init_rope()
+
+    def _init_rope(self):
+        if self.rope_scaling is None:
+            self.rotary_emb = Phi3RotaryEmbedding(
+                self.head_dim,
+                max_position_embeddings=self.max_position_embeddings,
+                base=self.rope_theta,
+            )
+        else:
+            scaling_type = self.config.rope_scaling["type"]
+            if scaling_type == "su":
+                self.rotary_emb = Phi3SuScaledRotaryEmbedding(self.head_dim, self.config)
+            elif scaling_type == "yarn":
+                self.rotary_emb = Phi3YarnScaledRotaryEmbedding(self.head_dim, self.config)
+            else:
+                raise ValueError(f"Unknown RoPE scaling type {scaling_type}")
+
+    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,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        logger.warning_once("You are not running the flash-attention implementation, expect numerical differences.")
+
+        bsz, q_len, _ = hidden_states.size()
+
+        qkv = self.qkv_proj(hidden_states)
+        query_pos = self.num_heads * self.head_dim
+        query_states = qkv[..., :query_pos]
+        key_states = qkv[..., query_pos : query_pos + self.num_key_value_heads * self.head_dim]
+        value_states = qkv[..., query_pos + self.num_key_value_heads * self.head_dim :]
+
+        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)
+        cos, sin = self.rotary_emb(value_states, position_ids, 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)
+
+        # 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)
+
+        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
+
+        # upcast attention to fp32
+        attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(value_states.dtype)
+        attn_weights = nn.functional.dropout(attn_weights, p=self.attention_dropout, training=self.training)
+
+        attn_output = torch.matmul(attn_weights, value_states)
+
+        if attn_output.size() != (bsz, self.num_heads, q_len, self.head_dim):
+            raise ValueError(
+                f"`attn_output` should be of size {(bsz, self.num_heads, q_len, self.head_dim)}, but is"
+                f" {attn_output.size()}"
+            )
+
+        attn_output = attn_output.transpose(1, 2).contiguous()
+        attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
+
+        attn_output = self.o_proj(attn_output)
+
+        if not output_attentions:
+            attn_weights = None
+
+        return attn_output, attn_weights, past_key_value
+
+
+class Phi3FlashAttention2(Phi3Attention):
+    """
+    Phi-3 flash attention module. This module inherits from `Phi3Attention` 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.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_value: Optional[Cache] = None,
+        output_attentions: bool = False,
+        use_cache: bool = False,
+        **kwargs,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        # Phi3FlashAttention2 attention does not support output_attentions
+
+        if not _flash_supports_window_size:
+            logger.warning_once(
+                "The current flash attention version does not support sliding window attention. Please use `attn_implementation='eager'` or upgrade flash-attn library."
+            )
+            raise ValueError("The current flash attention version does not support sliding window attention.")
+
+        output_attentions = False
+
+        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()
+
+        qkv = self.qkv_proj(hidden_states)
+        query_pos = self.num_heads * self.head_dim
+        query_states = qkv[..., :query_pos]
+        key_states = qkv[..., query_pos : query_pos + self.num_key_value_heads * self.head_dim]
+        value_states = qkv[..., query_pos + self.num_key_value_heads * self.head_dim :]
+
+        # Flash attention requires the input to have the shape
+        # batch_size x seq_length x head_dim x hidden_dim
+        # therefore we just need to keep the original shape
+        query_states = query_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
+        key_states = key_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+        value_states = value_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
+
+        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, position_ids, 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 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)
+
+        attn_dropout = self.attention_dropout if self.training else 0.0
+
+        # In PEFT, usually we cast the layer norms in float32 for training stability reasons
+        # therefore the input hidden states gets silently casted in float32. Hence, we need
+        # cast them back in the correct dtype just to be sure everything works as expected.
+        # This might slowdown training & inference so it is recommended to not cast the LayerNorms
+        # in fp32.
+
+        if query_states.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.qkv_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=attn_dropout,
+            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
+
+    # Copied from transformers.models.mistral.modeling_mistral.MistralFlashAttention2._flash_attention_forward
+    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
+
+    # Copied from transformers.models.mistral.modeling_mistral.MistralFlashAttention2._upad_input
+    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->Phi3
+# TODO @Arthur no longer copied from LLama after static cache
+class Phi3SdpaAttention(Phi3Attention):
+    """
+    Phi3 attention module using torch.nn.functional.scaled_dot_product_attention. This module inherits from
+    `Phi3Attention` as the weights of the module stays untouched. The only changes are on the forward pass to adapt to
+    SDPA API.
+    """
+
+    # Adapted from Phi3Attention.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,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        if output_attentions:
+            # TODO: Improve this warning with e.g. `model.config.attn_implementation = "manual"` once this is implemented.
+            logger.warning_once(
+                "Phi3Model is using Phi3SdpaAttention, 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,
+            )
+
+        bsz, q_len, _ = hidden_states.size()
+
+        qkv = self.qkv_proj(hidden_states)
+        query_pos = self.num_heads * self.head_dim
+        query_states = qkv[..., :query_pos]
+        key_states = qkv[..., query_pos : query_pos + self.num_key_value_heads * self.head_dim]
+        value_states = qkv[..., query_pos + self.num_key_value_heads * self.head_dim :]
+
+        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, position_ids, 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()}"
+                )
+
+        # SDPA with memory-efficient backend is currently (torch==2.1.2) bugged with non-contiguous inputs with custom attn_mask,
+        # Reference: https://github.com/pytorch/pytorch/issues/112577.
+        if query_states.device.type == "cuda" and 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,
+        )
+
+        attn_output = attn_output.transpose(1, 2).contiguous()
+        attn_output = attn_output.view(bsz, q_len, self.hidden_size)
+
+        attn_output = self.o_proj(attn_output)
+
+        return attn_output, None, past_key_value
+
+
+PHI3_ATTENTION_CLASSES = {
+    "eager": Phi3Attention,
+    "flash_attention_2": Phi3FlashAttention2,
+    "sdpa": Phi3SdpaAttention,
+}
+
+
+class Phi3DecoderLayer(nn.Module):
+    def __init__(self, config: Phi3Config, layer_idx: int):
+        super().__init__()
+
+        self.config = config
+        self.self_attn = PHI3_ATTENTION_CLASSES[config._attn_implementation](config, layer_idx=layer_idx)
+
+        self.mlp = Phi3MLP(config)
+        self.input_layernorm = Phi3RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+
+        self.resid_attn_dropout = nn.Dropout(config.resid_pdrop)
+        self.resid_mlp_dropout = nn.Dropout(config.resid_pdrop)
+        self.post_attention_layernorm = Phi3RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_value: Optional[Tuple[torch.Tensor]] = None,
+        output_attentions: Optional[bool] = False,
+        use_cache: Optional[bool] = False,
+        **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, 1, tgt_len, src_len)` where padding elements are indicated by very large negative values.
+            position_ids (`torch.LongTensor` of shape `({0})`, *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)
+            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
+        """
+
+        residual = hidden_states
+
+        hidden_states = self.input_layernorm(hidden_states)
+
+        # Self Attention
+        attn_outputs, 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_states = residual + self.resid_attn_dropout(attn_outputs)
+
+        residual = hidden_states
+        hidden_states = self.post_attention_layernorm(hidden_states)
+        hidden_states = self.mlp(hidden_states)
+        hidden_states = residual + self.resid_mlp_dropout(hidden_states)
+
+        outputs = (hidden_states,)
+
+        if output_attentions:
+            outputs += (self_attn_weights,)
+
+        if use_cache:
+            outputs += (present_key_value,)
+
+        return outputs
+
+
+PHI3_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 ([`Phi3Config`]):
+            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 Phi-3 model outputting raw hidden-states without any specific head on top.",
+    PHI3_START_DOCSTRING,
+)
+class Phi3PreTrainedModel(PreTrainedModel):
+    config_class = Phi3Config
+    base_model_prefix = "model"
+    supports_gradient_checkpointing = True
+    _no_split_modules = ["Phi3DecoderLayer"]
+    _skip_keys_device_placement = "past_key_values"
+    _supports_flash_attn_2 = True
+    _supports_sdpa = False
+    _supports_cache_class = True
+
+    _version = "0.0.5"
+
+    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_()
+
+
+PHI3_INPUTS_DOCSTRING = r"""
+    Args:
+        input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
+            Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
+            it.
+
+            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            [What are input IDs?](../glossary#input-ids)
+        attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
+
+            - 1 for tokens that are **not masked**,
+            - 0 for tokens that are **masked**.
+
+            [What are attention masks?](../glossary#attention-mask)
+
+            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
+            [`PreTrainedTokenizer.__call__`] for details.
+
+            If `past_key_values` is used, optionally only the last `input_ids` have to be input (see
+            `past_key_values`).
+
+            If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`]
+            and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more
+            information on the default strategy.
+
+            - 1 indicates the head is **not masked**,
+            - 0 indicates the head is **masked**.
+        position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
+            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
+            config.n_positions - 1]`.
+
+            [What are position IDs?](../glossary#position-ids)
+        past_key_values (`Cache` or `tuple(tuple(torch.FloatTensor))`, *optional*):
+            Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
+            blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values`
+            returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`.
+
+            Two formats are allowed:
+            - a [`~cache_utils.Cache`] instance;
+            - Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of
+            shape `(batch_size, num_heads, sequence_length, embed_size_per_head)`). This is also known as the legacy
+            cache format.
+
+            The model will output the same cache format that is fed as input. If no `past_key_values` are passed, the
+            legacy cache format will be returned.
+
+            If `past_key_values` are used, the user can optionally input only the last `input_ids` (those that don't
+            have their past key value states given to this model) of shape `(batch_size, 1)` instead of all `input_ids`
+            of shape `(batch_size, sequence_length)`.
+        inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
+            Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
+            is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
+            model's internal embedding lookup matrix.
+        use_cache (`bool`, *optional*):
+            If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see
+            `past_key_values`).
+        output_attentions (`bool`, *optional*):
+            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
+            tensors for more detail.
+        output_hidden_states (`bool`, *optional*):
+            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
+            more detail.
+        return_dict (`bool`, *optional*):
+            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
+"""
+
+
+@add_start_docstrings(
+    "The bare Phi-3 model outputting raw hidden-states without any specific head on top.",
+    PHI3_START_DOCSTRING,
+)
+class Phi3Model(Phi3PreTrainedModel):
+    """
+    Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`Phi3DecoderLayer`]
+
+    Args:
+        config: Phi3Config
+    """
+
+    def __init__(self, config: Phi3Config):
+        super().__init__(config)
+        self.padding_idx = config.pad_token_id
+        self.vocab_size = config.vocab_size
+
+        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
+        self.embed_dropout = nn.Dropout(config.embd_pdrop)
+        self.layers = nn.ModuleList(
+            [Phi3DecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
+        )
+        self._attn_implementation = config._attn_implementation
+        self.norm = Phi3RMSNorm(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
+
+    @add_start_docstrings_to_model_forward(PHI3_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 input_ids and inputs_embeds at the same time")
+        elif input_ids is not None:
+            batch_size, seq_length = input_ids.shape[:2]
+        elif inputs_embeds is not None:
+            batch_size, seq_length = inputs_embeds.shape[:2]
+        else:
+            raise ValueError("You have to specify either input_ids or inputs_embeds")
+
+        past_key_values_length = 0
+
+        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
+
+        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()
+
+        if inputs_embeds is None:
+            inputs_embeds = self.embed_tokens(input_ids)
+
+        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 Phi3. Make sure to "
+                    " call `tokenizer.padding_side  = 'left'` before tokenizing the input. "
+                )
+
+        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
+        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,
+            )
+
+        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
+
+        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_states = layer_outputs[0]
+
+            if use_cache:
+                next_decoder_cache = layer_outputs[2 if output_attentions else 1]
+
+            if output_attentions:
+                all_self_attns += (layer_outputs[1],)
+
+        hidden_states = self.norm(hidden_states)
+
+        # add hidden states from the last decoder layer
+        if output_hidden_states:
+            all_hidden_states += (hidden_states,)
+
+        next_cache = None
+        if use_cache:
+            next_cache = next_decoder_cache.to_legacy_cache() if 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 Phi3ForCausalLM(Phi3PreTrainedModel):
+    _tied_weights_keys = ["lm_head.weight"]
+
+    # Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.__init__ with Llama->Phi3
+    def __init__(self, config):
+        super().__init__(config)
+        self.model = Phi3Model(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()
+
+    # Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.get_input_embeddings
+    def get_input_embeddings(self):
+        return self.model.embed_tokens
+
+    # Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.set_input_embeddings
+    def set_input_embeddings(self, value):
+        self.model.embed_tokens = value
+
+    # Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.get_output_embeddings
+    def get_output_embeddings(self):
+        return self.lm_head
+
+    # Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.set_output_embeddings
+    def set_output_embeddings(self, new_embeddings):
+        self.lm_head = new_embeddings
+
+    # Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.set_decoder
+    def set_decoder(self, decoder):
+        self.model = decoder
+
+    # Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.get_decoder
+    def get_decoder(self):
+        return self.model
+
+    # Ignore copy
+    @add_start_docstrings_to_model_forward(PHI3_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, Phi3ForCausalLM
+
+        >>> model = Phi3ForCausalLM.from_pretrained("microsoft/phi-3-mini-4k-instruct")
+        >>> tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-3-mini-4k-instruct")
+
+        >>> prompt = "This is an example script ."
+        >>> 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]
+        'This is an example script .\n Certainly! Below is a sample script that demonstrates a simple task, such as calculating the sum'
+        ```"""
+
+        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
+            loss_fct = CrossEntropyLoss()
+            shift_logits = shift_logits.view(-1, self.config.vocab_size)
+            shift_labels = shift_labels.view(-1)
+            # Enable model parallelism
+            shift_labels = shift_labels.to(shift_logits.device)
+            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,
+        )
+
+    # Copied from transformers.models.persimmon.modeling_persimmon.PersimmonForCausalLM.prepare_inputs_for_generation
+    def prepare_inputs_for_generation(
+        self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs
+    ):
+        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
+    # Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM._reorder_cache
+    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 [`Phi3Model`] with a sequence classification head on top (linear layer).
+
+    [`Phi3ForSequenceClassification`] 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).
+    """,
+    PHI3_START_DOCSTRING,
+)
+# Copied from transformers.models.llama.modeling_llama.LlamaForSequenceClassification with Llama->Phi3, LLAMA->PHI3, self.transformer->self.model, transformer_outputs->model_outputs
+class Phi3ForSequenceClassification(Phi3PreTrainedModel):
+    def __init__(self, config):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+        self.model = Phi3Model(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(PHI3_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,
+        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
+
+        model_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 = model_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,) + model_outputs[1:]
+            return ((loss,) + output) if loss is not None else output
+
+        return SequenceClassifierOutputWithPast(
+            loss=loss,
+            logits=pooled_logits,
+            past_key_values=model_outputs.past_key_values,
+            hidden_states=model_outputs.hidden_states,
+            attentions=model_outputs.attentions,
+        )
+
+
+@add_start_docstrings(
+    """
+    [`Phi3Model`] with a token classification head on top (a linear layer on top of the hidden-states output) e.g. for
+    Named-Entity-Recognition (NER) tasks.
+    """,
+    PHI3_START_DOCSTRING,
+)
+# Copied from transformers.models.mpt.modeling_mpt.MptForTokenClassification with Mpt->Phi3,MPT->PHI3,self.transformer->self.model,transformer_outputs->model_outputs
+class Phi3ForTokenClassification(Phi3PreTrainedModel):
+    def __init__(self, config: Phi3Config):
+        super().__init__(config)
+        self.num_labels = config.num_labels
+
+        self.model = Phi3Model(config)
+        if hasattr(config, "classifier_dropout") and config.classifier_dropout is not None:
+            classifier_dropout = config.classifier_dropout
+        elif hasattr(config, "hidden_dropout") and config.hidden_dropout is not None:
+            classifier_dropout = config.hidden_dropout
+        else:
+            classifier_dropout = 0.1
+        self.dropout = nn.Dropout(classifier_dropout)
+        self.classifier = nn.Linear(config.hidden_size, config.num_labels)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    @add_start_docstrings_to_model_forward(PHI3_INPUTS_DOCSTRING)
+    @add_code_sample_docstrings(
+        checkpoint=_CHECKPOINT_FOR_DOC,
+        output_type=TokenClassifierOutput,
+        config_class=_CONFIG_FOR_DOC,
+    )
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.Tensor, torch.Tensor], ...]] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
+        labels: Optional[torch.Tensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        **deprecated_arguments,
+    ) -> Union[Tuple[torch.Tensor], TokenClassifierOutput]:
+        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
+
+        model_outputs = self.model(
+            input_ids,
+            past_key_values=past_key_values,
+            attention_mask=attention_mask,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        hidden_states = model_outputs[0]
+        hidden_states = self.dropout(hidden_states)
+        logits = self.classifier(hidden_states)
+
+        loss = None
+        if labels is not None:
+            # move labels to correct device to enable model parallelism
+            labels = labels.to(logits.device)
+            batch_size, seq_length = labels.shape
+            loss_fct = CrossEntropyLoss()
+            loss = loss_fct(
+                logits.view(batch_size * seq_length, self.num_labels), labels.view(batch_size * seq_length)
+            )
+
+        if not return_dict:
+            output = (logits,) + model_outputs[2:]
+            return ((loss,) + output) if loss is not None else output
+
+        return TokenClassifierOutput(
+            loss=loss,
+            logits=logits,
+            hidden_states=model_outputs.hidden_states,
+            attentions=model_outputs.attentions,
+        )