Update modeling_mobillama.py

modeling_mobillama.py

@@ -1,10 +1,10 @@
 # coding=utf-8
 # Copyright 2022 EleutherAI and the HuggingFace Inc. team. All rights reserved.
+# Copyright 2023 MobiLLaMA team.
+# Copyright 2023 vLLM team.
 #
 # 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.
+# and OPT implementations in this library. It has been modified for vLLM's model execution architecture.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -17,852 +17,337 @@
 # 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 LLaMA model."""
-import math
-from typing import List, Optional, Tuple, Union
+"""vLLM implementation of MobiLLaMA model."""
+
+from typing import Dict, List, Optional, Tuple
+from vllm.config import CacheConfig
+from vllm.model_executor.layers.quantization import QuantizationConfig
 
 import torch
-import torch.utils.checkpoint
 from torch import nn
-from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
-
-from transformers.activations import ACT2FN
-from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast, SequenceClassifierOutputWithPast
-from transformers.modeling_utils import PreTrainedModel
-from transformers.utils import add_start_docstrings, add_start_docstrings_to_model_forward, logging, replace_return_docstrings
-from transformers.models.llama.configuration_llama import LlamaConfig
-
-# from .configuration_mobillama import MobiLlamaConfig
-
-from flash_attn import flash_attn_func
-
-
-logger = logging.get_logger(__name__)
-
-_CONFIG_FOR_DOC = "LlamaConfig"
-
-
-# Copied from transformers.models.bart.modeling_bart._make_causal_mask
-def _make_causal_mask(
-    input_ids_shape: torch.Size, dtype: torch.dtype, device: torch.device, past_key_values_length: int = 0
-):
-    """
-    Make causal mask used for bi-directional self-attention.
-    """
-    bsz, tgt_len = input_ids_shape
-    mask = torch.full((tgt_len, tgt_len), torch.tensor(torch.finfo(dtype).min, device=device), device=device)
-    mask_cond = torch.arange(mask.size(-1), device=device)
-    mask.masked_fill_(mask_cond < (mask_cond + 1).view(mask.size(-1), 1), 0)
-    mask = mask.to(dtype)
-
-    if past_key_values_length > 0:
-        mask = torch.cat([torch.zeros(tgt_len, past_key_values_length, dtype=dtype, device=device), mask], dim=-1)
-    return mask[None, None, :, :].expand(bsz, 1, tgt_len, tgt_len + past_key_values_length)
-
-
-# Copied from transformers.models.bart.modeling_bart._expand_mask
-def _expand_mask(mask: torch.Tensor, dtype: torch.dtype, tgt_len: Optional[int] = None):
-    """
-    Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.
-    """
-    bsz, src_len = mask.size()
-    tgt_len = tgt_len if tgt_len is not None else src_len
+from transformers import LlamaConfig
 
-    expanded_mask = mask[:, None, None, :].expand(bsz, 1, tgt_len, src_len).to(dtype)
+from vllm.model_executor.layers.activation import SiluAndMul
+from vllm.attention import Attention
+from vllm.model_executor.layers.linear import (QKVParallelLinear, RowParallelLinear,
+                               ColumnParallelLinear)
+from vllm.model_executor.layers.layernorm import RMSNorm
+from vllm.model_executor.layers.rotary_embedding import get_rope
+from vllm.model_executor.layers.sampler import Sampler
 
-    inverted_mask = 1.0 - expanded_mask
-
-    return inverted_mask.masked_fill(inverted_mask.to(torch.bool), torch.finfo(dtype).min)
-
-
-class MobiLlamaRMSNorm(nn.Module):
-    def __init__(self, hidden_size, eps=1e-6):
-        """
-        MobiLlamaRMSNorm 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
-        variance = hidden_states.to(torch.float32).pow(2).mean(-1, keepdim=True)
-        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
-
-        return (self.weight * hidden_states).to(input_dtype)
-
-
-class MobiLlamaRotaryEmbedding(torch.nn.Module):
-    def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):
-        super().__init__()
-        inv_freq = 1.0 / (base ** (torch.arange(0, dim, 2).float().to(device) / dim))
-        self.register_buffer("inv_freq", inv_freq)
-
-        # Build here to make `torch.jit.trace` work.
-        self.max_seq_len_cached = max_position_embeddings
-        t = torch.arange(self.max_seq_len_cached, device=self.inv_freq.device, dtype=self.inv_freq.dtype)
-        freqs = torch.einsum("i,j->ij", t, self.inv_freq)
-        # Different from paper, but it uses a different permutation in order to obtain the same calculation
-        emb = torch.cat((freqs, freqs), dim=-1)
-        self.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):
-        # x: [bs, num_attention_heads, seq_len, head_size]
-        # This `if` block is unlikely to be run after we build sin/cos in `__init__`. Keep the logic here just in case.
-        if seq_len > self.max_seq_len_cached:
-            self.max_seq_len_cached = seq_len
-            t = torch.arange(self.max_seq_len_cached, device=x.device, dtype=self.inv_freq.dtype)
-            freqs = torch.einsum("i,j->ij", t, self.inv_freq)
-            # Different from paper, but it uses a different permutation in order to obtain the same calculation
-            emb = torch.cat((freqs, freqs), dim=-1).to(x.device)
-            self.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),
-        )
-
-
-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):
-    # The first two dimensions of cos and sin are always 1, so we can `squeeze` them.
-    cos = cos.squeeze(1).squeeze(0)  # [seq_len, dim]
-    sin = sin.squeeze(1).squeeze(0)  # [seq_len, dim]
-    cos = cos[position_ids].unsqueeze(1)  # [bs, 1, seq_len, dim]
-    sin = sin[position_ids].unsqueeze(1)  # [bs, 1, seq_len, dim]
-    q_embed = (q * cos) + (rotate_half(q) * sin)
-    k_embed = (k * cos) + (rotate_half(k) * sin)
-    return q_embed, k_embed
+from vllm.model_executor.sampling_metadata import SamplingMetadata
+from vllm.model_executor.model_loader.weight_utils import (default_weight_loader,
+                                 pt_weights_iterator)
+from vllm.model_executor.layers.sampler import SamplerOutput
 
+class MobiLlamaAttention(nn.Module):
+    """Multi-headed attention from the paper 'Attention Is All You Need'"""
 
-class MobiLlamaMLP(nn.Module):
     def __init__(
         self,
-        hidden_size: int,
-        intermediate_size: int,
-        hidden_act: str,
+        config: LlamaConfig,
+        layer_idx: int,
     ):
-        super().__init__()
-        self.gate_proj = nn.Linear(hidden_size, intermediate_size, bias=False)
-        self.down_proj = nn.Linear(intermediate_size, hidden_size, bias=False)
-        self.up_proj = nn.Linear(hidden_size, intermediate_size, bias=False)
-        self.act_fn = ACT2FN[hidden_act]
-
-    def forward(self, x):
-        return self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
-
-
-class MobiLlamaAttention(nn.Module):
-    """Multi-headed attention from 'Attention Is All You Need' paper"""
-
-    def __init__(self, config: LlamaConfig):
         super().__init__()
         self.config = config
+        self.layer_idx = layer_idx
+        
         self.hidden_size = config.hidden_size
         self.num_heads = config.num_attention_heads
         self.head_dim = self.hidden_size // self.num_heads
         self.max_position_embeddings = config.max_position_embeddings
+        self.scaling = self.head_dim**-0.5
 
         if (self.head_dim * self.num_heads) != self.hidden_size:
             raise ValueError(
                 f"hidden_size must be divisible by num_heads (got `hidden_size`: {self.hidden_size}"
                 f" and `num_heads`: {self.num_heads})."
             )
-        self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=False)
-        self.k_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=False)
-        self.v_proj = nn.Linear(self.hidden_size, self.num_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 = MobiLlamaRotaryEmbedding(self.head_dim, max_position_embeddings=self.max_position_embeddings)
-
-    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()
+        
+        self.num_key_value_heads = getattr(config, "num_key_value_heads", self.num_heads)
+        self.num_key_value_groups = self.num_heads // self.num_key_value_heads
+        
+        # In vLLM implementation, we use combined QKV projection
+        self.qkv_proj = QKVParallelLinear(
+            self.hidden_size,
+            self.head_dim,
+            self.num_heads,
+            self.num_key_value_heads,
+            bias=False,
+        )
+        
+        self.o_proj = RowParallelLinear(
+            self.hidden_size,
+            self.hidden_size,
+            bias=False,
+        )
+        
+        # Set up rotary embedding
+        rope_theta = getattr(config, "rope_theta", 10000)
+        max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
+        self.rotary_emb = get_rope(
+            self.head_dim,
+            rotary_dim=self.head_dim,
+            max_position=max_position_embeddings,
+            base=rope_theta,
+        )
+        
+        self.attn = Attention(
+            self.num_heads,
+            self.head_dim,
+            self.scaling,
+            num_kv_heads=self.num_key_value_heads
+        )
 
     def forward(
         self,
+        positions: torch.Tensor,
         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: bool = False,
-        use_cache: bool = False,
-    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
-        bsz, q_len, _ = hidden_states.size()
-
-        query_states = self.q_proj(hidden_states).view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
-        key_states = self.k_proj(hidden_states).view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
-        value_states = self.v_proj(hidden_states).view(bsz, q_len, self.num_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[0].shape[-2]
-        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)
-        # [bsz, nh, t, hd]
-
-        if past_key_value is not None:
-            # reuse k, v, self_attention
-            key_states = torch.cat([past_key_value[0], key_states], dim=2)
-            value_states = torch.cat([past_key_value[1], value_states], dim=2)
-
-        past_key_value = (key_states, value_states) if use_cache else None
-
-        attn_output = flash_attn_func(
-            q=query_states.transpose(1, 2).to(torch.bfloat16),
-            k=key_states.transpose(1, 2).to(torch.bfloat16),
-            v=value_states.transpose(1, 2).to(torch.bfloat16),
-            causal=True)
-        attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
-        attn_output = attn_output.to(query_states.dtype)
-
+        kv_cache: torch.Tensor,
+        attn_metadata: Dict,
+    ) -> torch.Tensor:
+        qkv = self.qkv_proj(hidden_states)
+        q, k, v = qkv.split([
+            self.num_heads * self.head_dim,
+            self.num_key_value_heads * self.head_dim,
+            self.num_key_value_heads * self.head_dim
+        ], dim=-1)
+        
+        # Reshape for rotary embedding
+        q = q.view(-1, self.num_heads, self.head_dim)
+        k = k.view(-1, self.num_key_value_heads, self.head_dim)
+        v = v.view(-1, self.num_key_value_heads, self.head_dim)
+        
+        # Apply rotary embedding
+        q, k = self.rotary_emb(positions, q, k)
+        
+        # Run attention
+        attn_output = self.attn(q, k, v, kv_cache, attn_metadata)
+        
+        # Reshape output and project back to hidden size
+        attn_output = attn_output.reshape(*hidden_states.shape[:-1], self.hidden_size)
         attn_output = self.o_proj(attn_output)
+        
+        return attn_output
 
-        # if not output_attentions:
-        #     attn_weights = None
-        assert not output_attentions
-        attn_weights = None
-
-        return attn_output, attn_weights, past_key_value
+class MobiLlamaMLP(nn.Module):
+    def __init__(
+        self,
+        config: LlamaConfig,
+    ):
+        super().__init__()
+        hidden_size = config.hidden_size
+        intermediate_size = config.intermediate_size
+        
+        # In vLLM, we use ColumnParallelLinear for gate_proj and up_proj
+        self.gate_proj = ColumnParallelLinear(
+            hidden_size, 
+            intermediate_size, 
+            bias=False,
+        )
+        self.up_proj = ColumnParallelLinear(
+            hidden_size, 
+            intermediate_size, 
+            bias=False,
+        )
+        self.down_proj = RowParallelLinear(
+            intermediate_size, 
+            hidden_size, 
+            bias=False,
+        )
+        self.act_fn = SiluAndMul()
 
+    def forward(self, x):
+        gate_output = self.gate_proj(x)
+        up_output = self.up_proj(x)
+        
+        # Apply SiLU activation and multiply
+        intermediate_output = self.act_fn(gate_output, up_output)
+        
+        # Project back to hidden size
+        output = self.down_proj(intermediate_output)
+        return output
 
 class MobiLlamaDecoderLayer(nn.Module):
-    def __init__(self, config: LlamaConfig):
+    def __init__(
+        self,
+        config: LlamaConfig,
+        layer_idx: int,
+    ):
         super().__init__()
         self.hidden_size = config.hidden_size
-        self.self_attn = MobiLlamaAttention(config=config)
-        self.mlp = MobiLlamaMLP(
-            hidden_size=self.hidden_size,
-            intermediate_size=config.intermediate_size,
-            hidden_act=config.hidden_act,
-        )
-        self.input_layernorm = MobiLlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
-        self.post_attention_layernorm = MobiLlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        
+        # Layer norms
+        self.input_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.post_attention_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        
+        # Self-attention
+        self.self_attn = MobiLlamaAttention(config=config, layer_idx=layer_idx)
+        
+        # MLP
+        self.mlp = MobiLlamaMLP(config)
 
     def forward(
         self,
+        positions: torch.Tensor,
         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,
-    ) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
-        """
-        Args:
-            hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)`
-            attention_mask (`torch.FloatTensor`, *optional*): attention mask of size
-                `(batch, 1, tgt_len, src_len)` where padding elements are indicated by very large negative 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.
-            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
-        """
-
+        kv_cache: List[torch.Tensor],
+        attn_metadata: Dict,
+    ) -> torch.Tensor:
+        # Layernorm before self-attention
         residual = hidden_states
-
         hidden_states = self.input_layernorm(hidden_states)
-
-        # Self Attention
-        hidden_states, self_attn_weights, present_key_value = self.self_attn(
+        
+        # Self-attention
+        hidden_states = self.self_attn(
+            positions=positions,
             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,
+            kv_cache=kv_cache[0],
+            attn_metadata=attn_metadata,
         )
+        
+        # First residual connection
         hidden_states = residual + hidden_states
-
-        # Fully Connected
+        
+        # Layernorm before MLP
         residual = hidden_states
         hidden_states = self.post_attention_layernorm(hidden_states)
+        
+        # MLP
         hidden_states = self.mlp(hidden_states)
+        
+        # Second residual connection
         hidden_states = residual + hidden_states
+        
+        return hidden_states
 
-        outputs = (hidden_states,)
-
-        if output_attentions:
-            outputs += (self_attn_weights,)
-
-        if use_cache:
-            outputs += (present_key_value,)
-
-        return outputs
-
-
-MOBILLAMA_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 ([`LlamaConfig`]):
-            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 LLaMA Model outputting raw hidden-states without any specific head on top.",
-    MOBILLAMA_START_DOCSTRING,
-)
-class MobiLlamaPreTrainedModel(PreTrainedModel):
-    config_class = LlamaConfig
-    base_model_prefix = "model"
-    supports_gradient_checkpointing = True
-    _no_split_modules = ["MobiLlamaDecoderLayer"]
-    _skip_keys_device_placement = "past_key_values"
-    _keys_to_ignore_on_load_unexpected = [r"decoder\.version"]
-
-    def _init_weights(self, module):
-        std = self.config.initializer_range
-        if isinstance(module, nn.Linear):
-            module.weight.data.normal_(mean=0.0, std=std)
-            if module.bias is not None:
-                module.bias.data.zero_()
-        elif isinstance(module, nn.Embedding):
-            module.weight.data.normal_(mean=0.0, std=std)
-            if module.padding_idx is not None:
-                module.weight.data[module.padding_idx].zero_()
-
-    def _set_gradient_checkpointing(self, module, value=False):
-        if isinstance(module, MobiLlamaModel):
-            module.gradient_checkpointing = value
-
-
-MOBILLAMA_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 (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
-            Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of shape
-            `(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional tensors of shape
-            `(batch_size, num_heads, encoder_sequence_length, embed_size_per_head)`.
-
-            Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
-            blocks) that can be used (see `past_key_values` input) to speed up sequential decoding.
-
-            If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that
-            don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all
-            `decoder_input_ids` of shape `(batch_size, sequence_length)`.
-        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 LLaMA Model outputting raw hidden-states without any specific head on top.",
-    MOBILLAMA_START_DOCSTRING,
-)
-class MobiLlamaModel(MobiLlamaPreTrainedModel):
-    """
-    Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`MobiLlamaDecoderLayer`]
-
-    Args:
-        config: LlamaConfig
-    """
-
+class MobiLlamaModel(nn.Module):
     def __init__(self, config: LlamaConfig):
-        super().__init__(config)
+        super().__init__()
+        self.config = config
         self.padding_idx = config.pad_token_id
         self.vocab_size = config.vocab_size
+        
+        # Token embedding
+        self.embed_tokens = nn.Embedding(
+            config.vocab_size, 
+            config.hidden_size, 
+            self.padding_idx
+        )
+        
+        # Decoder layers
+        self.layers = nn.ModuleList([
+            MobiLlamaDecoderLayer(config, i) 
+            for i in range(config.num_hidden_layers)
+        ])
+        
+        # Final layernorm
+        self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
 
-        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
-        self.layers = nn.ModuleList([MobiLlamaDecoderLayer(config) for _ in range(config.num_hidden_layers)])
-        self.norm = MobiLlamaRMSNorm(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
-
-    # Copied from transformers.models.bart.modeling_bart.BartDecoder._prepare_decoder_attention_mask
-    def _prepare_decoder_attention_mask(self, attention_mask, input_shape, inputs_embeds, past_key_values_length):
-        # create causal mask
-        # [bsz, seq_len] -> [bsz, 1, tgt_seq_len, src_seq_len]
-        combined_attention_mask = None
-        if input_shape[-1] > 1:
-            combined_attention_mask = _make_causal_mask(
-                input_shape,
-                inputs_embeds.dtype,
-                device=inputs_embeds.device,
-                past_key_values_length=past_key_values_length,
-            )
-
-        if attention_mask is not None:
-            # [bsz, seq_len] -> [bsz, 1, tgt_seq_len, src_seq_len]
-            expanded_attn_mask = _expand_mask(attention_mask, inputs_embeds.dtype, tgt_len=input_shape[-1]).to(
-                inputs_embeds.device
-            )
-            combined_attention_mask = (
-                expanded_attn_mask if combined_attention_mask is None else expanded_attn_mask + combined_attention_mask
-            )
-
-        return combined_attention_mask
-
-    @add_start_docstrings_to_model_forward(MOBILLAMA_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")
-
-        seq_length_with_past = seq_length
-        past_key_values_length = 0
-
-        if past_key_values is not None:
-            past_key_values_length = past_key_values[0][0].shape[2]
-            seq_length_with_past = seq_length_with_past + past_key_values_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)
-        # embed positions
-        if attention_mask is None:
-            attention_mask = torch.ones(
-                (batch_size, seq_length_with_past), dtype=torch.bool, device=inputs_embeds.device
+        input_ids: torch.Tensor,
+        positions: torch.Tensor,
+        kv_caches: List[List[torch.Tensor]],
+        attn_metadata: Dict,
+    ) -> torch.Tensor:
+        # Get token embeddings
+        hidden_states = self.embed_tokens(input_ids)
+        
+        # Forward through each decoder layer
+        for i, layer in enumerate(self.layers):
+            hidden_states = layer(
+                positions=positions,
+                hidden_states=hidden_states,
+                kv_cache=kv_caches[i],
+                attn_metadata=attn_metadata,
             )
-        attention_mask = self._prepare_decoder_attention_mask(
-            attention_mask, (batch_size, seq_length), inputs_embeds, past_key_values_length
-        )
-
-        hidden_states = 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
-
-        # decoder layers
-        all_hidden_states = () if output_hidden_states else None
-        all_self_attns = () if output_attentions else None
-        next_decoder_cache = () if use_cache else None
-
-        for idx, decoder_layer in enumerate(self.layers):
-            if output_hidden_states:
-                all_hidden_states += (hidden_states,)
-
-            past_key_value = past_key_values[idx] if past_key_values is not None else None
-
-            if self.gradient_checkpointing and self.training:
-
-                def create_custom_forward(module):
-                    def custom_forward(*inputs):
-                        # None for past_key_value
-                        return module(*inputs, output_attentions, None)
-
-                    return custom_forward
-
-                layer_outputs = torch.utils.checkpoint.checkpoint(
-                    create_custom_forward(decoder_layer),
-                    hidden_states,
-                    attention_mask,
-                    position_ids,
-                    None,
-                )
-            else:
-                layer_outputs = decoder_layer(
-                    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 = 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],)
-
+        
+        # Apply final layernorm
         hidden_states = self.norm(hidden_states)
+        
+        return hidden_states
 
-        # add hidden states from the last decoder layer
-        if output_hidden_states:
-            all_hidden_states += (hidden_states,)
-
-        next_cache = next_decoder_cache if use_cache else None
-        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 LlamaForCausalLM(MobiLlamaPreTrainedModel):
-    def __init__(self, config):
-        super().__init__(config)
+class MobiLlamaForCausalLM(nn.Module):
+    def __init__(self, config: LlamaConfig, quant_config: Optional[QuantizationConfig] = None,cache_config: Optional[CacheConfig] = None):
+        super().__init__()
+        self.config = config
+        
+        # Core MobiLLaMA model
         self.model = MobiLlamaModel(config)
+        
+        # LM head
+        self.lm_head = nn.Linear(
+            config.hidden_size, 
+            config.vocab_size, 
+            bias=False
+        )
+        
+        # Sampling module
+        self.sampler = Sampler()
 
-        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
-
-        # Initialize weights and apply final processing
-        self.post_init()
-
-    def get_input_embeddings(self):
-        return self.model.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
-
-    @add_start_docstrings_to_model_forward(MOBILLAMA_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, LlamaForCausalLM
-
-        >>> model = LlamaForCausalLM.from_pretrained(PATH_TO_CONVERTED_WEIGHTS)
-        >>> tokenizer = AutoTokenizer.from_pretrained(PATH_TO_CONVERTED_TOKENIZER)
-
-        >>> prompt = "Hey, are you consciours? 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 consciours? Can you talk to me?\nI'm not consciours, 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: torch.Tensor,
+        positions: torch.Tensor,
+        kv_caches: List[List[torch.Tensor]],
+        attn_metadata: Dict,
+    ) -> torch.Tensor:
+        hidden_states = 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,
+            positions=positions,
+            kv_caches=kv_caches,
+            attn_metadata=attn_metadata,
         )
-
-        hidden_states = outputs[0]
+        
+        # Apply LM head
         logits = self.lm_head(hidden_states)
+        
+        return logits
 
-        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,
-        )
-
-    def prepare_inputs_for_generation(
-        self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs
-    ):
-        if past_key_values:
-            input_ids = input_ids[:, -1:]
-
-        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[:, -1].unsqueeze(-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) for past_state in layer_past),)
-        return reordered_past
-
-
-@add_start_docstrings(
-    """
-    The LLaMa Model transformer with a sequence classification head on top (linear layer).
-
-    [`MobiLlamaForSequenceClassification`] 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).
-    """,
-    MOBILLAMA_START_DOCSTRING,
-)
-class MobiLlamaForSequenceClassification(MobiLlamaPreTrainedModel):
-    _keys_to_ignore_on_load_missing = [r"lm_head.weight"]
-
-    def __init__(self, config):
-        super().__init__(config)
-        self.num_labels = config.num_labels
-        self.model = MobiLlamaModel(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(MOBILLAMA_INPUTS_DOCSTRING)
-    def forward(
+    def sample(
         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
-
-        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:
-                sequence_lengths = (torch.ne(input_ids, self.config.pad_token_id).sum(-1) - 1).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,
-        )
+        logits: torch.Tensor,
+        sampling_metadata: SamplingMetadata,
+    ) -> SamplerOutput:
+        return self.sampler(logits, sampling_metadata)
+
+    def load_weights(self, weights):
+        # First use default loader for most weights
+        state_dict = self.state_dict()
+        for name, param in pt_weights_iterator(weights):
+            if "rotary_emb" in name:
+                # Skip rotary embedding weights as they're handled differently in vLLM
+                continue
+            
+            # vLLM uses a combined QKV projection
+            if any(n in name for n in ["q_proj", "k_proj", "v_proj"]):
+                # These weights will be loaded separately through the QKVParallelLinear
+                continue
+            
+            param_name = name
+            
+            # Handle mapping between HF and vLLM naming schemes
+            if "self_attn.o_proj" in name:
+                param_name = name.replace("self_attn.o_proj", "self_attn.o_proj.weight")
+            elif "mlp.gate_proj" in name:
+                param_name = name.replace("mlp.gate_proj", "mlp.gate_proj.weight")
+            elif "mlp.up_proj" in name:
+                param_name = name.replace("mlp.up_proj", "mlp.up_proj.weight")
+            elif "mlp.down_proj" in name:
+                param_name = name.replace("mlp.down_proj", "mlp.down_proj.weight")
+            
+            if param_name in state_dict:
+                state_dict[param_name].copy_(param)
+        
+        # Separately handle the QKV projections to combine them
+        for idx, layer in enumerate(self.model.layers):
+            # Get weights for q_proj, k_proj, v_proj
+            q_weight = weights[f"model.layers.{idx}.self_attn.q_proj.weight"]
+            k_weight = weights[f"model.layers.{idx}.self_attn.k_proj.weight"]
+            v_weight = weights[f"model.layers.{idx}.self_attn.v_proj.weight"]
+            
+            # Set the combined QKV weight
+            layer.self_attn.qkv_proj.weight.data.copy_(
+                torch.cat([q_weight, k_weight, v_weight], dim=0)
+            )