add qwen3_moe

2026-02-10 18:30:48 +08:00 · 2026-02-10 18:22:13 +08:00 · 2026-02-10 18:18:32 +08:00 · 2026-02-10 18:09:58 +08:00 · 2026-02-10 18:02:40 +08:00 · 2026-02-10 16:10:28 +08:00
20 changed files with 100 additions and 1671 deletions
--- a/README.md
+++ b/README.md
@@ -1,28 +1,12 @@
 # enginex-mlu370-vllm

-寒武纪 MLU370（X8/X4）加速卡上基于 vLLM 推理引擎的文本生成框架。
+# 寒武纪 mlu370 文本生成
+该模型测试框架在寒武纪mlu370 （X8/X4）加速卡上，基于vllm 推理引擎，适配了 Qwen1.5-1.8B-Chat 模型。

-## 版本更新记录

-**v0.0.6.2** — 2026-02-11 · Llama4 模型支持，含 sigmoid routing MoE、QK Norm、交替 dense/MoE 层；由于 MLU370（capability=3）限制，MoE 改为 dense 模式解决 graph capture 兼容性
-
-**v0.0.6.1** — 2026-02-11 · DeepSeek V3 MTP 推测解码，新建 MTP draft model 复用 DeepseekV2DecoderLayer，自动检测并启用 MTP speculative decoding
-
-**v0.0.6** — 2026-02-11 · DeepSeek V3 模型支持，复用 V2 实现，新增 `noaux_tc` 路由，修复 MLA unpaged 缓存算子
-
-**v0.0.5** — 2026-02-10 · Qwen3MoE 模型支持，修复 FusedMoE `forward_mlu` 签名 bug
-
-**v0.0.4.1** — 2026-02-10 · Gemma3 rope 兼容性修复，适配 MLU rotary_emb 接口
-
-**v0.0.4** — 2026-02-10 · Gemma3 模型支持，含 QK Norm、per-layer rope、滑动窗口
-
-**v0.0.3.1** — 2026-02-06 · CNNL Tensor 溢出修复，KV cache 元素数 int32 上限防护
-
-**v0.0.3** — 2026-02-06 · Transformers 通用后端，支持 `auto_map` 加载自定义 HF 模型
-
-**v0.0.2** — 2026-02-04 · Qwen3 模型支持，QK Norm 适配，修复 rope/tokenizer 兼容性
-
---
+* Qwen1.5-1.8B-Chat 是通义千问系列中一款约18亿参数、轻量级的中英文对话大模型，专为高效推理和多场景聊天交互设计。
+* Llama-2-7b-chat-hf：Meta 发布的 LLaMA 2 系列中 70 亿参数的对话优化版开源大模型，适合多轮聊天与通用任务。
+* ChatGLM3-6B：智谱 AI 推出的第 3 代 ChatGLM 系列中 60 亿参数的中英双语对话大模型，支持推理、代码和多任务能力。

 ## Quick Start
 1. 首先从modelscope上下载文本生成大模型，如`Qwen1.5-1.8B-Chat`  
@@ -181,3 +165,13 @@ curl http://localhost:80/v1/chat/completions \
 | ------------------- | ------------------- | -------------------| ------------------- | ------------------- | ------------------- | -------------------  | ------------------- | -------------------  |
 | Qwen/Qwen-1_8B   |0.203	| 13493.2	 |	119.2 |	10.0 | 0.052 | 25591.5 |	165.0 |	15.0|
 | Qwen/Qwen1.5-0.5B   |0.132	| 12366.6	 |	106.9 |	15.0 | 0.066 | 24935.4 |	151.4 |	10.0|
+
+## 版本更新记录
+
+| 版本 | 日期 | 更新内容 |
+|------|------|----------|
+| v0.0.2 | 2026-02-04 | **Qwen3 模型支持**：实现 QK Normalization 架构适配，修复 rope_scaling 和 tokenizer 兼容性问题，解决张量连续性导致的 view 操作失败 |
+| v0.0.3 | 2026-02-06 | **Transformers 通用后端**：支持通过 `auto_map` 加载任意自定义 HuggingFace 模型，新增 registry 回退逻辑、Linear 返回值处理、RMSNorm 维度恢复等 |
+| v0.0.3.1 | 2026-02-06 | **CNNL Tensor 溢出修复**：解决极小模型在大显存设备上部署时 KV cache 元素数超过 int32 限制的问题，在 mlu_worker 和 cache_engine 中添加双重防护 |
+| v0.0.4 | 2026-02-10 | **Gemma3 模型支持**：新增 Gemma3ForCausalLM 模型实现（含 QK Normalization、per-layer rope 配置、layer_types 滑动窗口），修复 `patch_rope_scaling_dict` 在 rope_scaling 缺少 `rope_type` 键时崩溃的问题，更新模型注册表及 config.py 中 interleaved attention 和 dtype 自动处理逻辑 |
+| v0.0.4.1 | 2026-02-10 | **Gemma3 rope 兼容性修复**：修复新版 transformers `Gemma3TextConfig` 缺少 `rope_theta` 属性的问题，从 `rope_parameters` 字典兼容提取 rope 配置（支持 Transformers v4/v5）；修复 `rope_scaling` 嵌套字典导致 `get_rope` 缓存 unhashable 的问题；适配 MLU `forward_mlu` 接口，将 q/k 合并为单张量调用 rotary_emb 后再拆分 |
--- a/vllm-v0.6.2/vllm/config.py
+++ b/vllm-v0.6.2/vllm/config.py
@@ -1403,18 +1403,6 @@ class SpeculativeConfig:

            draft_hf_config = draft_model_config.hf_config

-            # Detect DeepSeek V3 MTP: same model path with
-            # num_nextn_predict_layers > 0
-            num_nextn = getattr(draft_hf_config,
-                                "num_nextn_predict_layers", 0)
-            if (num_nextn and num_nextn > 0
-                    and getattr(draft_hf_config, "model_type", "")
-                    in ("deepseek_v3",)):
-                draft_hf_config.model_type = "deepseek_mtp"
-                draft_hf_config.architectures = ["DeepSeekMTPModel"]
-                if num_speculative_tokens is None:
-                    num_speculative_tokens = num_nextn
-
            if (num_speculative_tokens is not None
                    and hasattr(draft_hf_config, "num_lookahead_tokens")):
                draft_hf_config.num_lookahead_tokens = num_speculative_tokens
@@ -1433,7 +1421,7 @@ class SpeculativeConfig:
                        f"{num_speculative_tokens=} was provided.")

            if enable_chunked_prefill and draft_hf_config.model_type in (
-                    "medusa", "mlp_speculator", "eagle", "deepseek_mtp"):
+                    "medusa", "mlp_speculator", "eagle"):
                raise ValueError(
                    "Chunked prefill and hidden-state based draft models are "
                    "not compatible.")
--- a/vllm-v0.6.2/vllm/model_executor/layers/layernorm.py
+++ b/vllm-v0.6.2/vllm/model_executor/layers/layernorm.py
@@ -143,14 +143,11 @@ class RMSNorm(CustomOp):
        from vllm import _mlu_ops as mlu_ops

        x = x.view(-1, self.weight.data.shape[0])
-        weight = self.weight.data
-        if weight.dtype != x.dtype:
-            weight = weight.to(x.dtype)
        if residual is not None:
            residual = residual.view(-1, self.weight.data.shape[0])
-            return mlu_ops.fused_rms_norm(x, residual, weight, None, None, self.variance_epsilon, True)
+            return mlu_ops.fused_rms_norm(x, residual, self.weight.data, None, None, self.variance_epsilon, True)
        else:
-            return mlu_ops.fused_rms_norm(x, residual, weight, None, None, self.variance_epsilon, False)
+            return mlu_ops.fused_rms_norm(x, residual, self.weight.data, None, None, self.variance_epsilon, False)

    def extra_repr(self) -> str:
        s = f"hidden_size={self.weight.data.size(0)}"
--- a/vllm-v0.6.2/vllm/model_executor/layers/vocab_parallel_embedding.py
+++ b/vllm-v0.6.2/vllm/model_executor/layers/vocab_parallel_embedding.py
@@ -38,9 +38,6 @@ class UnquantizedEmbeddingMethod(QuantizeMethodBase):
              layer: torch.nn.Module,
              x: torch.Tensor,
              bias: Optional[torch.Tensor] = None) -> torch.Tensor:
-        # MLU F.linear requires matching dtypes
-        if x.dtype != layer.weight.dtype:
-            x = x.to(layer.weight.dtype)
        return F.linear(x, layer.weight, bias)

    def embedding(self, layer: torch.nn.Module,
--- a/vllm-v0.6.2/vllm/model_executor/model_loader/utils.py
+++ b/vllm-v0.6.2/vllm/model_executor/model_loader/utils.py
@@ -20,7 +20,7 @@ def set_default_torch_dtype(dtype: torch.dtype):

 def get_model_architecture(
        model_config: ModelConfig) -> Tuple[Type[nn.Module], str]:
-    architectures = getattr(model_config.hf_config, "architectures", None) or []
+    architectures = getattr(model_config.hf_config, "architectures", [])
    # Special handling for quantized Mixtral.
    # FIXME(woosuk): This is a temporary hack.
    mixtral_supported = [
--- a/vllm-v0.6.2/vllm/model_executor/models/deepseek_mtp.py
+++ b/vllm-v0.6.2/vllm/model_executor/models/deepseek_mtp.py
@@ -1,291 +0,0 @@
-"""Inference-only DeepSeek V3 Multi-Token Prediction (MTP) model."""
-from typing import Iterable, List, Optional, Tuple
-
-import torch
-import torch.nn as nn
-
-from vllm.attention.backends.abstract import AttentionMetadata
-from vllm.config import CacheConfig, VllmConfig
-from vllm.model_executor.layers.layernorm import RMSNorm
-from vllm.model_executor.layers.logits_processor import LogitsProcessor
-from vllm.model_executor.layers.quantization import QuantizationConfig
-from vllm.model_executor.layers.sampler import SamplerOutput, get_sampler
-from vllm.model_executor.layers.vocab_parallel_embedding import (
-    ParallelLMHead, VocabParallelEmbedding)
-from vllm.model_executor.model_loader.weight_utils import default_weight_loader
-from vllm.model_executor.sampling_metadata import SamplingMetadata
-from vllm.sequence import IntermediateTensors
-
-from .deepseek_v2 import DeepseekV2DecoderLayer
-
-
-class SharedHead(nn.Module):
-    """Shared head for MTP: norm + lm_head."""
-
-    def __init__(self, config, prefix: str = ""):
-        super().__init__()
-        self.norm = RMSNorm(config.hidden_size,
-                            eps=getattr(config, "rms_norm_eps", 1e-6))
-        self.head = ParallelLMHead(
-            config.vocab_size,
-            config.hidden_size,
-        )
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        return self.norm(hidden_states)
-
-
-class DeepSeekMultiTokenPredictorLayer(nn.Module):
-    """Single MTP layer: enorm + hnorm + eh_proj + shared_head + mtp_block."""
-
-    def __init__(
-        self,
-        config,
-        layer_idx: int,
-        cache_config: Optional[CacheConfig] = None,
-        quant_config: Optional[QuantizationConfig] = None,
-        prefix: str = "",
-    ):
-        super().__init__()
-        self.enorm = RMSNorm(config.hidden_size,
-                             eps=getattr(config, "rms_norm_eps", 1e-6))
-        self.hnorm = RMSNorm(config.hidden_size,
-                             eps=getattr(config, "rms_norm_eps", 1e-6))
-        self.eh_proj = nn.Linear(config.hidden_size * 2,
-                                 config.hidden_size,
-                                 bias=False)
-        self.shared_head = SharedHead(config,
-                                      prefix=f"{prefix}.shared_head")
-        # Reuse DeepseekV2DecoderLayer (MLU hijack auto-applies)
-        self.mtp_block = DeepseekV2DecoderLayer(
-            config,
-            prefix=f"model.layers.{layer_idx}",
-            cache_config=cache_config,
-            quant_config=quant_config,
-        )
-
-    def forward(
-        self,
-        input_ids: torch.Tensor,
-        positions: torch.Tensor,
-        previous_hidden_states: torch.Tensor,
-        kv_caches: List[torch.Tensor],
-        attn_metadata: AttentionMetadata,
-        inputs_embeds: Optional[torch.Tensor] = None,
-    ) -> torch.Tensor:
-        assert inputs_embeds is not None
-        # Mask inputs at position 0
-        inputs_embeds = torch.where(
-            positions.unsqueeze(-1) == 0, 0, inputs_embeds)
-        inputs_embeds = self.enorm(inputs_embeds)
-        previous_hidden_states = self.hnorm(previous_hidden_states)
-
-        hidden_states = self.eh_proj(
-            torch.cat([inputs_embeds, previous_hidden_states], dim=-1))
-
-        hidden_states, residual = self.mtp_block(
-            positions, hidden_states, kv_caches[0], attn_metadata,
-            residual=None)
-        hidden_states = residual + hidden_states
-        return hidden_states
-
-
-def get_spec_layer_idx_from_weight_name(config, weight_name: str):
-    """Check if weight belongs to a speculative (MTP) layer.
-    Returns the layer index if so, None otherwise."""
-    num_nextn = getattr(config, "num_nextn_predict_layers", 0)
-    if num_nextn and num_nextn > 0:
-        layer_idx = config.num_hidden_layers
-        for i in range(num_nextn):
-            if weight_name.startswith(f"model.layers.{layer_idx + i}."):
-                return layer_idx + i
-    return None
-
-
-def _rewrite_spec_layer_name(config, spec_layer: int, name: str) -> str:
-    """Rewrite weight name for MTP layer.
-    Add .mtp_block for transformer block weights,
-    rename shared weights to top level."""
-    spec_layer_weight_names = [
-        "embed_tokens", "enorm", "hnorm", "eh_proj", "shared_head",
-    ]
-    shared_weight_names = ["embed_tokens"]
-    spec_layer_weight = False
-    shared_weight = False
-    for weight_name in spec_layer_weight_names:
-        if weight_name in name:
-            spec_layer_weight = True
-            if weight_name in shared_weight_names:
-                shared_weight = True
-            break
-    if not spec_layer_weight:
-        # Transformer block weights -> add .mtp_block prefix
-        name = name.replace(
-            f"model.layers.{spec_layer}.",
-            f"model.layers.{spec_layer}.mtp_block.")
-    elif shared_weight:
-        # Shared weights -> top level
-        name = name.replace(f"model.layers.{spec_layer}.", "model.")
-    return name
-
-
-class DeepSeekMTP(nn.Module):
-    """DeepSeek V3 Multi-Token Prediction draft model.
-    Uses hidden states from the target model to predict the next token
-    via a single additional decoder layer."""
-
-    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
-        super().__init__()
-        config = vllm_config.model_config.hf_config
-        cache_config = vllm_config.cache_config
-        quant_config = vllm_config.quant_config
-        self.config = config
-
-        self.embed_tokens = VocabParallelEmbedding(
-            config.vocab_size,
-            config.hidden_size,
-        )
-
-        self.mtp_start_layer_idx = config.num_hidden_layers
-        num_mtp = getattr(config, "num_nextn_predict_layers", 1)
-
-        self.layers = nn.ModuleDict()
-        for i in range(num_mtp):
-            layer_idx = self.mtp_start_layer_idx + i
-            self.layers[str(layer_idx)] = DeepSeekMultiTokenPredictorLayer(
-                config=config,
-                layer_idx=layer_idx,
-                cache_config=cache_config,
-                quant_config=quant_config,
-                prefix=f"model.layers.{layer_idx}",
-            )
-
-        self.logits_processor = LogitsProcessor(config.vocab_size)
-        self.sampler = get_sampler()
-
-    def forward(
-        self,
-        input_ids: torch.Tensor,
-        positions: torch.Tensor,
-        kv_caches: List[torch.Tensor],
-        attn_metadata: AttentionMetadata,
-        previous_hidden_states: torch.Tensor,
-        intermediate_tensors: Optional[IntermediateTensors] = None,
-    ) -> torch.Tensor:
-        inputs_embeds = self.embed_tokens(input_ids)
-        # Use the first MTP layer (DeepSeek V3 only has 1)
-        layer = self.layers[str(self.mtp_start_layer_idx)]
-        hidden_states = layer(
-            input_ids, positions, previous_hidden_states,
-            kv_caches, attn_metadata, inputs_embeds)
-        return hidden_states
-
-    def compute_logits(
-        self,
-        hidden_states: torch.Tensor,
-        sampling_metadata: SamplingMetadata,
-    ) -> Optional[torch.Tensor]:
-        layer = self.layers[str(self.mtp_start_layer_idx)]
-        normed = layer.shared_head(hidden_states)
-        logits = self.logits_processor(
-            layer.shared_head.head, normed, sampling_metadata)
-        return logits
-
-    def sample(
-        self,
-        logits: Optional[torch.Tensor],
-        sampling_metadata: SamplingMetadata,
-    ) -> Optional[SamplerOutput]:
-        next_tokens = self.sampler(logits, sampling_metadata)
-        return next_tokens
-
-    def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
-        # MLU SparseMoeMlp needs pack_params() before loading
-        try:
-            from vllm_mlu.model_executor.layers.sparse_moe_mlp import (
-                SparseMoeMlp)
-            for name, m in self.named_modules():
-                if isinstance(m, SparseMoeMlp):
-                    m.pack_params()
-        except ImportError:
-            pass
-
-        stacked_params_mapping = [
-            # (param_name, shard_name, shard_id)
-            ("gate_up_proj", "gate_proj", 0),
-            ("gate_up_proj", "up_proj", 1),
-        ]
-
-        params_dict = dict(self.named_parameters())
-
-        for name, loaded_weight in weights:
-            if "rotary_emb.inv_freq" in name:
-                continue
-
-            # Only load MTP layer weights
-            spec_layer = get_spec_layer_idx_from_weight_name(
-                self.config, name)
-            if spec_layer is None:
-                continue
-
-            # Rewrite weight name for MTP structure
-            name = _rewrite_spec_layer_name(
-                self.config, spec_layer, name)
-
-            # Only load shared weights (embed_tokens) from first
-            # MTP layer, per DeepSeek V3 Technical Report
-            if (spec_layer != self.mtp_start_layer_idx
-                    and ".layers" not in name):
-                continue
-
-            # Strip "model." prefix since DeepSeekMTP holds
-            # embed_tokens and layers directly (no .model wrapper)
-            if name.startswith("model."):
-                name = name[len("model."):]
-
-            self._load_single_weight(
-                name, loaded_weight, stacked_params_mapping,
-                params_dict)
-
-    def _load_single_weight(
-        self,
-        name: str,
-        loaded_weight: torch.Tensor,
-        stacked_params_mapping: List[Tuple[str, str, int]],
-        params_dict: dict,
-    ):
-        for (param_name, weight_name, shard_id) in stacked_params_mapping:
-            if weight_name not in name:
-                continue
-            name = name.replace(weight_name, param_name)
-            # Skip expert weights not in params_dict
-            if (("mlp.experts." in name
-                    or "mlp.shared_experts." in name
-                    or "mlp.shared_expert_gate." in name
-                    or "e_score_correction_bias" in name)
-                    and name not in params_dict):
-                return
-            if name.endswith(".bias") and name not in params_dict:
-                return
-            if name not in params_dict:
-                return
-            param = params_dict[name]
-            weight_loader = param.weight_loader
-            weight_loader(param, loaded_weight, shard_id)
-            return
-
-        # Non-stacked weights
-        if name.endswith(".bias") and name not in params_dict:
-            return
-        if (("mlp.experts." in name
-                or "mlp.shared_experts." in name
-                or "mlp.shared_expert_gate." in name
-                or "e_score_correction_bias" in name)
-                and name not in params_dict):
-            return
-        if name not in params_dict:
-            return
-        param = params_dict[name]
-        weight_loader = getattr(param, "weight_loader",
-                                default_weight_loader)
-        weight_loader(param, loaded_weight)
--- a/vllm-v0.6.2/vllm/model_executor/models/deepseek_v2.py
+++ b/vllm-v0.6.2/vllm/model_executor/models/deepseek_v2.py
@@ -611,7 +611,3 @@ class DeepseekV2ForCausalLM(nn.Module, SupportsPP):
                    weight_loader = getattr(param, "weight_loader",
                                            default_weight_loader)
                    weight_loader(param, loaded_weight)
-
-
-class DeepseekV3ForCausalLM(DeepseekV2ForCausalLM):
-    pass
--- a/vllm-v0.6.2/vllm/model_executor/models/llama4.py
+++ b/vllm-v0.6.2/vllm/model_executor/models/llama4.py
@@ -1,580 +0,0 @@
-# Copyright 2025 the LLAMA4, Meta Inc., vLLM, and 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.
-"""Inference-only Llama4 model compatible with HuggingFace weights."""
-import re
-from typing import Any, Dict, Iterable, List, Optional, Tuple, Union
-
-import torch
-from torch import nn
-
-from vllm.attention import Attention, AttentionMetadata
-from vllm.config import CacheConfig, VllmConfig
-from vllm.distributed import (get_pp_group,
-                              get_tensor_model_parallel_world_size,
-                              tensor_model_parallel_all_reduce)
-from vllm.logger import init_logger
-from vllm.model_executor.layers.fused_moe import FusedMoE
-from vllm.model_executor.layers.layernorm import RMSNorm
-from vllm.model_executor.layers.linear import (QKVParallelLinear,
-                                               ReplicatedLinear,
-                                               RowParallelLinear)
-from vllm.model_executor.layers.logits_processor import LogitsProcessor
-from vllm.model_executor.layers.quantization import QuantizationConfig
-from vllm.model_executor.layers.rotary_embedding import get_rope
-from vllm.model_executor.layers.sampler import SamplerOutput, get_sampler
-from vllm.model_executor.layers.vocab_parallel_embedding import (
-    DEFAULT_VOCAB_PADDING_SIZE, ParallelLMHead, VocabParallelEmbedding)
-from vllm.model_executor.model_loader.weight_utils import (
-    default_weight_loader, maybe_remap_kv_scale_name)
-from vllm.model_executor.sampling_metadata import SamplingMetadata
-from vllm.sequence import IntermediateTensors
-
-from .interfaces import SupportsPP
-from .llama import LlamaMLP
-from .utils import (AutoWeightsLoader, PPMissingLayer, is_pp_missing_parameter,
-                    make_empty_intermediate_tensors_factory, make_layers,
-                    maybe_prefix)
-
-logger = init_logger(__name__)
-
-
-def _extract_layer_index(prefix: str) -> int:
-    """Extract layer index from prefix string like 'model.layers.0.self_attn'."""
-    match = re.search(r'layers\.(\d+)', prefix)
-    if match is None:
-        raise ValueError(f"Cannot extract layer index from prefix: {prefix}")
-    return int(match.group(1))
-
-
-class Llama4MoE(nn.Module):
-    """Llama4 Mixture of Experts with shared expert."""
-
-    @staticmethod
-    def custom_routing_function(
-        hidden_states: torch.Tensor,
-        gating_output: torch.Tensor,
-        topk: int,
-        renormalize: bool,
-    ) -> Tuple[torch.Tensor, torch.Tensor]:
-        router_scores, router_indices = torch.topk(
-            gating_output, topk, dim=-1)
-        router_scores = torch.sigmoid(router_scores.float())
-        return (router_scores, router_indices.to(torch.int32))
-
-    def __init__(
-        self,
-        config,
-        quant_config: Optional[QuantizationConfig] = None,
-        prefix: str = "",
-    ) -> None:
-        super().__init__()
-        self.tp_size = get_tensor_model_parallel_world_size()
-        self.top_k = getattr(config, "num_experts_per_tok", 1)
-        self.num_local_experts = getattr(config, "num_local_experts", 8)
-        self.hidden_size = getattr(config, "hidden_size", 4096)
-        intermediate_size_moe = getattr(config, "intermediate_size", 8192)
-
-        self.router = ReplicatedLinear(
-            self.hidden_size,
-            self.num_local_experts,
-            bias=False,
-            quant_config=None,
-            prefix=f"{prefix}.router",
-        )
-
-        self.experts = FusedMoE(
-            num_experts=self.num_local_experts,
-            top_k=self.top_k,
-            hidden_size=self.hidden_size,
-            intermediate_size=intermediate_size_moe,
-            reduce_results=False,
-            renormalize=False,
-            quant_config=quant_config,
-            custom_routing_function=Llama4MoE.custom_routing_function,
-            prefix=f"{prefix}.experts",
-        )
-
-        self.shared_expert = LlamaMLP(
-            hidden_size=self.hidden_size,
-            intermediate_size=intermediate_size_moe,
-            hidden_act="silu",
-            quant_config=quant_config,
-            bias=False,
-            prefix=f"{prefix}.shared_expert",
-        )
-
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
-        orig_shape = hidden_states.shape
-        hidden_states = hidden_states.view(-1, self.hidden_size)
-        # router_logits: (num_tokens, n_experts)
-        router_logits, _ = self.router(hidden_states)
-        # routed experts
-        routed_out = self.experts(hidden_states, router_logits)
-        # shared expert
-        shared_out = self.shared_expert(hidden_states)
-        # combine and all-reduce
-        experts_out = routed_out + shared_out
-        if self.tp_size > 1:
-            experts_out = tensor_model_parallel_all_reduce(experts_out)
-        return experts_out.view(orig_shape)
-
-
-class Llama4Attention(nn.Module):
-
-    def __init__(
-        self,
-        config,
-        hidden_size: int,
-        num_heads: int,
-        num_kv_heads: int,
-        max_position_embeddings: int = 8192,
-        quant_config: Optional[QuantizationConfig] = None,
-        bias: bool = False,
-        cache_config: Optional[CacheConfig] = None,
-        prefix: str = "",
-    ) -> None:
-        super().__init__()
-        self.layer_idx = _extract_layer_index(prefix)
-        self.hidden_size = hidden_size
-        self.no_rope_layers = getattr(config, "no_rope_layers", None)
-        self.nope = (self.no_rope_layers is not None
-                     and self.no_rope_layers[self.layer_idx] == 0)
-        self.use_qk_norm = getattr(config, "use_qk_norm", False) and not self.nope
-        tp_size = get_tensor_model_parallel_world_size()
-        self.total_num_heads = num_heads
-        assert self.total_num_heads % tp_size == 0
-        self.num_heads = self.total_num_heads // tp_size
-        self.total_num_kv_heads = num_kv_heads
-        if self.total_num_kv_heads >= tp_size:
-            assert self.total_num_kv_heads % tp_size == 0
-        else:
-            assert tp_size % self.total_num_kv_heads == 0
-        self.num_kv_heads = max(1, self.total_num_kv_heads // tp_size)
-        self.head_dim = getattr(config, "head_dim",
-                                self.hidden_size // self.total_num_heads)
-        self.q_size = self.num_heads * self.head_dim
-        self.kv_size = self.num_kv_heads * self.head_dim
-        self.scaling = self.head_dim**-0.5
-        self.max_position_embeddings = max_position_embeddings
-
-        # Temperature tuning for NoPE layers
-        self.attn_temperature_tuning = (
-            self.nope and getattr(config, "attn_temperature_tuning", False))
-        self.floor_scale = getattr(config, "floor_scale", 8192.0)
-        self.attn_scale = getattr(config, "attn_scale", 0.1)
-
-        # QK norm
-        rms_norm_eps = getattr(config, "rms_norm_eps", 1e-5)
-        if self.use_qk_norm:
-            self.qk_norm = RMSNorm(self.head_dim, eps=rms_norm_eps)
-            # v0.6.2 RMSNorm doesn't support has_weight=False,
-            # so we set weight to ones and make it non-trainable
-            self.qk_norm.weight.data.fill_(1.0)
-            self.qk_norm.weight.requires_grad = False
-        else:
-            self.qk_norm = None
-
-        self.qkv_proj = QKVParallelLinear(
-            hidden_size=hidden_size,
-            head_size=self.head_dim,
-            total_num_heads=self.total_num_heads,
-            total_num_kv_heads=self.total_num_kv_heads,
-            bias=bias,
-            quant_config=quant_config,
-            prefix=f"{prefix}.qkv_proj",
-        )
-        self.o_proj = RowParallelLinear(
-            input_size=self.total_num_heads * self.head_dim,
-            output_size=hidden_size,
-            bias=bias,
-            quant_config=quant_config,
-            prefix=f"{prefix}.o_proj",
-        )
-
-        # RoPE (None for NoPE layers)
-        rope_theta = getattr(config, "rope_theta", 10000)
-        rope_scaling = getattr(config, "rope_scaling", None)
-        if not self.nope:
-            self.rotary_emb = get_rope(
-                self.head_dim,
-                rotary_dim=self.head_dim,
-                max_position=max_position_embeddings,
-                base=rope_theta,
-                rope_scaling=rope_scaling,
-                is_neox_style=True,
-            )
-        else:
-            self.rotary_emb = None
-
-        self.attn = Attention(
-            self.num_heads,
-            self.head_dim,
-            self.scaling,
-            num_kv_heads=self.num_kv_heads,
-            cache_config=cache_config,
-            quant_config=quant_config,
-        )
-
-    def _get_attn_scale(self, positions: torch.Tensor) -> torch.Tensor:
-        floor = torch.floor((positions + 1.0) / self.floor_scale)
-        attn_scale = torch.log(floor + 1.0) * self.attn_scale + 1.0
-        return attn_scale.unsqueeze(-1)
-
-    def forward(
-        self,
-        positions: torch.Tensor,
-        hidden_states: torch.Tensor,
-        kv_cache: torch.Tensor,
-        attn_metadata: AttentionMetadata,
-    ) -> torch.Tensor:
-        qkv, _ = self.qkv_proj(hidden_states)
-        q, k, v = qkv.split([self.q_size, self.kv_size, self.kv_size],
-                             dim=-1)
-
-        if self.rotary_emb is not None:
-            q, k = self.rotary_emb(positions, q, k)
-
-        if self.qk_norm is not None:
-            q = q.reshape(-1, self.head_dim)
-            q = self.qk_norm(q.float()).reshape(-1, self.q_size).to(q.dtype)
-            k = k.reshape(-1, self.head_dim)
-            k = self.qk_norm(k.float()).reshape(-1, self.kv_size).to(k.dtype)
-
-        if self.attn_temperature_tuning and self.nope:
-            attn_scale = self._get_attn_scale(positions)
-            q = (q * attn_scale).to(q.dtype)
-
-        attn_output = self.attn(q, k, v, kv_cache, attn_metadata)
-        output, _ = self.o_proj(attn_output)
-        return output
-
-
-class Llama4DecoderLayer(nn.Module):
-
-    def __init__(
-        self,
-        config,
-        cache_config: Optional[CacheConfig] = None,
-        quant_config: Optional[QuantizationConfig] = None,
-        prefix: str = "",
-    ) -> None:
-        super().__init__()
-        self.layer_idx = _extract_layer_index(prefix)
-        self.hidden_size = getattr(config, "hidden_size", 4096)
-        max_position_embeddings = getattr(config, "max_position_embeddings",
-                                          8192)
-
-        self.self_attn = Llama4Attention(
-            config=config,
-            hidden_size=self.hidden_size,
-            num_heads=getattr(config, "num_attention_heads", 32),
-            num_kv_heads=getattr(config, "num_key_value_heads",
-                                 getattr(config, "num_attention_heads", 32)),
-            max_position_embeddings=max_position_embeddings,
-            quant_config=quant_config,
-            bias=False,
-            cache_config=cache_config,
-            prefix=f"{prefix}.self_attn",
-        )
-
-        # Interleaved MoE/dense layers
-        interleave_moe_layer_step = getattr(config,
-                                            "interleave_moe_layer_step", 0)
-        is_moe_layer = (interleave_moe_layer_step > 0
-                        and (self.layer_idx + 1)
-                        % interleave_moe_layer_step == 0)
-
-        if is_moe_layer:
-            self.feed_forward = Llama4MoE(
-                config=config,
-                quant_config=quant_config,
-                prefix=f"{prefix}.feed_forward",
-            )
-        else:
-            intermediate_size_mlp = getattr(config, "intermediate_size_mlp",
-                                            getattr(config,
-                                                    "intermediate_size", 8192))
-            self.feed_forward = LlamaMLP(
-                hidden_size=self.hidden_size,
-                intermediate_size=intermediate_size_mlp,
-                hidden_act="silu",
-                quant_config=quant_config,
-                bias=False,
-                prefix=f"{prefix}.feed_forward",
-            )
-
-        rms_norm_eps = getattr(config, "rms_norm_eps", 1e-5)
-        self.input_layernorm = RMSNorm(self.hidden_size, eps=rms_norm_eps)
-        self.post_attention_layernorm = RMSNorm(self.hidden_size,
-                                                eps=rms_norm_eps)
-
-    def forward(
-        self,
-        positions: torch.Tensor,
-        hidden_states: torch.Tensor,
-        kv_cache: torch.Tensor,
-        attn_metadata: AttentionMetadata,
-        residual: Optional[torch.Tensor],
-    ) -> Tuple[torch.Tensor, torch.Tensor]:
-        # Self Attention
-        if residual is None:
-            residual = hidden_states
-            hidden_states = self.input_layernorm(hidden_states)
-        else:
-            hidden_states, residual = self.input_layernorm(
-                hidden_states, residual)
-        hidden_states = self.self_attn(positions=positions,
-                                       hidden_states=hidden_states,
-                                       kv_cache=kv_cache,
-                                       attn_metadata=attn_metadata)
-
-        # Fully Connected
-        hidden_states, residual = self.post_attention_layernorm(
-            hidden_states, residual)
-        hidden_states = self.feed_forward(hidden_states)
-        return hidden_states, residual
-
-
-class Llama4Model(nn.Module):
-    """Llama4 model - independent implementation to avoid pad_token_id issue."""
-
-    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
-        super().__init__()
-
-        config = vllm_config.model_config.hf_config
-        cache_config = vllm_config.cache_config
-        quant_config = vllm_config.quant_config
-        lora_config = vllm_config.lora_config
-
-        self.config = config
-        # Defensive access - Llama4Config may not have pad_token_id
-        self.padding_idx = getattr(config, "pad_token_id", None)
-        lora_vocab = (lora_config.lora_extra_vocab_size *
-                      (lora_config.max_loras or 1)) if lora_config else 0
-        self.vocab_size = config.vocab_size + lora_vocab
-        self.org_vocab_size = config.vocab_size
-
-        if get_pp_group().is_first_rank or (
-                getattr(config, "tie_word_embeddings", False)
-                and get_pp_group().is_last_rank):
-            self.embed_tokens = VocabParallelEmbedding(
-                self.vocab_size,
-                config.hidden_size,
-                org_num_embeddings=config.vocab_size,
-                quant_config=quant_config,
-            )
-        else:
-            self.embed_tokens = PPMissingLayer()
-
-        self.start_layer, self.end_layer, self.layers = make_layers(
-            config.num_hidden_layers,
-            lambda prefix: Llama4DecoderLayer(
-                config=config,
-                cache_config=cache_config,
-                quant_config=quant_config,
-                prefix=prefix),
-            prefix=f"{prefix}.layers",
-        )
-
-        rms_norm_eps = getattr(config, "rms_norm_eps", 1e-5)
-        if get_pp_group().is_last_rank:
-            self.norm = RMSNorm(config.hidden_size, eps=rms_norm_eps)
-        else:
-            self.norm = PPMissingLayer()
-
-        self.make_empty_intermediate_tensors = (
-            make_empty_intermediate_tensors_factory(
-                ["hidden_states", "residual"], config.hidden_size))
-
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
-        return self.embed_tokens(input_ids)
-
-    def forward(
-        self,
-        input_ids: Optional[torch.Tensor],
-        positions: torch.Tensor,
-        kv_caches: List[torch.Tensor],
-        attn_metadata: AttentionMetadata,
-        intermediate_tensors: Optional[IntermediateTensors],
-        inputs_embeds: Optional[torch.Tensor] = None,
-    ) -> Union[torch.Tensor, IntermediateTensors]:
-        if get_pp_group().is_first_rank:
-            if inputs_embeds is not None:
-                hidden_states = inputs_embeds
-            else:
-                hidden_states = self.get_input_embeddings(input_ids)
-            residual = None
-        else:
-            assert intermediate_tensors is not None
-            hidden_states = intermediate_tensors["hidden_states"]
-            residual = intermediate_tensors["residual"]
-
-        for i in range(self.start_layer, self.end_layer):
-            layer = self.layers[i]
-            hidden_states, residual = layer(positions, hidden_states,
-                                            kv_caches[i - self.start_layer],
-                                            attn_metadata, residual)
-
-        if not get_pp_group().is_last_rank:
-            return IntermediateTensors({
-                "hidden_states": hidden_states,
-                "residual": residual
-            })
-
-        hidden_states, _ = self.norm(hidden_states, residual)
-        return hidden_states
-
-
-class Llama4ForCausalLM(nn.Module, SupportsPP):
-    packed_modules_mapping = {
-        "qkv_proj": ["q_proj", "k_proj", "v_proj"],
-        "gate_up_proj": ["gate_proj", "up_proj"],
-    }
-
-    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
-        super().__init__()
-        config = vllm_config.model_config.hf_config
-        # Llama4ForConditionalGeneration uses top-level Llama4Config
-        # which has text_config sub-config. Extract it for text model.
-        text_config = getattr(config, "text_config", None)
-        if text_config is not None:
-            orig_archs = getattr(config, "architectures", None)
-            vllm_config.model_config.hf_config = text_config
-            if orig_archs and not getattr(text_config, "architectures", None):
-                text_config.architectures = orig_archs
-            config = text_config
-        quant_config = vllm_config.quant_config
-        lora_config = vllm_config.lora_config
-        self.config = config
-        self.lora_config = lora_config
-
-        self.model = Llama4Model(vllm_config=vllm_config,
-                                 prefix=maybe_prefix(prefix, "model"))
-
-        if get_pp_group().is_last_rank:
-            self.unpadded_vocab_size = config.vocab_size
-            if lora_config:
-                self.unpadded_vocab_size += lora_config.lora_extra_vocab_size
-            self.lm_head = ParallelLMHead(
-                self.unpadded_vocab_size,
-                config.hidden_size,
-                org_num_embeddings=config.vocab_size,
-                padding_size=(
-                    DEFAULT_VOCAB_PADDING_SIZE if not lora_config
-                    else lora_config.lora_vocab_padding_size),
-                quant_config=quant_config,
-                prefix=maybe_prefix(prefix, "lm_head"),
-            )
-            if getattr(config, "tie_word_embeddings", False):
-                self.lm_head = self.lm_head.tie_weights(
-                    self.model.embed_tokens)
-
-            logit_scale = getattr(config, "logit_scale", 1.0)
-            self.logits_processor = LogitsProcessor(
-                self.unpadded_vocab_size,
-                config.vocab_size,
-                logit_scale)
-            self.sampler = get_sampler()
-        else:
-            self.lm_head = PPMissingLayer()
-
-        self.make_empty_intermediate_tensors = (
-            self.model.make_empty_intermediate_tensors)
-
-    def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
-        return self.model.get_input_embeddings(input_ids)
-
-    def forward(
-        self,
-        input_ids: torch.Tensor,
-        positions: torch.Tensor,
-        kv_caches: List[torch.Tensor],
-        attn_metadata: AttentionMetadata,
-        intermediate_tensors: Optional[IntermediateTensors] = None,
-        inputs_embeds: Optional[torch.Tensor] = None,
-    ) -> Union[torch.Tensor, IntermediateTensors]:
-        model_output = self.model(input_ids, positions, kv_caches,
-                                  attn_metadata, intermediate_tensors,
-                                  inputs_embeds)
-        return model_output
-
-    def compute_logits(
-        self,
-        hidden_states: torch.Tensor,
-        sampling_metadata: SamplingMetadata,
-    ) -> Optional[torch.Tensor]:
-        logits = self.logits_processor(self.lm_head, hidden_states,
-                                       sampling_metadata)
-        return logits
-
-    def sample(self, logits: torch.Tensor,
-               sampling_metadata: SamplingMetadata) -> Optional[SamplerOutput]:
-        next_tokens = self.sampler(logits, sampling_metadata)
-        return next_tokens
-
-    def permute_qk_weight_for_rotary(
-        self,
-        name: str,
-        loaded_weight: torch.Tensor,
-    ) -> Tuple[str, torch.Tensor]:
-        """Permute Q/K weights for rotary embedding compatibility."""
-        def permute(w: torch.Tensor, n_heads: int):
-            attn_in = getattr(self.config, "head_dim", 128) * n_heads
-            attn_out = getattr(self.config, "hidden_size", 4096)
-            return (w.contiguous()
-                    .view(n_heads, attn_in // n_heads // 2, 2, attn_out)
-                    .transpose(1, 2).reshape(attn_in, attn_out))
-
-        modules = name.split(".")
-        is_weight = modules[-1] == "weight"
-
-        if is_weight:
-            if "k_proj" in modules:
-                loaded_weight = permute(
-                    loaded_weight,
-                    getattr(self.config, "num_key_value_heads", 8))
-            elif "q_proj" in modules:
-                loaded_weight = permute(
-                    loaded_weight,
-                    getattr(self.config, "num_attention_heads", 32))
-
-        return name, loaded_weight
-
-    def load_weights(
-        self, weights: Iterable[Tuple[str, torch.Tensor]],
-    ):
-        loader = AutoWeightsLoader(
-            self,
-            skip_prefixes=(
-                ["lm_head."]
-                if getattr(self.config, "tie_word_embeddings", False)
-                else None),
-        )
-
-        def _process_weights(weights):
-            for name, loaded_weight in weights:
-                # Strip language_model. prefix for Llama4ForConditionalGeneration
-                if name.startswith("language_model."):
-                    name = name[len("language_model."):]
-                # Skip vision encoder weights
-                elif name.startswith("multi_modal_projector.") or \
-                     name.startswith("vision_encoder.") or \
-                     name.startswith("vision_model."):
-                    continue
-                name, loaded_weight = self.permute_qk_weight_for_rotary(
-                    name, loaded_weight)
-                yield name, loaded_weight
-
-        loader.load_weights(_process_weights(weights))
--- a/vllm-v0.6.2/vllm/model_executor/models/registry.py
+++ b/vllm-v0.6.2/vllm/model_executor/models/registry.py
@@ -48,7 +48,6 @@ _TEXT_GENERATION_MODELS = {
    "DeciLMForCausalLM": ("decilm", "DeciLMForCausalLM"),
    "DeepseekForCausalLM": ("deepseek", "DeepseekForCausalLM"),
    "DeepseekV2ForCausalLM": ("deepseek_v2", "DeepseekV2ForCausalLM"),
-    "DeepseekV3ForCausalLM": ("deepseek_v2", "DeepseekV3ForCausalLM"),
    "ExaoneForCausalLM": ("exaone", "ExaoneForCausalLM"),
    "FalconForCausalLM": ("falcon", "FalconForCausalLM"),
    "GemmaForCausalLM": ("gemma", "GemmaForCausalLM"),
@@ -65,8 +64,6 @@ _TEXT_GENERATION_MODELS = {
    "InternLM2VEForCausalLM": ("internlm2_ve", "InternLM2VEForCausalLM"),
    "JAISLMHeadModel": ("jais", "JAISLMHeadModel"),
    "JambaForCausalLM": ("jamba", "JambaForCausalLM"),
-    "Llama4ForCausalLM": ("llama4", "Llama4ForCausalLM"),
-    "Llama4ForConditionalGeneration": ("llama4", "Llama4ForCausalLM"),
    "LlamaForCausalLM": ("llama", "LlamaForCausalLM"),
    # For decapoda-research/llama-*
    "LLaMAForCausalLM": ("llama", "LlamaForCausalLM"),
@@ -166,7 +163,6 @@ _SPECULATIVE_DECODING_MODELS = {
    "EAGLEModel": ("eagle", "EAGLE"),
    "MedusaModel": ("medusa", "Medusa"),
    "MLPSpeculatorPreTrainedModel": ("mlp_speculator", "MLPSpeculator"),
-    "DeepSeekMTPModel": ("deepseek_mtp", "DeepSeekMTP"),
 }

 # Transformers backend models - wrapper classes for custom HuggingFace models
@@ -484,7 +480,6 @@ class _ModelRegistry:
            architectures = [architectures]
        if not architectures:
            logger.warning("No model architectures are specified")
-            return []

        return architectures

--- a/vllm-v0.6.2/vllm/spec_decode/mlu_spec_decode_worker.py
+++ b/vllm-v0.6.2/vllm/spec_decode/mlu_spec_decode_worker.py
@@ -159,11 +159,9 @@ class MLUSpecDecodeWorker(LoraNotSupportedWorkerBase):
                    draft_worker_kwargs[
                        "model_runner_cls"] = MLUTP1DraftModelRunner
                else:
-                    if draft_model_config.hf_config.model_type in (
-                            "eagle", "deepseek_mtp"):
+                    if draft_model_config.hf_config.model_type == "eagle":
                        raise NotImplementedError(
-                            f"{draft_model_config.hf_config.model_type} "
-                            "does not support TP > 1 yet")
+                            "EAGLE does not support TP > 1 yet")

                    allow_zero_draft_token_step = False
                proposer_worker = MLUMultiStepWorker(**draft_worker_kwargs)
--- a/vllm-v0.6.2/vllm/transformers_utils/config.py
+++ b/vllm-v0.6.2/vllm/transformers_utils/config.py
@@ -13,7 +13,7 @@ from transformers import GenerationConfig, PretrainedConfig
 from transformers.models.auto.image_processing_auto import (
    get_image_processor_config)
 from transformers.models.auto.modeling_auto import (
-    MODEL_FOR_CAUSAL_LM_MAPPING_NAMES, MODEL_MAPPING_NAMES)
+    MODEL_FOR_CAUSAL_LM_MAPPING_NAMES)
 from transformers.utils import CONFIG_NAME as HF_CONFIG_NAME

 from vllm.envs import VLLM_USE_MODELSCOPE
@@ -229,17 +229,6 @@ def get_config(
        model_type = MODEL_FOR_CAUSAL_LM_MAPPING_NAMES[config.model_type]
        config.update({"architectures": [model_type]})

-    # Architecture mapping for models without explicit architectures field
-    if not getattr(config, "architectures", None):
-        if config.model_type not in MODEL_MAPPING_NAMES:
-            logger.warning(
-                "Model config does not have a top-level 'architectures' "
-                "field: expecting `hf_overrides={'architectures': "
-                "['...']}` to be passed in engine args.")
-        else:
-            model_type = MODEL_MAPPING_NAMES[config.model_type]
-            config.update({"architectures": [model_type]})
-
    patch_rope_scaling(config)

    if trust_remote_code:
--- a/vllm-v0.6.2/vllm/worker/mlu_worker.py
+++ b/vllm-v0.6.2/vllm/worker/mlu_worker.py
@@ -59,20 +59,12 @@ class MLUWorker(Worker):
        # mlp_speculator
        speculative_config = self.speculative_config
        model_config = self.model_config
-        is_mtp = (speculative_config is not None
-                  and model_config.task != "draft"
-                  and getattr(
-                      speculative_config.draft_model_config.hf_config,
-                      "model_type", None) == "deepseek_mtp")
-        speculative_args = (
-            {"return_hidden_states": True} if is_mtp else
-            ({} if speculative_config is None
-             or (speculative_config.draft_model_config.model ==
-                 model_config.model)
-             or (speculative_config.draft_model_config.hf_config.model_type
-                 not in ["medusa", "mlp_speculator", "eagle"])
-             else {"return_hidden_states": True})
-        )
+        speculative_args = {} if speculative_config is None \
+            or (speculative_config.draft_model_config.model ==
+                model_config.model) \
+            or (speculative_config.draft_model_config.hf_config.model_type
+                not in ["medusa", "mlp_speculator", "eagle"]) \
+                    else {"return_hidden_states": True}

        ModelRunnerClass: Type[MLUModelRunnerBase] = MLUModelRunner
        if model_runner_cls is not None:
--- a/vllm-v0.6.2/vllm_mlu/vllm_mlu/attention/backends/mlu_attn.py
+++ b/vllm-v0.6.2/vllm_mlu/vllm_mlu/attention/backends/mlu_attn.py
@@ -580,58 +580,34 @@ def unified_flash_attention_v2(
                                                    value_cache,
                                                    updated_slot_mapping.flatten())
            else:
-                # unpaged (linear cache) path
-                if use_mla:
-                    # MLA: 镜像 paged 路径的处理方式
-                    # key_cache: (num_blocks, 1, block_size, 576)
-                    value_to_cache = None
-                    if attn_metadata.prefill_metadata:
-                        # MLA prefill cache 已在 forward_prefill 中写入，跳过
-                        pass
-                    else:
-                        if kv_cache_dtype == 'int8':
-                            mlu_ops.quant_to_linear_cache(
-                                key, value_to_cache,
-                                key_cache, value_cache,
-                                key_cache_scale, value_cache_scale,
-                                attn_metadata.cu_seq_lens,
-                                attn_metadata.max_seq_len,
-                                True, None,
-                                attn_metadata.batch_ids,
-                                attn_metadata.slot_mapping_unpaged)
-                        else:
-                            mlu_ops.reshape_linear_cache(
-                                key, value_to_cache,
-                                key_cache, value_cache,
-                                attn_metadata.cu_seq_lens,
-                                attn_metadata.max_seq_len,
-                                True, None,
-                                attn_metadata.batch_ids,
-                                attn_metadata.slot_mapping_unpaged)
+                # FIXME: After TMO-1496 is completed, remove this code.
+                if key.stride() != value.stride():
+                    key = key.contiguous()
+                    value = value.contiguous()
+                if kv_cache_dtype == 'int8':
+                    mlu_ops.quant_to_linear_cache(key,
+                                                    value,
+                                                    key_cache,
+                                                    value_cache,
+                                                    key_cache_scale,
+                                                    value_cache_scale,
+                                                    attn_metadata.cu_seq_lens,
+                                                    attn_metadata.max_seq_len,
+                                                    True, # packed
+                                                    None, # context_seq_offset
+                                                    attn_metadata.batch_ids,
+                                                    attn_metadata.slot_mapping_unpaged)
                else:
-                    # FIXME: After TMO-1496 is completed, remove this code.
-                    if key.stride() != value.stride():
-                        key = key.contiguous()
-                        value = value.contiguous()
-                    if kv_cache_dtype == 'int8':
-                        mlu_ops.quant_to_linear_cache(
-                            key, value,
-                            key_cache, value_cache,
-                            key_cache_scale, value_cache_scale,
-                            attn_metadata.cu_seq_lens,
-                            attn_metadata.max_seq_len,
-                            True, None,
-                            attn_metadata.batch_ids,
-                            attn_metadata.slot_mapping_unpaged)
-                    else:
-                        mlu_ops.reshape_linear_cache(
-                            key, value,
-                            key_cache, value_cache,
-                            attn_metadata.cu_seq_lens,
-                            attn_metadata.max_seq_len,
-                            True, None,
-                            attn_metadata.batch_ids,
-                            attn_metadata.slot_mapping_unpaged)
+                    mlu_ops.reshape_linear_cache(key,
+                                                 value,
+                                                 key_cache,
+                                                 value_cache,
+                                                 attn_metadata.cu_seq_lens,
+                                                 attn_metadata.max_seq_len,
+                                                 True, # packed
+                                                 None, # context_seq_offset
+                                                 attn_metadata.batch_ids,
+                                                 attn_metadata.slot_mapping_unpaged)
    if use_mla and attn_metadata.prefill_metadata:
        output = torch.empty(query.shape[0], query.shape[1], v_head_size, dtype=query.dtype, device="mlu")
    else:
--- a/vllm-v0.6.2/vllm_mlu/vllm_mlu/config.py
+++ b/vllm-v0.6.2/vllm_mlu/vllm_mlu/config.py
@@ -37,7 +37,7 @@ def vllm__config__CacheConfig___verify_cache_dtype(self) -> None:

 def vllm__config__ModelConfig__get_num_kv_heads(self, parallel_config: "ParallelConfig") -> int:
    """Returns the number of KV heads per GPU."""
-    if hasattr(self.hf_text_config,"model_type") and self.hf_text_config.model_type in ('deepseek_v2', 'deepseek_v3', 'deepseek_mtp'):
+    if hasattr(self.hf_text_config,"model_type") and self.hf_text_config.model_type == 'deepseek_v2':
        # feature flag MLA
        return 1
    total_num_kv_heads = self.get_total_num_kv_heads()
@@ -51,7 +51,7 @@ def vllm__config__ModelConfig__get_num_kv_heads(self, parallel_config: "Parallel
 def vllm__config__ModelConfig__get_head_size(self) -> int:
    # TODO remove hard code
    if hasattr(self.hf_text_config, "model_type"
-               ) and self.hf_text_config.model_type in ('deepseek_v2', 'deepseek_v3', 'deepseek_mtp'):
+               ) and self.hf_text_config.model_type == 'deepseek_v2':
        '''
        =============================
        Modify by vllm_mlu
@@ -109,7 +109,7 @@ def vllm__config__LoRAConfig__verify_with_model_config(self, model_config: Model
 def vllm__config__ModelConfig__is_deepseek_v2(self) -> bool:
    result = hasattr(
        self.hf_text_config,
-        "model_type") and self.hf_text_config.model_type in ('deepseek_v2', 'deepseek_v3', 'deepseek_mtp')
+        "model_type") and self.hf_text_config.model_type == 'deepseek_v2'
    return result

 MluHijackObject.apply_hijack(ModelConfig,
--- a/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/layers/linear.py
+++ b/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/layers/linear.py
@@ -26,12 +26,6 @@ def vllm__module_executor__layers__linear__UnquantizedLinearMethod__apply(
         beta = 1.0
         residual = residual.view(-1, residual.shape[-1])
    res_shape = x.shape[0:-1] + (layer.weight.shape[0], )
-    # MLU matmul requires all tensors to have matching dtypes
-    target_dtype = layer.weight.dtype
-    if x.dtype != target_dtype:
-        x = x.to(target_dtype)
-    if residual is not None and residual.dtype != target_dtype:
-        residual = residual.to(target_dtype)
    return mlu_ops.matmul(x.view(-1, x.shape[-1]), layer.weight, bias, residual, 'none', 1.0, beta).view(res_shape)


--- a/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/layers/sparse_moe_mlp.py
+++ b/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/layers/sparse_moe_mlp.py
@@ -73,24 +73,6 @@ class SparseMoeMlp(nn.Module):
        self.expert_group = expert_group
        self.topk_group = topk_group
        if get_device_major_capability() == 3:
-            # WARNING: MLU370 (capability=3) 不支持 fused_moe 算子，强制关闭。
-            #
-            # 背景：原始 forward_experts_nofused 包含 torch.unique、torch.tensor([0], ...)、
-            # 数据依赖分支等 graph capture 不兼容操作，导致 MLU370 上所有走 SparseMoeMlp
-            # 的 MoE 模型必须加 --enforce-eager 才能运行。当前已将 forward_experts_nofused
-            # 改为 dense 模式（每个 expert 处理全部 token，用路由权重 mask），解决了
-            # graph capture 兼容性问题，所有 MoE 模型无需 --enforce-eager 即可运行。
-            #
-            # 性能代价：dense 模式计算量为 O(num_experts * num_tokens)，相比稀疏路由的
-            # O(topk * num_tokens) 增大了 num_experts/topk 倍。prefill 阶段对 expert
-            # 数量多的模型会明显变慢，decode 阶段（token 少）影响可忽略。
-            # 已知受影响模型：Mixtral (8)、Qwen2-MoE (60)、HunYuan (16)、Llama4 (16) 等。
-            # DeepSeek V2/V3 不受影响（有独立的 MLU MoE hijack 实现）。
-            #
-            # TODO: MLU370 已有完整的 MoE 算子链（moe_gen_idx、moe_expand_input、
-            # group_gemm、moe_active、moe_combine_result），与 forward_group_experts
-            # 使用的算子相同。后续应拆分 is_use_fused_moe 标志，让 MLU370 走
-            # forward_group_experts 路径以避免 dense 模式的性能开销。
            self.is_use_fused_moe = False

        if params_dtype is None:
@@ -302,28 +284,34 @@ class SparseMoeMlp(nn.Module):


    def forward_experts_nofused(self, hidden_states, expert_logits):
-        # Dense approach: each expert processes ALL tokens, then mask by routing
-        # weights. This avoids data-dependent control flow (variable-size slicing,
-        # conditional branches, torch.unique, torch.tensor creation) that is
-        # incompatible with MLU graph capture.
-        num_tokens, hidden_size = hidden_states.shape
+        hidden_states_shape = hidden_states.shape
        topk_values, topk_indices = self.topk_softmax(expert_logits)
+        expand_gather_idx, scatter_idx, expand_token_count, cusum_token_count = self.generate_gather_idx(
+            topk_indices)
+        # no expert is routed, then expand_gather_idx, expand_scatter_idx has no item,
+        # expand_token_count and expand_cusum_token_count has item but the value is all zero
+        # so this rank should only return final_hidden_states with zero value
+        if expand_gather_idx.numel() == 0:
+            final_hidden_states = torch.zeros_like(hidden_states,
+                                                   dtype=hidden_states.dtype,
+                                                   device=hidden_states.device)
+            return final_hidden_states

-        final_hidden_states = torch.zeros(
-            num_tokens, hidden_size,
-            dtype=hidden_states.dtype, device=hidden_states.device)
+        expand_hidden_states = self.expand_input(hidden_states, expand_gather_idx)

-        for expert_idx in range(self.num_experts_per_rank):
-            global_expert_idx = self.start_expert_id + expert_idx
-            expert_output = self.experts[expert_idx](hidden_states)
-            expert_output = expert_output[0] if isinstance(
-                expert_output, (tuple, list)) else expert_output
-
-            # Routing weight per token for this expert
-            expert_mask = (topk_indices == global_expert_idx).to(topk_values.dtype)
-            expert_weights = (topk_values * expert_mask).sum(dim=-1, keepdim=True)
-
-            final_hidden_states = final_hidden_states + expert_output * expert_weights
+        expand_output_list = []
+        expand_cusum_token_count = cusum_token_count[self.start_expert_id:self.end_expert_id +
+                                                     1] - cusum_token_count[self.start_expert_id]
+        for expert_idx, num_tokens_per_expert in enumerate(expand_token_count):
+            if num_tokens_per_expert > 0:
+                expert_hidden_states = expand_hidden_states[
+                    expand_cusum_token_count[expert_idx]:expand_cusum_token_count[expert_idx + 1]]
+                expert_output = self.experts[expert_idx](expert_hidden_states)
+                expert_output = expert_output[0] if isinstance(expert_output, (tuple, list)) else expert_output
+                expand_output_list.append(expert_output)
+        expand_output = torch.cat(expand_output_list, dim=0)
+        final_hidden_states = self.combine_moe(expand_output, scatter_idx, cusum_token_count, hidden_states_shape,
+                                               topk_values)

        return final_hidden_states

@@ -437,9 +425,9 @@ class SparseMoeMlp(nn.Module):
        scatter_idx=torch.zeros((indices.numel(),), dtype=seqs.dtype, device=seqs.device).scatter(0, indices, seqs)

        # token_count: [self.num_experts_per_rank]
-        # Use scatter_add_ instead of torch.unique for MLU graph capture compatibility
-        token_count = torch.zeros(self.num_total_experts, dtype=sorted_expert_id.dtype, device=device)
-        token_count.scatter_add_(0, sorted_expert_id, torch.ones_like(sorted_expert_id))
+        partial_token_index, partial_token_count = sorted_expert_id.unique(sorted=True, return_counts=True)
+        zero_token_count = torch.zeros(self.num_total_experts, dtype=partial_token_count.dtype, device=device)
+        token_count = zero_token_count.scatter(dim=0, index=partial_token_index, src=partial_token_count)
        # cusum_token_count: [self.num_experts_per_rank + 1]
        cusum_token_count = torch.cat(
            [torch.tensor([0], dtype=token_count.dtype, device=device),
--- a/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/models/init.py
+++ b/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/models/init.py
@@ -39,9 +39,3 @@ try:
 except ImportError as e:
    import logging
    logging.warning(f"Failed to import mllama hijack: {e}")
-
-try:
-    import vllm_mlu.model_executor.models.llama4
-except ImportError as e:
-    import logging
-    logging.warning(f"Failed to import llama4 hijack: {e}")
--- a/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/models/deepseek_v2.py
+++ b/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/models/deepseek_v2.py
@@ -28,7 +28,6 @@ from vllm.model_executor.layers.sampler import SamplerOutput
 from vllm_mlu.model_executor.layers.feed_forward import FeedForward
 from vllm_mlu.mlu_hijack_utils import MluHijackObject
 from vllm_mlu.model_executor.layers.sparse_moe_mlp import SparseMoeMlp
-from vllm import _mlu_ops as mlu_ops
 from vllm.utils import print_warning_once
 from vllm.model_executor.models.utils import is_pp_missing_parameter
 from vllm_mlu.model_executor.models.layer_utils import quant_fusion_with_rmsnorm
@@ -78,12 +77,6 @@ class DeepseekV2MoE(SparseMoeMlp):
                                     bias=False,
                                     quant_config=None,
                                     prefix=f"{prefix}.gate")
-        if getattr(config, "topk_method", None) == "noaux_tc":
-            self.gate.e_score_correction_bias = nn.Parameter(
-                torch.empty(config.n_routed_experts, dtype=torch.float32)
-            )
-        else:
-            self.gate.e_score_correction_bias = None
        if config.n_shared_experts is not None:
            intermediate_size = (config.moe_intermediate_size *
                                 config.n_shared_experts)
@@ -111,7 +104,6 @@ class DeepseekV2MoE(SparseMoeMlp):
    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
        num_tokens, hidden_dim = hidden_states.shape
        hidden_states = hidden_states.view(-1, hidden_dim)
-        shared_output = None
        if self.n_shared_experts is not None:
            shared_output = self.shared_experts(hidden_states)
        # router_logits: (num_tokens, n_experts)
@@ -121,25 +113,9 @@ class DeepseekV2MoE(SparseMoeMlp):
        Modify by vllm_mlu
        =============================
        @brief: replace experts() with forward_experts, which defined by SparseMoeMlp.
-               For noaux_tc (DeepSeek V3), do manual routing with e_score_correction_bias.
        '''
-        if self.gate.e_score_correction_bias is not None:
-            # noaux_tc routing: softmax → add bias for topk selection → use original scores
-            scores = router_logits.float().softmax(dim=-1)
-            scores_for_choice = scores + self.gate.e_score_correction_bias.unsqueeze(0)
-            topk_weights, topk_indices = torch.topk(
-                scores_for_choice, k=self.top_k, dim=-1)
-            # Use original softmax scores (without bias) as weights
-            topk_weights = scores.gather(1, topk_indices)
-            if self.renormalize:
-                topk_weights = topk_weights / topk_weights.sum(
-                    dim=-1, keepdim=True)
-            final_hidden_states = self.forward_experts_with_precomputed_routing(
-                hidden_states, topk_weights, topk_indices
-            ) * self.routed_scaling_factor
-        else:
-            final_hidden_states = self.forward_experts(
-                hidden_states, router_logits) * self.routed_scaling_factor
+        final_hidden_states = self.forward_experts(
+            hidden_states, router_logits) * self.routed_scaling_factor
        '''
        ==================
        End of MLU Hijack
@@ -153,55 +129,6 @@ class DeepseekV2MoE(SparseMoeMlp):

        return final_hidden_states.view(num_tokens, hidden_dim)

-    def forward_experts_with_precomputed_routing(
-        self, hidden_states, topk_weights, topk_indices
-    ):
-        """使用预计算的路由结果执行 MoE 前向传播"""
-        self.pack_params()
-        ori_input_shape = hidden_states.shape
-        expert_num = self.num_total_experts
-        expert_size = self.w13.size(0)
-        max_m = hidden_states.shape[0]
-        hidden_states = hidden_states.view(-1, hidden_states.size(-1))
-
-        reduce_weight = topk_weights.to(torch.float32)
-        expert_id = topk_indices.to(torch.int32)
-
-        # gen_idx
-        expand_idx, combine_idx, token_count, cusum_token_count = (
-            mlu_ops.moe_gen_idx(expert_id, expert_num)
-        )
-
-        start_expert_id = self.start_expert_id
-        # gemm1
-        expand_hidden_states = mlu_ops.moe_expand_input(
-            hidden_states, expand_idx, cusum_token_count,
-            start_expert_id, expert_size
-        )
-        gemm1_out = mlu_ops.group_gemm(
-            expand_hidden_states, self.w13,
-            token_count[start_expert_id:start_expert_id + expert_size],
-            None, None, None, None, max_m
-        )
-        # activation
-        act_out = mlu_ops.moe_active(
-            gemm1_out, self.hidden_act, self.is_gated, None, self.b13,
-            cusum_token_count, start_expert_id, expert_size
-        )
-        # gemm2
-        gemm2_out = mlu_ops.group_gemm(
-            act_out, self.w2,
-            token_count[start_expert_id:start_expert_id + expert_size],
-            None, None, None, None, max_m
-        )
-        # combine
-        output = mlu_ops.moe_combine_result(
-            gemm2_out, reduce_weight, combine_idx,
-            None, cusum_token_count, start_expert_id,
-            expert_size, self.b2
-        )
-        return output.view(ori_input_shape)
-
 def forward_prefill(
    self,
    positions: torch.Tensor,
@@ -252,27 +179,19 @@ def forward_prefill(
        updated_slot_mapping = attn_metadata.slot_mapping
        if self.attn.kv_cache_dtype == 'int8':
            key_cache_scale = kv_cache[1][0]
-            mlu_ops.quant_to_linear_cache(key_value,
-                                          None,
-                                          key_cache,
-                                          None,
-                                          key_cache_scale,
-                                          None,
-                                          attn_metadata.cu_seq_lens,
-                                          attn_metadata.max_seq_len,
-                                          True, None,
-                                          attn_metadata.batch_ids,
-                                          attn_metadata.slot_mapping_unpaged)
-        else:
-            mlu_ops.reshape_linear_cache(key_value,
+            mlu_ops.quant_to_paged_cache(key_value,
                                         None,
                                         key_cache,
                                         None,
-                                         attn_metadata.cu_seq_lens,
-                                         attn_metadata.max_seq_len,
-                                         True, None,
-                                         attn_metadata.batch_ids,
-                                         attn_metadata.slot_mapping_unpaged)
+                                         key_cache_scale,
+                                         None,
+                                         updated_slot_mapping.flatten())
+        else:
+            mlu_ops.reshape_paged_cache(key_value,
+                                        None,
+                                        key_cache,
+                                        None,
+                                        updated_slot_mapping.flatten())
    '''
    ==================
    End of MLU Hijack
@@ -572,15 +491,6 @@ def vllm__module_executor__models__deepseek_v2__DeepseekV2DecoderLayer__init__(
    self.post_attention_layernorm = RMSNorm(config.hidden_size,
                                            eps=config.rms_norm_eps)

-def get_spec_layer_idx_from_weight_name(config, weight_name):
-    num_nextn = getattr(config, "num_nextn_predict_layers", 0)
-    if num_nextn and num_nextn > 0:
-        layer_idx = config.num_hidden_layers
-        for i in range(num_nextn):
-            if weight_name.startswith(f"model.layers.{layer_idx + i}."):
-                return layer_idx + i
-    return None
-
 def vllm__module_executor__models__deepseek_v2__DeepseekV2ForCausalLM__load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
    '''
    =============================
@@ -620,10 +530,6 @@ def vllm__module_executor__models__deepseek_v2__DeepseekV2ForCausalLM__load_weig
    for name, loaded_weight in weights:
        if "rotary_emb.inv_freq" in name:
            continue
-        # Skip MTP speculative decoding layer weights
-        spec_layer = get_spec_layer_idx_from_weight_name(self.config, name)
-        if spec_layer is not None:
-            continue
        '''
        =============================
        Modify by vllm_mlu
@@ -659,9 +565,7 @@ def vllm__module_executor__models__deepseek_v2__DeepseekV2ForCausalLM__load_weig
            @brief: add expert skiped condition and delete useless if name not in params_dict: continue condition
            '''
            name = name.replace(weight_name, param_name)
-            if (("mlp.experts." in name or "mlp.shared_experts." in name
-                    or "mlp.shared_expert_gate." in name
-                    or "e_score_correction_bias" in name)
+            if (("mlp.experts." in name or "mlp.shared_experts." in name or "mlp.shared_expert_gate." in name)
                    and name not in params_dict):
                continue
            '''
@@ -691,9 +595,7 @@ def vllm__module_executor__models__deepseek_v2__DeepseekV2ForCausalLM__load_weig
            if name.endswith(".bias") and name not in params_dict:
                continue

-            if (("mlp.experts." in name or "mlp.shared_experts." in name
-                    or "mlp.shared_expert_gate." in name
-                    or "e_score_correction_bias" in name)
+            if (("mlp.experts." in name or "mlp.shared_experts." in name or "mlp.shared_expert_gate." in name)
                    and name not in params_dict):
                continue
            if is_pp_missing_parameter(name, self):
--- a/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/models/layer_utils.py
+++ b/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/models/layer_utils.py
@@ -194,11 +194,6 @@ def decoder_model_forward_base_pp(
            hidden_states = inputs_embeds
        else:
            hidden_states = get_input_embeddings(input_ids)
-        # MLU F.embedding may output float32 even with float16 weights;
-        # cast to model dtype to avoid dtype mismatches downstream.
-        target_dtype = next(layers[start_layer].parameters()).dtype
-        if hidden_states.dtype != target_dtype:
-            hidden_states = hidden_states.to(target_dtype)
    else:
        assert intermediate_tensors is not None
        hidden_states = intermediate_tensors["hidden_states"]
--- a/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/models/llama4.py
+++ b/vllm-v0.6.2/vllm_mlu/vllm_mlu/model_executor/models/llama4.py
@@ -1,495 +0,0 @@
-import torch
-import re
-
-from typing import List, Optional, Tuple, Union, Iterable
-from vllm.attention import AttentionMetadata
-from vllm.config import CacheConfig
-from vllm.model_executor.layers.quantization.base_config import QuantizationConfig
-from vllm.model_executor.layers.layernorm import RMSNorm
-from vllm_mlu.model_executor.layers.feed_forward import FeedForward
-from vllm_mlu.mlu_hijack_utils import MluHijackObject
-from vllm.distributed import tensor_model_parallel_all_reduce
-from vllm.model_executor.model_loader.weight_utils import (
-    default_weight_loader, maybe_remap_kv_scale_name)
-from vllm.model_executor.models.llama4 import (
-    Llama4Attention, Llama4DecoderLayer, Llama4ForCausalLM,
-    Llama4Model, Llama4MoE)
-from vllm_mlu.model_executor.layers.sparse_moe_mlp import SparseMoeMlp
-from vllm.model_executor.models.utils import is_pp_missing_parameter
-from vllm.sequence import IntermediateTensors
-
-from vllm_mlu.model_executor.models.layer_utils import (
-    decoder_layer_forward_base, decoder_model_forward_base_pp,
-    is_per_tensor_smoothquant, is_per_token_smoothquant,
-    quant_fusion_with_rmsnorm)
-
-from vllm.logger import init_logger
-
-logger = init_logger(__name__)
-
-# ============================================================
-# Llama4MoE MLU replacement: SparseMoeMlp + shared expert
-# ============================================================
-
-class Llama4MoEMlu(SparseMoeMlp):
-    """MLU replacement for Llama4MoE using SparseMoeMlp + shared expert."""
-
-    def __init__(self, config, quant_config=None, prefix=""):
-        num_local_experts = getattr(config, "num_local_experts", 8)
-        top_k = getattr(config, "num_experts_per_tok", 1)
-        hidden_size = getattr(config, "hidden_size", 4096)
-        intermediate_size = getattr(config, "intermediate_size", 8192)
-
-        super().__init__(
-            num_experts=num_local_experts,
-            top_k=top_k,
-            hidden_size=hidden_size,
-            intermediate_size=intermediate_size,
-            up_proj_name="gate_up_proj",
-            is_gated=True,
-            down_proj_name="down_proj",
-            has_bias=False,
-            skip_bias_add=False,
-            renormalize=False,
-            hidden_act="silu",
-            params_dtype=None,
-            quant_config=quant_config,
-            is_use_fused_moe=True,
-        )
-
-        # Llama4 uses sigmoid routing, not softmax
-        # Override topk_softmax to use sigmoid
-        self._use_sigmoid_routing = True
-
-        # Shared expert (independent from routed experts)
-        self.shared_expert = FeedForward(
-            hidden_size=hidden_size,
-            intermediate_size=intermediate_size,
-            hidden_act="silu",
-            up_proj_name="gate_up_proj",
-            is_gated=True,
-            down_proj_name="down_proj",
-            bias=False,
-            quant_config=quant_config,
-            reduce_results=False,
-            prefix=f"{prefix}.shared_expert",
-        )
-
-    def topk_softmax(self, expert_logits):
-        """Override: Llama4 uses sigmoid routing instead of softmax."""
-        topk_values, topk_indices = torch.topk(
-            expert_logits, self.top_k, dim=-1)
-        topk_values = torch.sigmoid(topk_values.float())
-        return topk_values, topk_indices
-
-    def forward(self, hidden_states, residual=None):
-        orig_shape = hidden_states.shape
-        hidden_states = hidden_states.view(-1, self.hidden_size)
-
-        # Shared expert output
-        shared_out = self.shared_expert(hidden_states)
-
-        # Router logits
-        router_logits, _ = self.gate(hidden_states)
-
-        # Routed experts
-        routed_out = self.forward_experts(hidden_states, router_logits, None)
-
-        # Combine
-        final_out = routed_out + shared_out
-        if self.tp_size > 1:
-            final_out = tensor_model_parallel_all_reduce(final_out)
-
-        return final_out.view(orig_shape)
-
-
-# ============================================================
-# Llama4Attention hijack
-# ============================================================
-
-vllm__llama4__Llama4Attention__init__org = Llama4Attention.__init__
-
-
-def vllm__llama4__Llama4Attention____init__(
-    self,
-    config,
-    hidden_size: int,
-    num_heads: int,
-    num_kv_heads: int,
-    max_position_embeddings: int = 8192,
-    quant_config: Optional[QuantizationConfig] = None,
-    bias: bool = False,
-    cache_config: Optional[CacheConfig] = None,
-    prefix: str = "",
-) -> None:
-    vllm__llama4__Llama4Attention__init__org(
-        self, config, hidden_size, num_heads, num_kv_heads,
-        max_position_embeddings, quant_config, bias, cache_config, prefix)
-    '''
-    =============================
-    Modify by vllm_mlu
-    =============================
-    @brief: save rope_scaling for MLU RoPE dispatch
-    '''
-    self.rope_scaling = getattr(config, "rope_scaling", None)
-    '''
-    ==================
-    End of MLU Hijack
-    ==================
-    '''
-
-
-def vllm__llama4__Llama4Attention__forward(
-    self,
-    positions: torch.Tensor,
-    hidden_states: torch.Tensor,
-    kv_cache: torch.Tensor,
-    attn_metadata: AttentionMetadata,
-    residual: Optional[torch.Tensor] = None,
-    smooth_quant_scale: Optional[torch.Tensor] = None,
-) -> torch.Tensor:
-    qkv, _ = self.qkv_proj(hidden_states, smooth_quant_scale)
-    q, k, v = qkv.split([self.q_size, self.kv_size, self.kv_size], dim=-1)
-
-    '''
-    =============================
-    Modify by vllm_mlu
-    =============================
-    @brief: MLU RoPE: merge q/k, apply rotary, split back (教训 #3)
-    For NoPE layers (self.rotary_emb is None), skip RoPE entirely.
-    '''
-    if self.rotary_emb is not None:
-        if (self.rope_scaling is not None
-                and self.rope_scaling.get("rope_type") == "longrope"):
-            q, k = self.rotary_emb(positions, q, k)
-        else:
-            qk, _ = qkv.split(
-                [self.q_size + self.kv_size, self.kv_size], dim=-1)
-            self.rotary_emb(
-                positions,
-                qk.view(-1, self.num_heads + self.num_kv_heads,
-                         self.head_dim))
-            q, k = qk.split([self.q_size, self.kv_size], dim=-1)
-    '''
-    ==================
-    End of MLU Hijack
-    ==================
-    '''
-
-    # QK norm (MLU fused_rms_norm requires matching dtypes, skip .float())
-    if self.qk_norm is not None:
-        q = q.contiguous().reshape(-1, self.head_dim)
-        q = (self.qk_norm(q)
-             .contiguous().reshape(-1, self.q_size))
-        k = k.contiguous().reshape(-1, self.head_dim)
-        k = (self.qk_norm(k)
-             .contiguous().reshape(-1, self.kv_size))
-
-    # Temperature tuning for NoPE layers
-    if self.attn_temperature_tuning and self.nope:
-        attn_scale = self._get_attn_scale(positions)
-        q = (q * attn_scale).to(q.dtype)
-
-    attn_output = self.attn(q, k, v, kv_cache, attn_metadata)
-
-    '''
-    =============================
-    Modify by vllm_mlu
-    =============================
-    @brief: add residual in o_proj
-    '''
-    output, _ = self.o_proj(attn_output, residual)
-    '''
-    ==================
-    End of MLU Hijack
-    ==================
-    '''
-    return output
-
-
-# ============================================================
-# Llama4DecoderLayer hijack
-# ============================================================
-
-def vllm__llama4__Llama4DecoderLayer____init__(
-    self,
-    config,
-    cache_config: Optional[CacheConfig] = None,
-    quant_config: Optional[QuantizationConfig] = None,
-    prefix: str = "",
-) -> None:
-    super(Llama4DecoderLayer, self).__init__()
-    from vllm.model_executor.models.llama4 import (
-        _extract_layer_index, Llama4Attention)
-
-    self.layer_idx = _extract_layer_index(prefix)
-    self.hidden_size = getattr(config, "hidden_size", 4096)
-    max_position_embeddings = getattr(
-        config, "max_position_embeddings", 8192)
-
-    self.self_attn = Llama4Attention(
-        config=config,
-        hidden_size=self.hidden_size,
-        num_heads=getattr(config, "num_attention_heads", 32),
-        num_kv_heads=getattr(config, "num_key_value_heads",
-                             getattr(config, "num_attention_heads", 32)),
-        max_position_embeddings=max_position_embeddings,
-        quant_config=quant_config,
-        bias=False,
-        cache_config=cache_config,
-        prefix=f"{prefix}.self_attn",
-    )
-
-    interleave_moe_layer_step = getattr(
-        config, "interleave_moe_layer_step", 0)
-    is_moe_layer = (interleave_moe_layer_step > 0
-                    and (self.layer_idx + 1)
-                    % interleave_moe_layer_step == 0)
-
-    '''
-    =============================
-    Modify by vllm_mlu
-    =============================
-    @brief: Replace MoE with Llama4MoEMlu (SparseMoeMlp + shared expert),
-            Replace dense MLP with FeedForward.
-    '''
-    if is_moe_layer:
-        self.feed_forward = Llama4MoEMlu(
-            config=config,
-            quant_config=quant_config,
-            prefix=f"{prefix}.feed_forward",
-        )
-    else:
-        intermediate_size_mlp = getattr(
-            config, "intermediate_size_mlp",
-            getattr(config, "intermediate_size", 8192))
-        self.feed_forward = FeedForward(
-            hidden_size=self.hidden_size,
-            intermediate_size=intermediate_size_mlp,
-            hidden_act="silu",
-            up_proj_name="gate_up_proj",
-            is_gated=True,
-            down_proj_name="down_proj",
-            bias=False,
-            quant_config=quant_config,
-            prefix=f"{prefix}.feed_forward",
-        )
-    '''
-    ==================
-    End of MLU Hijack
-    ==================
-    '''
-
-    rms_norm_eps = getattr(config, "rms_norm_eps", 1e-5)
-    self.input_layernorm = RMSNorm(self.hidden_size, eps=rms_norm_eps)
-    self.post_attention_layernorm = RMSNorm(
-        self.hidden_size, eps=rms_norm_eps)
-
-    self.is_per_tesnor_sq_perf_cases = is_per_tensor_smoothquant(
-        quant_config)
-    self.is_per_token_sq_perf_cases = is_per_token_smoothquant(
-        quant_config)
-    if self.is_per_tesnor_sq_perf_cases or self.is_per_token_sq_perf_cases:
-        self.self_attn.qkv_proj.quant_method.skip_quant_input = True
-        self.quant_fusion_attn_layernorm = None
-
-
-def vllm__llama4__Llama4DecoderLayer__forward(
-    self,
-    positions: torch.Tensor,
-    hidden_states: torch.Tensor,
-    kv_cache: torch.Tensor,
-    attn_metadata: AttentionMetadata,
-) -> torch.Tensor:
-    '''
-    =============================
-    Modify by vllm_mlu
-    =============================
-    @brief: use decoder_layer_forward_base with residual-in-matmul
-            and optional quant fusion.
-    '''
-    attn_layernorm = self.input_layernorm
-    if self.is_per_tesnor_sq_perf_cases or self.is_per_token_sq_perf_cases:
-        if self.quant_fusion_attn_layernorm is None:
-            if self.is_per_token_sq_perf_cases:
-                attn_quant_scale = self.self_attn.qkv_proj.smooth
-            else:
-                attn_quant_scale = self.self_attn.qkv_proj.scale_to_int
-            self.quant_fusion_attn_layernorm = quant_fusion_with_rmsnorm(
-                self.input_layernorm, attn_quant_scale,
-                dynamic_quant=self.is_per_token_sq_perf_cases)
-        attn_layernorm = self.quant_fusion_attn_layernorm
-
-    return decoder_layer_forward_base(
-        positions, hidden_states, kv_cache, attn_metadata,
-        attn_layernorm,
-        self.self_attn,
-        self.post_attention_layernorm,
-        self.feed_forward,
-        input_norm_fuse_en=self.is_per_token_sq_perf_cases)
-
-
-# ============================================================
-# Llama4Model hijack
-# ============================================================
-
-def vllm__llama4__Llama4Model__forward(
-    self,
-    input_ids: Optional[torch.Tensor],
-    positions: torch.Tensor,
-    kv_caches: List[torch.Tensor],
-    attn_metadata: AttentionMetadata,
-    intermediate_tensors: Optional[IntermediateTensors],
-    inputs_embeds: Optional[torch.Tensor] = None,
-) -> Union[torch.Tensor, IntermediateTensors]:
-    return decoder_model_forward_base_pp(
-        input_ids, positions, kv_caches, attn_metadata,
-        intermediate_tensors,
-        self.layers, self.start_layer, self.end_layer,
-        self.get_input_embeddings,
-        self.norm,
-        inputs_embeds)
-
-
-# ============================================================
-# Llama4ForCausalLM load_weights hijack
-# ============================================================
-
-def vllm__llama4__Llama4ForCausalLM__load_weights(
-    self,
-    weights: Iterable[Tuple[str, torch.Tensor]],
-):
-    '''
-    =============================
-    Modify by vllm_mlu
-    =============================
-    @brief: pack params for SparseMoeMlp (MoE layers)
-    '''
-    for name, m in self.model.named_modules():
-        if isinstance(m, SparseMoeMlp):
-            m.pack_params()
-
-    start_expert_id = 0
-    '''
-    ==================
-    End of MLU Hijack
-    ==================
-    '''
-
-    stacked_params_mapping = [
-        # (param_name, shard_name, shard_id)
-        ("qkv_proj", "q_proj", "q"),
-        ("qkv_proj", "k_proj", "k"),
-        ("qkv_proj", "v_proj", "v"),
-        ("gate_up_proj", "gate_proj", 0),
-        ("gate_up_proj", "up_proj", 1),
-    ]
-
-    params_dict = dict(self.named_parameters())
-    for name, loaded_weight in weights:
-        if "rotary_emb.inv_freq" in name:
-            continue
-
-        # Strip language_model. prefix for Llama4ForConditionalGeneration
-        if name.startswith("language_model."):
-            name = name[len("language_model."):]
-        # Skip vision encoder weights
-        elif (name.startswith("multi_modal_projector.")
-              or name.startswith("vision_encoder.")
-              or name.startswith("vision_model.")):
-            continue
-
-        # Permute Q/K weights for rotary embedding
-        name, loaded_weight = self.permute_qk_weight_for_rotary(
-            name, loaded_weight)
-
-        '''
-        =============================
-        Modify by vllm_mlu
-        =============================
-        @brief: remap expert_id for distributed inference
-        '''
-        if (start_expert_id > 0
-                and "feed_forward.experts." in name):
-            match = re.search(r'experts\.\d+', name)
-            if match:
-                expert_str = match.group(0)
-                expert_id = int(expert_str.split(".")[1])
-                named_expert_id = expert_id - start_expert_id
-                name = name.replace(
-                    f"experts.{expert_id}",
-                    f"experts.{named_expert_id}")
-        '''
-        ==================
-        End of MLU Hijack
-        ==================
-        '''
-
-        for (param_name, weight_name, shard_id) in stacked_params_mapping:
-            if weight_name not in name:
-                continue
-            name = name.replace(weight_name, param_name)
-            if ((name.endswith(".bias") or name.endswith("_bias"))
-                    and name not in params_dict):
-                continue
-            if is_pp_missing_parameter(name, self):
-                continue
-            # Skip experts not assigned to this worker
-            if ("feed_forward.experts." in name
-                    and name not in params_dict):
-                continue
-            param = params_dict[name]
-            weight_loader = param.weight_loader
-            weight_loader(param, loaded_weight, shard_id)
-            break
-        else:
-            if ((name.endswith(".bias") or name.endswith("_bias"))
-                    and name not in params_dict):
-                continue
-            if is_pp_missing_parameter(name, self):
-                continue
-            name = maybe_remap_kv_scale_name(name, params_dict)
-            if name is None:
-                continue
-            # Skip experts not assigned to this worker
-            if ("feed_forward.experts." in name
-                    and name not in params_dict):
-                continue
-            if name not in params_dict:
-                logger.warning(
-                    "Skipping weight %s not present in the model", name)
-                continue
-            param = params_dict[name]
-            weight_loader = getattr(
-                param, "weight_loader", default_weight_loader)
-            weight_loader(param, loaded_weight)
-
-
-# ============================================================
-# Apply all hijacks
-# ============================================================
-
-MluHijackObject.apply_hijack(
-    Llama4Attention,
-    Llama4Attention.__init__,
-    vllm__llama4__Llama4Attention____init__)
-MluHijackObject.apply_hijack(
-    Llama4Attention,
-    Llama4Attention.forward,
-    vllm__llama4__Llama4Attention__forward)
-MluHijackObject.apply_hijack(
-    Llama4DecoderLayer,
-    Llama4DecoderLayer.__init__,
-    vllm__llama4__Llama4DecoderLayer____init__)
-MluHijackObject.apply_hijack(
-    Llama4DecoderLayer,
-    Llama4DecoderLayer.forward,
-    vllm__llama4__Llama4DecoderLayer__forward)
-MluHijackObject.apply_hijack(
-    Llama4Model,
-    Llama4Model.forward,
-    vllm__llama4__Llama4Model__forward)
-MluHijackObject.apply_hijack(
-    Llama4ForCausalLM,
-    Llama4ForCausalLM.load_weights,
-    vllm__llama4__Llama4ForCausalLM__load_weights)
Author	SHA1	Message	Date
Chranos	934ed88691	add qwen3_moe	2026-02-10 18:30:48 +08:00
Chranos	fa0219fbf8	add qwen3_moe	2026-02-10 18:22:13 +08:00
Chranos	efbb06147a	add qwen3_moe	2026-02-10 18:18:32 +08:00
Chranos	a26729bf7f	add qwen3_moe	2026-02-10 18:09:58 +08:00
Chranos	8a613d15bd	add qwen3_moe	2026-02-10 18:02:40 +08:00
Chranos	a6f39375e5	debugging	2026-02-10 16:10:28 +08:00
Chranos	afc34d988e	debugging	2026-02-10 15:47:48 +08:00
Chranos	fa194c215b	add gemma3	2026-02-10 14:52:56 +08:00
Chranos	5fbe8b20a7	add gemma3	2026-02-10 14:26:03 +08:00
Chranos	2dad4e71c5	add gemma3	2026-02-10 14:15:33 +08:00
Chranos	cb1846cd4f	add gemma3	2026-02-10 14:10:04 +08:00
Chranos	81fc273396	add gemma3	2026-02-10 14:06:26 +08:00
Chranos	3ef89630ab	add gemma3	2026-02-10 13:00:25 +08:00
Chranos	40dee08f7b	fix: handle missing tie_word_embeddings attr in MPTConfig Use getattr with default True for MPTConfig.tie_word_embeddings, as some MPT model configs lack this attribute. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-02-09 17:47:18 +08:00
Chranos	1d70f93cfc	debugging	2026-02-09 15:24:55 +08:00
Chranos	8ecba6115e	fix: add logger import to llama.py for unknown weight skip warning The previous commit added a warning log for skipping unknown weights (e.g. embed_tokens.biases) but missed importing the logger. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-02-09 13:13:56 +08:00
Chranos	65ad893ee7	debugging	2026-02-09 13:00:35 +08:00
Chranos	d08217307d	update README	2026-02-09 11:46:04 +08:00
Chranos	8ac4215755	update README	2026-02-09 11:44:52 +08:00
Chranos	a095dede48	fixed kvcache bug	2026-02-06 17:10:36 +08:00
Chranos	374826c841	fixing kvcache bug	2026-02-06 16:25:54 +08:00
Chranos	ebdc6fed03	fix: pass lm_head to LogitsProcessor instead of calling forward() In vLLM v0.6.2, ParallelLMHead.forward() raises RuntimeError since its weights should be used through LogitsProcessor.linear_method.apply(). Pass lm_head as first arg to LogitsProcessor which handles the hidden_states -> logits projection internally.	2026-02-06 14:21:14 +08:00
Chranos	b702adf015	testing dynamic register	2026-02-06 14:17:06 +08:00
Chranos	fba02652c8	testing dynamic register	2026-02-06 14:04:04 +08:00
Chranos	5d2f4000cc	testing dynamic register	2026-02-06 13:51:02 +08:00
Chranos	f088a6b45d	testing dynamic register	2026-02-06 13:39:13 +08:00
Chranos	d31ace279b	testing dynamic register	2026-02-05 18:57:04 +08:00
Chranos	ac2082ff36	testing dynamic register	2026-02-05 18:48:11 +08:00
Chranos	2068984bde	testing dynamic register	2026-02-05 18:36:03 +08:00
Chranos	df848b4284	testing dynamic register	2026-02-05 18:24:33 +08:00
Chranos	4d0da98b9e	testing dynamic register	2026-02-05 18:21:31 +08:00
Chranos	05605419e3	testing dynamic register	2026-02-05 18:08:05 +08:00
Chranos	332e5f71a6	testing dynamic register	2026-02-05 18:02:59 +08:00
Chranos	6e38461af6	testing dynamic register	2026-02-05 17:11:09 +08:00
Chranos	b399840b8d	testing dynamic register	2026-02-05 16:30:44 +08:00
Chranos	808b9b7c97	删除 .DS_Store	2026-02-05 16:20:54 +08:00