[Model Refactoring] Migrate DeepSeek V4 to vllm/models/ [1/N]

tests/models/test_deepseek_v4_mega_moe.py

@@ -6,7 +6,7 @@
 import pytest
 import torch
 
-from vllm.model_executor.models.deepseek_v4 import (
+from vllm.models.deepseek_v4.nvidia.deepseek_v4 import (
     DeepseekV4MegaMoEExperts,
     _stage_deepseek_v4_mega_moe_inputs,
     make_deepseek_v4_expert_params_mapping,

vllm/model_executor/layers/quantization/__init__.py

@@ -113,7 +113,7 @@ def get_quantization_config(quantization: str) -> type[QuantizationConfig]:
     # lazy import to avoid triggering `torch.compile` too early
     from vllm.config.quantization import _ONLINE_SHORTHANDS
     from vllm.model_executor.layers.quantization.quark.quark import QuarkConfig
-    from vllm.model_executor.models.deepseek_v4 import DeepseekV4FP8Config
+    from vllm.models.deepseek_v4 import DeepseekV4FP8Config
 
     from .auto_gptq import AutoGPTQConfig
     from .awq import AWQConfig

vllm/model_executor/models/registry.py

@@ -6,6 +6,7 @@
 """
 
 import importlib
+import importlib.util
 import json
 import os
 import pickle
@@ -97,7 +98,7 @@
     "DeepseekV2ForCausalLM": ("deepseek_v2", "DeepseekV2ForCausalLM"),
     "DeepseekV3ForCausalLM": ("deepseek_v2", "DeepseekV3ForCausalLM"),
     "DeepseekV32ForCausalLM": ("deepseek_v2", "DeepseekV3ForCausalLM"),
-    "DeepseekV4ForCausalLM": ("deepseek_v4", "DeepseekV4ForCausalLM"),
+    "DeepseekV4ForCausalLM": ("vllm.models.deepseek_v4", "DeepseekV4ForCausalLM"),
     "Dots1ForCausalLM": ("dots1", "Dots1ForCausalLM"),
     "Ernie4_5ForCausalLM": ("ernie45", "Ernie4_5ForCausalLM"),
     "Ernie4_5_MoeForCausalLM": ("ernie45_moe", "Ernie4_5_MoeForCausalLM"),
@@ -611,7 +612,7 @@
     "Eagle3DeepseekV3ForCausalLM": ("deepseek_eagle3", "Eagle3DeepseekV2ForCausalLM"),
     "EagleDeepSeekMTPModel": ("deepseek_eagle", "EagleDeepseekV3ForCausalLM"),
     "DeepSeekMTPModel": ("deepseek_mtp", "DeepSeekMTP"),
-    "DeepSeekV4MTPModel": ("deepseek_v4_mtp", "DeepSeekV4MTP"),
+    "DeepSeekV4MTPModel": ("vllm.models.deepseek_v4", "DeepSeekV4MTP"),
     "Gemma4MTPModel": ("gemma4_mtp", "Gemma4MTP"),
     "ErnieMTPModel": ("ernie_mtp", "ErnieMTP"),
     "ExaoneMoeMTP": ("exaone_moe_mtp", "ExaoneMoeMTP"),
@@ -870,10 +871,21 @@ def _save_modelinfo_to_cache(self, mi: _ModelInfo, module_hash: str) -> None:
 
     @logtime(logger=logger, msg="Registry inspect model class")
     def inspect_model_cls(self) -> _ModelInfo:
-        model_path = Path(__file__).parent / f"{self.module_name.split('.')[-1]}.py"
+        # Modules registered with a non-default location (e.g. the
+        # hardware-isolated ``vllm.models.<name>`` layout) live outside
+        # ``vllm/model_executor/models``. Resolve the module spec directly
+        # so the file-hash cache stays warm for them.
+        if self.module_name.startswith("vllm.model_executor.models."):
+            model_path = Path(__file__).parent / f"{self.module_name.split('.')[-1]}.py"
+        else:
+            try:
+                spec = importlib.util.find_spec(self.module_name)
+            except (ImportError, ValueError):
+                spec = None
+            model_path = Path(spec.origin) if spec is not None and spec.origin else None
         module_hash = None
 
-        if model_path.exists():
+        if model_path is not None and model_path.exists():
             with open(model_path, "rb") as f:
                 module_hash = safe_hash(f.read(), usedforsecurity=False).hexdigest()
 
@@ -1328,10 +1340,19 @@ def is_transcription_only_model(
         return model_cls.supports_transcription_only
 
 
+def _resolve_module_name(mod_relname: str) -> str:
+    # Allow registry entries to point at fully-qualified module paths (e.g.
+    # ``vllm.models.deepseek_v4``) for models that live outside the legacy
+    # ``vllm.model_executor.models`` flat layout.
+    if mod_relname.startswith("vllm."):
+        return mod_relname
+    return f"vllm.model_executor.models.{mod_relname}"
+
+
 ModelRegistry = _ModelRegistry(
     {
         model_arch: _LazyRegisteredModel(
-            module_name=f"vllm.model_executor.models.{mod_relname}",
+            module_name=_resolve_module_name(mod_relname),
             class_name=cls_name,
         )
         for model_arch, (mod_relname, cls_name) in _VLLM_MODELS.items()

vllm/models/__init__.py

@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project

vllm/models/deepseek_v4/__init__.py

@@ -0,0 +1,31 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+"""DeepSeek V4 model — hardware-isolated entry point.
+
+The actual implementation lives under ``nvidia/`` and ``amd/``; this module
+picks the right one for the current platform and re-exports the public
+classes used by the model registry and quantization config lookup.
+"""
+
+from typing import TYPE_CHECKING
+
+from vllm.platforms import current_platform
+
+# Pick the per-platform implementation. The NVIDIA branch is the static
+# default that mypy sees; the ROCm branch overrides it at runtime and is
+# kept type-compatible via ``# type: ignore[assignment]``.
+if TYPE_CHECKING or not current_platform.is_rocm():
+    from .nvidia.deepseek_v4 import DeepseekV4ForCausalLM, DeepseekV4FP8Config
+    from .nvidia.deepseek_v4_mtp import DeepSeekV4MTP
+else:
+    from .amd.deepseek_v4 import (  # type: ignore[assignment]
+        DeepseekV4ForCausalLM,
+        DeepseekV4FP8Config,
+    )
+    from .amd.deepseek_v4_mtp import DeepSeekV4MTP  # type: ignore[assignment]
+
+__all__ = [
+    "DeepSeekV4MTP",
+    "DeepseekV4FP8Config",
+    "DeepseekV4ForCausalLM",
+]

vllm/models/deepseek_v4/amd/__init__.py

@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project

vllm/models/deepseek_v4/amd/deepseek_v4.py

@@ -0,0 +1 @@
+../nvidia/deepseek_v4.py

vllm/models/deepseek_v4/amd/deepseek_v4_mtp.py

@@ -0,0 +1 @@
+../nvidia/deepseek_v4_mtp.py

vllm/models/deepseek_v4/nvidia/__init__.py

@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project

vllm/model_executor/models/deepseek_v4.py → vllm/models/deepseek_v4/nvidia/deepseek_v4.py

@@ -60,13 +60,7 @@
 )
 from vllm.model_executor.model_loader.weight_utils import default_weight_loader
 from vllm.model_executor.models.interfaces import SupportsPP
-from vllm.model_executor.utils import set_weight_attrs
-from vllm.platforms import current_platform
-from vllm.sequence import IntermediateTensors
-from vllm.triton_utils import tl, triton
-from vllm.utils.torch_utils import direct_register_custom_op
-
-from .utils import (
+from vllm.model_executor.models.utils import (
     AutoWeightsLoader,
     PPMissingLayer,
     WeightsMapper,
@@ -75,6 +69,11 @@
     make_layers,
     maybe_prefix,
 )
+from vllm.model_executor.utils import set_weight_attrs
+from vllm.platforms import current_platform
+from vllm.sequence import IntermediateTensors
+from vllm.triton_utils import tl, triton
+from vllm.utils.torch_utils import direct_register_custom_op
 
 _DEEPSEEK_V4_EXPERT_DTYPES = ("fp4", "fp8")
 
@@ -1535,7 +1534,7 @@ def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
                     ):
                         loaded_weight = loaded_weight.view(torch.uint8)
                     for mapping in expert_mapping:
-                        param_name, weight_name, expert_id, shard_id = mapping
+                        param_name, weight_name, expert_id, shard_id = mapping  # type: ignore[assignment]
                         if weight_name not in name:
                             continue
                         name_mapped = name.replace(weight_name, param_name)
@@ -1673,7 +1672,7 @@ def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
         else:
             self.lm_head = PPMissingLayer()
         self.logits_processor = LogitsProcessor(config.vocab_size)
-        self.make_empty_intermediate_tensors = (
+        self.make_empty_intermediate_tensors = (  # type: ignore[method-assign]
             self.model.make_empty_intermediate_tensors
         )
 

vllm/model_executor/models/deepseek_v4_mtp.py → vllm/models/deepseek_v4/nvidia/deepseek_v4_mtp.py

@@ -34,16 +34,16 @@
     VocabParallelEmbedding,
 )
 from vllm.model_executor.model_loader.weight_utils import default_weight_loader
+from vllm.model_executor.models.deepseek_mtp import SharedHead
+from vllm.model_executor.models.deepseek_v2 import get_spec_layer_idx_from_weight_name
+from vllm.model_executor.models.utils import maybe_prefix
 from vllm.platforms import current_platform
 from vllm.sequence import IntermediateTensors
 
-from .deepseek_mtp import SharedHead
-from .deepseek_v2 import get_spec_layer_idx_from_weight_name
 from .deepseek_v4 import (
     DeepseekV4DecoderLayer,
     make_deepseek_v4_expert_params_mapping,
 )
-from .utils import maybe_prefix
 
 logger = init_logger(__name__)
 
@@ -68,7 +68,7 @@ def __init__(
     ) -> None:
         super().__init__()
 
-        config = vllm_config.speculative_config.draft_model_config.hf_config
+        config = vllm_config.speculative_config.draft_model_config.hf_config  # type: ignore[union-attr]
         self.config = config
         quant_config = vllm_config.quant_config
         self.rms_norm_eps = config.rms_norm_eps
@@ -407,7 +407,7 @@ def _find_mtp_layer_idx(name: str) -> int:
                     ):
                         loaded_weight = loaded_weight.view(torch.uint8)
                     for mapping in expert_mapping:
-                        param_name, weight_name, expert_id, shard_id = mapping
+                        param_name, weight_name, expert_id, shard_id = mapping  # type: ignore[assignment]
                         if weight_name not in name:
                             continue
                         name_mapped = name.replace(weight_name, param_name)