[Feature][EPLB] Add support for Qwen3 EPLB

vllm/model_executor/models/qwen3_moe.py

@@ -22,17 +22,19 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """Inference-only Qwen3MoE model compatible with HuggingFace weights."""
-from collections.abc import Iterable
+import typing
+from collections.abc import Callable, Iterable
 from typing import Any, Optional, Union
 
 import torch
-from torch import nn
+from torch import Tensor, nn
 from transformers import PretrainedConfig
 
 from vllm.attention import Attention
 from vllm.compilation.decorators import support_torch_compile
-from vllm.config import CacheConfig, VllmConfig
-from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
+from vllm.config import CacheConfig, VllmConfig, get_current_vllm_config
+from vllm.distributed import (get_ep_group, get_pp_group,
+                              get_tensor_model_parallel_world_size)
 from vllm.logger import init_logger
 from vllm.model_executor.layers.activation import SiluAndMul
 from vllm.model_executor.layers.fused_moe import FusedMoE
@@ -50,8 +52,8 @@
 from vllm.model_executor.sampling_metadata import SamplingMetadata
 from vllm.sequence import IntermediateTensors
 
-from .interfaces import SupportsLoRA, SupportsPP
-from .utils import (AutoWeightsLoader, extract_layer_index,
+from .interfaces import MixtureOfExperts, SupportsLoRA, SupportsPP
+from .utils import (AutoWeightsLoader, PPMissingLayer, extract_layer_index,
                     is_pp_missing_parameter,
                     make_empty_intermediate_tensors_factory, make_layers,
                     maybe_prefix)
@@ -101,6 +103,7 @@ def __init__(
         config: PretrainedConfig,
         quant_config: Optional[QuantizationConfig] = None,
         prefix: str = "",
+        enable_eplb: bool = False,
     ):
         super().__init__()
         self.tp_size = get_tensor_model_parallel_world_size()
@@ -110,14 +113,29 @@ def __init__(
                 f"Tensor parallel size {self.tp_size} is greater than "
                 f"the number of experts {config.num_experts}.")
 
+        self.ep_group = get_ep_group().device_group
+        self.ep_size = self.ep_group.size()
+
+        vllm_config = get_current_vllm_config()
+        parallel_config = vllm_config.parallel_config
+        self.n_routed_experts = config.num_experts
+        self.n_redundant_experts = parallel_config.num_redundant_experts
+        self.n_logical_experts = self.n_routed_experts
+        self.n_physical_experts = (self.n_logical_experts +
+                                   self.n_redundant_experts)
+        self.n_local_physical_experts = self.n_physical_experts // self.ep_size
+        self.enable_eplb = enable_eplb
+
         self.experts = FusedMoE(num_experts=config.num_experts,
                                 top_k=config.num_experts_per_tok,
                                 hidden_size=config.hidden_size,
                                 intermediate_size=config.moe_intermediate_size,
                                 reduce_results=False,
                                 renormalize=config.norm_topk_prob,
                                 quant_config=quant_config,
-                                prefix=f"{prefix}.experts")
+                                prefix=f"{prefix}.experts",
+                                enable_eplb=self.enable_eplb,
+                                num_redundant_experts=self.n_redundant_experts)
 
         self.gate = ReplicatedLinear(config.hidden_size,
                                      config.num_experts,
@@ -246,6 +264,7 @@ def __init__(
         cache_config: Optional[CacheConfig] = None,
         quant_config: Optional[QuantizationConfig] = None,
         prefix: str = "",
+        enable_eplb: bool = False,
     ) -> None:
         super().__init__()
         self.hidden_size = config.hidden_size
@@ -277,7 +296,8 @@ def __init__(
             (layer_idx + 1) % config.decoder_sparse_step == 0):
             self.mlp = Qwen3MoeSparseMoeBlock(config=config,
                                               quant_config=quant_config,
-                                              prefix=f"{prefix}.mlp")
+                                              prefix=f"{prefix}.mlp",
+                                              enable_eplb=enable_eplb)
         else:
             self.mlp = Qwen3MoeMLP(hidden_size=config.hidden_size,
                                    intermediate_size=config.intermediate_size,
@@ -323,6 +343,9 @@ def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
         config = vllm_config.model_config.hf_config
         cache_config = vllm_config.cache_config
         quant_config = vllm_config.quant_config
+        enable_eplb = vllm_config.parallel_config.enable_eplb
+        self.num_redundant_experts = (
+            vllm_config.parallel_config.num_redundant_experts)
 
         self.padding_idx = config.pad_token_id
         self.vocab_size = config.vocab_size
@@ -336,7 +359,8 @@ def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
             lambda prefix: Qwen3MoeDecoderLayer(config=config,
                                                 cache_config=cache_config,
                                                 quant_config=quant_config,
-                                                prefix=prefix),
+                                                prefix=prefix,
+                                                enable_eplb=enable_eplb),
             prefix=f"{prefix}.layers",
         )
         self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
@@ -382,7 +406,8 @@ def get_expert_mapping(self) -> list[tuple[str, str, int, str]]:
             ckpt_gate_proj_name="gate_proj",
             ckpt_down_proj_name="down_proj",
             ckpt_up_proj_name="up_proj",
-            num_experts=self.config.num_experts)
+            num_experts=self.config.num_experts,
+            num_redundant_experts=self.num_redundant_experts)
 
     def load_weights(self, weights: Iterable[tuple[str,
                                                    torch.Tensor]]) -> set[str]:
@@ -433,27 +458,38 @@ def load_weights(self, weights: Iterable[tuple[str,
                 weight_loader(param, loaded_weight, shard_id)
                 break
             else:
+                is_expert_weight = False
                 for mapping in expert_params_mapping:
                     param_name, weight_name, expert_id, shard_id = mapping
                     if weight_name not in name:
                         continue
-                    name = name.replace(weight_name, param_name)
+
+                    is_expert_weight = True
+
+                    name_mapped = name.replace(weight_name, param_name)
+
                     # Skip layers on other devices.
-                    if is_pp_missing_parameter(name, self):
+                    if is_pp_missing_parameter(name_mapped, self):
                         continue
                     # Skip loading extra parameters for GPTQ/modelopt models.
-                    if name.endswith(
+                    if name_mapped.endswith(
                             ignore_suffixes) and name not in params_dict:
                         continue
-                    param = params_dict[name]
-                    weight_loader = param.weight_loader
-                    weight_loader(param,
-                                  loaded_weight,
-                                  name,
-                                  shard_id=shard_id,
-                                  expert_id=expert_id)
-                    break
+                    param = params_dict[name_mapped]
+                    weight_loader = typing.cast(Callable[..., bool],
+                                                param.weight_loader)
+                    success = weight_loader(param,
+                                            loaded_weight,
+                                            name_mapped,
+                                            shard_id=shard_id,
+                                            expert_id=expert_id,
+                                            return_success=True)
+                    if success:
+                        name = name_mapped
+                        break
                 else:
+                    if is_expert_weight:
+                        continue
                     # Skip loading extra parameters for GPTQ/modelopt models.
                     if name.endswith(
                             ignore_suffixes) and name not in params_dict:
@@ -482,7 +518,8 @@ def load_weights(self, weights: Iterable[tuple[str,
         return loaded_params
 
 
-class Qwen3MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA):
+class Qwen3MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA,
+                          MixtureOfExperts):
     packed_modules_mapping = {
         "qkv_proj": [
             "q_proj",
@@ -514,6 +551,68 @@ def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
         self.make_empty_intermediate_tensors = (
             self.model.make_empty_intermediate_tensors)
 
+        # Implement the MixtureOfExperts protocol.
+        self.expert_weights = []
+
+        self.moe_layers: list[FusedMoE] = []
+        example_layer = None
+        for layer in self.model.layers:
+            if isinstance(layer, PPMissingLayer):
+                continue
+
+            assert isinstance(layer, Qwen3MoeDecoderLayer)
+            if isinstance(layer.mlp, Qwen3MoeSparseMoeBlock):
+                example_layer = layer.mlp
+                self.moe_layers.append(layer.mlp.experts)
+        self.num_moe_layers = len(self.moe_layers)
+
+        if example_layer is None:
+            raise RuntimeError("No Qwen3MoE layer found in model.layers.")
+
+        self.num_expert_groups = 1
+        self.num_logical_experts = example_layer.n_logical_experts
+        self.num_physical_experts = example_layer.n_physical_experts
+        self.num_local_physical_experts = example_layer.n_local_physical_experts
+        self.num_routed_experts = example_layer.n_routed_experts
+        self.num_shared_experts = 0
+        self.num_redundant_experts = example_layer.n_redundant_experts
+
+    def set_eplb_state(
+        self,
+        expert_load_view: Tensor,
+        logical_to_physical_map: Tensor,
+        logical_replica_count: Tensor,
+    ) -> None:
+        for layer_idx, layer in enumerate(self.moe_layers):
+            self.expert_weights.append(layer.get_expert_weights())
+            layer.set_eplb_state(
+                moe_layer_idx=layer_idx,
+                expert_load_view=expert_load_view,
+                logical_to_physical_map=logical_to_physical_map,
+                logical_replica_count=logical_replica_count,
+            )
+
+    def update_physical_experts_metadata(
+        self,
+        num_physical_experts: int,
+        num_local_physical_experts: int,
+    ) -> None:
+        if self.num_local_physical_experts != num_local_physical_experts:
+            raise ValueError(f"Expected num_local_physical_experts to be "
+                             f"{self.num_local_physical_experts}, got "
+                             f"{num_local_physical_experts}")
+        self.num_physical_experts = num_physical_experts
+        self.num_local_physical_experts = num_local_physical_experts
+        self.num_redundant_experts = (num_physical_experts -
+                                      self.num_logical_experts)
+        for layer in self.model.layers:
+            if isinstance(layer.mlp, Qwen3MoeSparseMoeBlock):
+                moe = layer.mlp
+                moe.n_local_physical_experts = num_local_physical_experts
+                moe.n_physical_experts = num_physical_experts
+                moe.n_redundant_experts = self.num_redundant_experts
+                moe.experts.update_expert_map()
+
     def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
         return self.model.get_input_embeddings(input_ids)