biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Elastic Expert Parallel Initial Support (#20775)

Browse Source 
Signed-off-by: Rui Qiao <ruisearch42@gmail.com>
This commit is contained in:
Rui Qiao
2025-07-18 17:46:09 -07:00
committed by GitHub
parent 5782581acf
commit 217937221b
24 changed files with 1659 additions and 68 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
vllm/model_executor/models/deepseek_v2.py
View File

@@ -776,6 +776,24 @@ class DeepseekV2ForCausalLM(nn.Module, SupportsPP, MixtureOfExperts):
logical_replica_count=logical_replica_count,
)
def update_physical_experts_metadata(
self,
num_physical_experts: int,
num_local_physical_experts: int,
) -> None:
assert self.num_local_physical_experts == 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, DeepseekV2MoE):
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)
@@ -931,9 +949,8 @@ class DeepseekV3ForCausalLM(DeepseekV2ForCausalLM):
def get_spec_layer_idx_from_weight_name(config: PretrainedConfig,
weight_name: str) -> Optional[int]:
if hasattr(config,
"num_nextn_predict_layers") and (config.num_nextn_predict_layers
> 0):
if (hasattr(config, "num_nextn_predict_layers")
and config.num_nextn_predict_layers > 0):
layer_idx = config.num_hidden_layers
for i in range(config.num_nextn_predict_layers):
if weight_name.startswith(f"model.layers.{layer_idx+i}."):