[Mypy Fix] Mypy fix for "vllm/model_executor/models/[eE][fF]" (#53381)

Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
Wentao Ye
2026-08-22 15:41:27 -04:00
committed by GitHub
parent 236f78cc5c
commit 811d12ee26
16 changed files with 71 additions and 84 deletions
+3 -20
View File
@@ -41,7 +41,6 @@ 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.vocab_parallel_embedding import (
DEFAULT_VOCAB_PADDING_SIZE,
ParallelLMHead,
VocabParallelEmbedding,
)
@@ -112,6 +111,7 @@ class ExaoneMoe(nn.Module):
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.shared_experts: ExaoneMoeGatedMLP | None
if getattr(config, "num_shared_experts", 0) > 0:
intermediate_size = config.moe_intermediate_size * config.num_shared_experts
self.shared_experts = ExaoneMoeGatedMLP(
@@ -292,7 +292,7 @@ class ExaoneMoeDecoderLayer(nn.Module):
config: PretrainedConfig,
cache_config: CacheConfig | None = None,
quant_config: QuantizationConfig | None = None,
is_mtp: bool = None,
is_mtp: bool = False,
prefix: str = "",
) -> None:
super().__init__()
@@ -381,26 +381,19 @@ class ExaoneMoeModel(nn.Module):
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.num_redundant_experts = (
vllm_config.parallel_config.eplb_config.num_redundant_experts
)
self.config = config
self.quant_config = quant_config
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.vocab_size = config.vocab_size
if get_pp_group().is_first_rank or (
config.tie_word_embeddings 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:
@@ -497,10 +490,8 @@ class ExaoneMoeForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
super().__init__()
config = vllm_config.model_config.hf_config.get_text_config()
quant_config = vllm_config.quant_config
lora_config = vllm_config.lora_config
self.config = config
self.lora_config = lora_config
self.quant_config = quant_config
self.model = ExaoneMoeModel(
@@ -509,17 +500,9 @@ class ExaoneMoeForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
)
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
# We need bigger padding if using lora for kernel
# compatibility
if not lora_config
else lora_config.lora_vocab_padding_size,
quant_config=quant_config,
)
if config.tie_word_embeddings: