xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 07:15:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bugfix] Fix Qwen3-VL regression from #24982 (#25814)

Browse Source 
Signed-off-by: Roger Wang <hey@rogerw.io>
This commit is contained in:
Roger Wang 2025-09-27 20:21:09 -07:00 committed by GitHub
parent 69311446ba
commit 6144754014
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
vllm/model_executor/models/qwen3_moe.py
View File

@ -107,7 +107,7 @@ class Qwen3MoeSparseMoeBlock(nn.Module):
): ):
super().__init__() super().__init__()
config = vllm_config.model_config.hf_config config = vllm_config.model_config.hf_text_config
parallel_config = vllm_config.parallel_config parallel_config = vllm_config.parallel_config
quant_config = vllm_config.quant_config quant_config = vllm_config.quant_config
@ -293,7 +293,7 @@ class Qwen3MoeDecoderLayer(nn.Module):
def __init__(self, vllm_config: VllmConfig, prefix: str = "") -> None: def __init__(self, vllm_config: VllmConfig, prefix: str = "") -> None:
super().__init__() super().__init__()
config = vllm_config.model_config.hf_config config = vllm_config.model_config.hf_text_config
cache_config = vllm_config.cache_config cache_config = vllm_config.cache_config
quant_config = vllm_config.quant_config quant_config = vllm_config.quant_config
@ -372,7 +372,7 @@ class Qwen3MoeModel(nn.Module):
def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""): def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
super().__init__() super().__init__()
config = vllm_config.model_config.hf_config.get_text_config() config = vllm_config.model_config.hf_text_config
quant_config = vllm_config.quant_config quant_config = vllm_config.quant_config
parallel_config = vllm_config.parallel_config parallel_config = vllm_config.parallel_config
eplb_config = parallel_config.eplb_config eplb_config = parallel_config.eplb_config
@ -586,7 +586,7 @@ class Qwen3MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA,
def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""): def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
super().__init__() super().__init__()
config = vllm_config.model_config.hf_config config = vllm_config.model_config.hf_text_config
quant_config = vllm_config.quant_config quant_config = vllm_config.quant_config
self.config = config self.config = config
self.quant_config = quant_config self.quant_config = quant_config