diff --git a/vllm/model_executor/models/qwen3_moe.py b/vllm/model_executor/models/qwen3_moe.py index fe6b303ba0b5a..51cfa5796187d 100644 --- a/vllm/model_executor/models/qwen3_moe.py +++ b/vllm/model_executor/models/qwen3_moe.py @@ -475,6 +475,17 @@ class Qwen3MoeModel(nn.Module): class Qwen3MoeForCausalLM(nn.Module, SupportsPP): + packed_modules_mapping = { + "qkv_proj": [ + "q_proj", + "k_proj", + "v_proj", + ], + "gate_up_proj": [ + "gate_proj", + "up_proj", + ], + } fall_back_to_pt_during_load = False