diff --git a/vllm/model_executor/models/deepseek_v2.py b/vllm/model_executor/models/deepseek_v2.py index 6ed7505c9e..3b3b7a1a37 100644 --- a/vllm/model_executor/models/deepseek_v2.py +++ b/vllm/model_executor/models/deepseek_v2.py @@ -716,7 +716,7 @@ class DeepSeekV2FusedQkvAProj(MergedColumnParallelLinear): def __init__( self, input_size: int, - output_size: int, + output_size: list[int], quant_config: QuantizationConfig | None = None, prefix: str = "", ): @@ -726,7 +726,7 @@ class DeepSeekV2FusedQkvAProj(MergedColumnParallelLinear): bias=False, quant_config=quant_config, disable_tp=True, - prefix=f"{prefix}.kv_a_proj_with_mqa", + prefix=prefix, ) # Check if the DeepSeek V3 fused A GEMM kernel can be used.