diff --git a/vllm/lora/layers/column_parallel_linear.py b/vllm/lora/layers/column_parallel_linear.py index 12151699ac..ea776ec2ea 100644 --- a/vllm/lora/layers/column_parallel_linear.py +++ b/vllm/lora/layers/column_parallel_linear.py @@ -265,17 +265,36 @@ class MergedColumnParallelLinearWithLoRA(ColumnParallelLinearWithLoRA): def expand_packed_lora( self, - lora_a: list[torch.Tensor], - lora_b: list[torch.Tensor], - ) -> tuple[list[torch.Tensor], list[torch.Tensor]]: + lora_a: list[torch.Tensor | None], + lora_b: list[torch.Tensor | None], + ) -> tuple[list[torch.Tensor | None], list[torch.Tensor | None]]: """ Expand packed adapter groups when they don't match n_slices. - E.g. in_proj_qkv (covers Q+K+V) + in_proj_z + E.g. in_proj_qkv (covers Q+K+V) + in_proj_z. + + A None group member means that member was not adapted; the slice(s) + it covers are emitted as None placeholders so subsequent groups stay + aligned and those slices are left at base weights. This matches the + None-tolerance already present in slice_lora_b() and the set_lora() + stacking loop. """ - expanded_a: list[torch.Tensor] = [] - expanded_b: list[torch.Tensor] = [] + expanded_a: list[torch.Tensor | None] = [] + expanded_b: list[torch.Tensor | None] = [] start_idx = 0 for a_i, b_i in zip(lora_a, lora_b): + if b_i is None: + # Unadapted group member: its row count is unknown (the tensor + # is missing), so infer its coverage as the remaining slices. + # This is exact for the only layout that reaches this path, + # the fused GDN in_proj_qkvz group, whose sole multi-slice + # member (in_proj_qkv, Q+K+V) leads and whose only optional + # member (in_proj_z) trails. + covered = self.n_slices - start_idx + for _ in range(covered): + expanded_a.append(None) + expanded_b.append(None) + start_idx += covered + continue # Determine which output slices this b_i covers. b_rows, cu_rows, covered = b_i.shape[0], 0, 0 for i in range(start_idx, self.n_slices):