[Misc][Model][Refactor] Pass the prefix into Linear layers (#31669)

Signed-off-by: Wang Kunpeng <1289706727@qq.com>
2026-01-06 04:03:18 +08:00
parent 02dbb933cb
commit 5708297e4e
17 changed files with 181 additions and 40 deletions
--- a/vllm/model_executor/models/phimoe.py
+++ b/vllm/model_executor/models/phimoe.py
@@ -272,6 +272,7 @@ class PhiMoE(nn.Module):
            bias=False,
            params_dtype=params_dtype,
            quant_config=None,
+            prefix=f"{prefix}.gate",
        )

        self.experts = FusedMoE(