From 1b0b565d68a93bad8a4944ab5b1c8500d6887f1d Mon Sep 17 00:00:00 2001
From: zxq <342239412@qq.com>
Date: Mon, 22 Dec 2025 17:18:36 +0800
Subject: [PATCH] =?UTF-8?q?=E3=80=90master=E3=80=91=E9=80=82=E9=85=8Dq=5Fl?=
 =?UTF-8?q?ora=5Frank=E4=B8=BAnull=E5=9C=BA=E6=99=AF=E7=9A=84=E6=9D=83?=
 =?UTF-8?q?=E9=87=8D=E5=8A=A0=E8=BD=BD?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 .../inference/transformer/multi_latent_attention.py              | 1 -
 1 file changed, 1 deletion(-)

diff --git a/mindformers/parallel_core/inference/transformer/multi_latent_attention.py b/mindformers/parallel_core/inference/transformer/multi_latent_attention.py
index bb704a603..07f95064f 100644
--- a/mindformers/parallel_core/inference/transformer/multi_latent_attention.py
+++ b/mindformers/parallel_core/inference/transformer/multi_latent_attention.py
@@ -292,7 +292,6 @@ class MLASelfAttention(MultiLatentAttention):
                 config=self.config,
                 bias=False,
                 skip_bias_add=False,
-                gather_output=False,
                 transpose_b=True,
                 compute_dtype=self.config.compute_dtype,
                 is_expert=False,
-- 
Gitee