From 91c40e6c58db0c7c59563c64dfa855830773b5c2 Mon Sep 17 00:00:00 2001
From: yuanwu <yuan.wu@intel.com>
Date: Wed, 11 Jun 2025 02:34:09 +0000
Subject: [PATCH] Fix crash

Signed-off-by: yuanwu <yuan.wu@intel.com>
---
 .../models/custom_modeling/flash_qwen3_moe_modeling.py           | 1 -
 1 file changed, 1 deletion(-)

diff --git a/backends/gaudi/server/text_generation_server/models/custom_modeling/flash_qwen3_moe_modeling.py b/backends/gaudi/server/text_generation_server/models/custom_modeling/flash_qwen3_moe_modeling.py
index f1e73f46..583c4ad0 100644
--- a/backends/gaudi/server/text_generation_server/models/custom_modeling/flash_qwen3_moe_modeling.py
+++ b/backends/gaudi/server/text_generation_server/models/custom_modeling/flash_qwen3_moe_modeling.py
@@ -185,7 +185,6 @@ class Qwen3MoeAttention(nn.Module):
                 seqlen=seqlen,
                 softmax_scale=self.scaling,
                 window_size_left=self.max_past,
-                num_key_value_groups=self.num_key_value_groups,
             )
         # Decode
         else: