From 1d713b3e51de5ca94c8d54b53fc03d84117339a8 Mon Sep 17 00:00:00 2001
From: manyizhang <56545151+manyizhang@users.noreply.github.com>
Date: Wed, 26 Nov 2025 19:32:21 +0800
Subject: [PATCH] Eliminate duplicate bias addition in linear function

---
 inference_lib/src/fp_quant/module/pseudoquant_linear_fns.py | 2 --
 1 file changed, 2 deletions(-)

diff --git a/inference_lib/src/fp_quant/module/pseudoquant_linear_fns.py b/inference_lib/src/fp_quant/module/pseudoquant_linear_fns.py
index ea4e2d4..6e6a1b1 100644
--- a/inference_lib/src/fp_quant/module/pseudoquant_linear_fns.py
+++ b/inference_lib/src/fp_quant/module/pseudoquant_linear_fns.py
@@ -152,8 +152,6 @@ def forward(
         y = torch.nn.functional.linear(x_flat_dq, weight_dq, bias)
 
         y = y.unflatten(dim=0, sizes=x.shape[:-1])
-        if bias is not None:
-            y += bias
 
         ctx.x_shape = x.shape
         ctx.dtype = dtype