From 38ea8c3931ce6e06fa2bccce41bff78d16d9af69 Mon Sep 17 00:00:00 2001
From: Baiju Meswani <bmeswani@microsoft.com>
Date: Mon, 28 Aug 2023 17:05:40 -0700
Subject: [PATCH] Increase max error tolerance for ConvTransposeGrad test
 (#17315)

---
 orttraining/orttraining/test/gradient/gradient_ops_test.cc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/orttraining/orttraining/test/gradient/gradient_ops_test.cc b/orttraining/orttraining/test/gradient/gradient_ops_test.cc
index d4e18dbfd2290..178d5db627888 100644
--- a/orttraining/orttraining/test/gradient/gradient_ops_test.cc
+++ b/orttraining/orttraining/test/gradient/gradient_ops_test.cc
@@ -3045,7 +3045,7 @@ void ConvTransposeGradientCheckerTest(std::vector<std::unique_ptr<IExecutionProv
   GradientChecker<float, float, float> gradient_checker;
   OpDef op_def{"ConvTranspose"};
 
-  float error_tolerance = 1e-1f;
+  float error_tolerance = 3e-1f;
 
   // 1D convolution
   {