From d2876b84f13cf3797c57810eb282614118bf2523 Mon Sep 17 00:00:00 2001
From: Guenther Schmuelling <guschmue@microsoft.com>
Date: Fri, 7 Feb 2025 13:12:03 -0800
Subject: [PATCH] works

---
 onnxruntime/contrib_ops/webgpu/quantization/matmul_nbits.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/onnxruntime/contrib_ops/webgpu/quantization/matmul_nbits.h b/onnxruntime/contrib_ops/webgpu/quantization/matmul_nbits.h
index a2470d9268..1a32ef4280 100644
--- a/onnxruntime/contrib_ops/webgpu/quantization/matmul_nbits.h
+++ b/onnxruntime/contrib_ops/webgpu/quantization/matmul_nbits.h
@@ -60,7 +60,7 @@ class MatMulNBits final : public WebGpuKernel {
     N_ = info.GetAttr<int64_t>("N");
     block_size_ = info.GetAttr<int64_t>("block_size");
     int64_t bits = info.GetAttr<int64_t>("bits");
-    accuracy_level_ = info.GetAttrOrDefault<int64_t>("accuracy_level", 4);
+    accuracy_level_ = 4; // info.GetAttrOrDefault<int64_t>("accuracy_level", 4);
     ORT_ENFORCE(bits == 4,
                 "Only 4b quantization is supported for MatMulNBits op, additional bits support is planned.");
   }