From c3aaf3978050c7d3834577ed1cedc2ed202dc248 Mon Sep 17 00:00:00 2001
From: Md Fahim Faysal Khan <mdfahimfaysa@nvidia.com>
Date: Wed, 3 Apr 2024 15:26:00 -0700
Subject: [PATCH] changed the default mode quantize qkv config

---
 MaxText/layers/models.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/MaxText/layers/models.py b/MaxText/layers/models.py
index d15f76444..cfb87a157 100644
--- a/MaxText/layers/models.py
+++ b/MaxText/layers/models.py
@@ -90,7 +90,7 @@ def __call__(self,
       dropout_rate=cfg.dropout_rate,
       name='self_attention',
       quant=self.quant,
-      quantize_kvcache=self.quantize_kvcache)
+      quantize_kvcache=cfg.quantize_kvcache)
 
 
     attention_lnx = attention_layer(