From b2eb9cd18665317ec7900364ef21a38c3edb9e3f Mon Sep 17 00:00:00 2001
From: yuehuayingxueluo <867460659@qq.com>
Date: Thu, 4 Jan 2024 15:09:06 +0800
Subject: [PATCH] Fixed a typo

---
 colossalai/inference/modeling/models/llama.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/colossalai/inference/modeling/models/llama.py b/colossalai/inference/modeling/models/llama.py
index 10b2134a3..1331cc021 100644
--- a/colossalai/inference/modeling/models/llama.py
+++ b/colossalai/inference/modeling/models/llama.py
@@ -159,7 +159,7 @@ def llama_attn_forward(
 
     _, _, _, block_size = k_cache.shape
 
-    # NOTE: context_attention_unpadded is unsed for testing accuracy and we can only use aligned inputs.
+    # NOTE: context_attention_unpadded is used for testing accuracy and we can only use aligned inputs.
     # The code below will be uncommented after the development of attention-related kernel is completed.
     if is_prompts:
         attn_output = context_attention_unpadded(