diff --git a/README.md b/README.md
index bd91b0ee43fbbbb8794cdeb4bfeb90b890681ba9..77460da347ea8596a388e41f273fef6f80e7e420 100644
--- a/README.md
+++ b/README.md
@@ -101,7 +101,7 @@ All data in **LongBench** are standardized to the following format:
 ```
 
 #### Evaluation
-We provide an evaluation code using ChatGLM2-6B as an example. First, run the [pred.py](pred.py) under the repository:
+Install the requirements with pip: `pip install -r requirements.txt`. We provide an evaluation code using ChatGLM2-6B as an example. First, run the [pred.py](pred.py) under the repository:
 ```bash
 CUDA_VISIBLE_DEVICES=0 python pred.py
 ```
diff --git a/README_ZH.md b/README_ZH.md
index a4474c2c3c19e90350bd7845db84192a499bfa00..6bb46f258800d9b23102728456cdba41c40fb499 100644
--- a/README_ZH.md
+++ b/README_ZH.md
@@ -97,7 +97,7 @@ for dataset in datasets:
 ```
 
 #### 评测
-我们以ChatGLM2-6B为例提供了一份评测代码。首先，运行仓库下的[pred.py](pred.py)
+通过pip安装依赖：`pip install -r requirements.txt`。我们以ChatGLM2-6B为例提供了一份评测代码。首先，运行仓库下的[pred.py](pred.py)
 ```bash
 CUDA_VISIBLE_DEVICES=0 python pred.py
 ```