diff --git a/PaddleCV/face_detection/README_cn.md b/PaddleCV/face_detection/README_cn.md
index 3329a914af4fd12ae943fdc09ba7a6a15995db77..7324ad98760170550db69bc57850c18b90cfafbb 100644
--- a/PaddleCV/face_detection/README_cn.md
+++ b/PaddleCV/face_detection/README_cn.md
@@ -92,8 +92,8 @@ tar -xf vgg_ilsvrc_16_fc_reduced.tar.gz && rm -f vgg_ilsvrc_16_fc_reduced.tar.gz
 ```bash
 python -u train.py --batch_size=16 --pretrained_model=vgg_ilsvrc_16_fc_reduced
 ```
-  - 可以通过设置 `export CUDA_VISIBLE_DEVICES=0,1,2,3` 指定想要使用的GPU数量，`batch_size`默认设置为12或16。
-  - **注意**: 在**Windows**机器上训练，需要设置 `--use_multiprocess=False`，因为在Windows上使用Python多进程加速训练时有错误。
+  - 可以通过设置 `export CUDA_VISIBLE_DEVICES=0,1,2,3` 指定想要使用的GPU数量，`batch_size`默认设置为12或16，如果GPU显存较小，请适当调低`batch_size`。
+  - **注意**: 在**Windows**机器上训练或者训练环境内存较小时，需要设置 `--use_multiprocess=False`，因为在Windows上使用Python多进程加速训练时有错误，内存不足时程序会崩溃。
   - 更多的可选参数见:
     ```bash
     python train.py --help
diff --git a/PaddleCV/face_detection/visualize.py b/PaddleCV/face_detection/visualize.py
index eee5d4db9ed7848cfeee1e92ec00ed489efd93dc..469696c6fc4f3670a3672635c1f1216f6d6b46e7 100644
--- a/PaddleCV/face_detection/visualize.py
+++ b/PaddleCV/face_detection/visualize.py
@@ -28,7 +28,7 @@ def draw_bbox(image, bbox):
         bbox (np.array|list|tuple): (xmin, ymin, xmax, ymax).
     """
     draw = ImageDraw.Draw(image)
-    xmin, ymin, xmax, ymax = box
+    xmin, ymin, xmax, ymax = bbox
     (left, right, top, bottom) = (xmin, xmax, ymin, ymax)
     draw.line(
         [(left, top), (left, bottom), (right, bottom), (right, top),
diff --git a/PaddleCV/face_detection/widerface_eval.py b/PaddleCV/face_detection/widerface_eval.py
index 80201486b70d49689a9063b518efd6d4223edefd..803655ddc2598b595ee45d3ed74a0a8861a03c51 100644
--- a/PaddleCV/face_detection/widerface_eval.py
+++ b/PaddleCV/face_detection/widerface_eval.py
@@ -67,7 +67,7 @@ def infer(args, config):
         image_path = args.image_path
         image = Image.open(image_path)
         if image.mode == 'L':
-            image = img.convert('RGB')
+            image = image.convert('RGB')
         shrink, max_shrink = get_shrink(image.size[1], image.size[0])
 
         det0 = detect_face(image, shrink)