modified readme

PaddleCV/yolov3/README.md

@@ -163,6 +163,8 @@ Inference is used to get prediction score or image features based on trained mod
         --image_name=000000000139.jpg \
         --draw_threshold=0.5
 
+- Set ```export CUDA_VISIBLE_DEVICES=0``` to specifiy one GPU to infer.
+
 Inference speed（Tesla P40）:
 
 

PaddleCV/yolov3/README_cn.md

@@ -165,6 +165,8 @@ Train Loss
         --image_name=000000000139.jpg \
         --draw_threshold=0.5
 
+- 通过设置export CUDA\_VISIBLE\_DEVICES=0指定单卡GPU预测。
+
 模型预测速度（Tesla P40）：