polish yolo readme (#2201)

PaddleCV/yolov3/README.md

@@ -161,7 +161,7 @@ Inference is used to get prediction score or image features based on trained mod
         --weights=${path_to_weights}  \
         --image_path=data/COCO17/val2017/  \
         --image_name=000000000139.jpg \
-        --draw_threshold=0.5
+        --draw_thresh=0.5
 
 - Set ```export CUDA_VISIBLE_DEVICES=0``` to specifiy one GPU to infer.
 

PaddleCV/yolov3/README_cn.md

@@ -163,7 +163,7 @@ Train Loss
         --weights=${path_to_weights}  \
         --image_path=data/COCO17/val2017/  \
         --image_name=000000000139.jpg \
-        --draw_threshold=0.5
+        --draw_thresh=0.5
 
 - 通过设置export CUDA\_VISIBLE\_DEVICES=0指定单卡GPU预测。