Refine readme in cycle gan. (#1972)

* 1. Refine readme
2. Fix some issues

* Refine README.md.

PaddleCV/gan/cycle_gan/README.md

-
 
 运行本目录下的程序示例需要使用PaddlePaddle develop最新版本。如果您的PaddlePaddle安装版本低于此要求，请按照[安装文档](http://www.paddlepaddle.org/docs/develop/documentation/zh/build_and_install/pip_install_cn.html)中的说明更新PaddlePaddle安装版本。
 
@@ -73,8 +72,8 @@ env CUDA_VISIBLE_DEVICES=0 python train.py
 执行以下命令读取多张图片进行预测：
 
 ```
-env CUDA_VISIBLE_DEVICE=0 python infer.py \
-    --init_model="checkpoints/1" --input="./data/inputA/*" \
+env CUDA_VISIBLE_DEVICES=0 python infer.py \
+    --init_model="output/checkpoints/1" --input="./data/horse2zebra/trainA/*" \
     --input_style A --output="./output"
 ```
 
@@ -89,3 +88,5 @@ env CUDA_VISIBLE_DEVICE=0 python infer.py \
 <img src="images/B2A.jpg" width="620" hspace='10'/> <br/>
 <strong>图 3</strong>
 </p>
+
+>在本文示例中，均可通过修改`CUDA_VISIBLE_DEVICES`改变使用的显卡号。

PaddleCV/gan/cycle_gan/infer.py

@@ -44,7 +44,6 @@ def infer(args):
     if not os.path.exists(args.output):
         os.makedirs(args.output)
     for file in glob.glob(args.input):
-        print "read %s" % file
         image_name = os.path.basename(file)
         image = Image.open(file)
         image = image.resize((256, 256))

PaddleCV/gan/cycle_gan/train.py

@@ -22,7 +22,7 @@ add_arg = functools.partial(add_arguments, argparser=parser)
 # yapf: disable
 add_arg('batch_size',        int,   1,          "Minibatch size.")
 add_arg('epoch',             int,   2,        "The number of epoched to be trained.")
-add_arg('output',            str,   "./output_0", "The directory the model and the test result to be saved to.")
+add_arg('output',            str,   "./output", "The directory the model and the test result to be saved to.")
 add_arg('init_model',        str,   None,       "The init model file of directory.")
 add_arg('save_checkpoints',  bool,  True,       "Whether to save checkpoints.")
 add_arg('run_test',          bool,  True,       "Whether to run test.")