Fix some doc in models (#1951)

Fix doc for human pose estimation and DeepLab v3+.

fluid/PaddleCV/deeplabv3+/README.md

@@ -90,7 +90,7 @@ python ./train.py \
 执行以下命令在`Cityscape`测试数据集上进行测试：
 ```
 python ./eval.py \
-    --init_weights=deeplabv3plus_gn \
+    --init_weights_path=deeplabv3plus_gn \
     --norm_type=gn \
     --dataset_path=$DATASET_PATH
 ```

fluid/PaddleCV/deeplabv3+/eval.py

@@ -23,7 +23,6 @@ add_arg = lambda *args: utility.add_arguments(*args, argparser=parser)
 add_arg('total_step',           int,    -1,     "Number of the step to be evaluated, -1 for full evaluation.")
 add_arg('init_weights_path',    str,    None,   "Path of the weights to evaluate.")
 add_arg('dataset_path',         str,    None,   "Cityscape dataset path.")
-add_arg('verbose',              bool,   False,  "Print mIoU for each step if verbose.")
 add_arg('use_gpu',              bool,   True,   "Whether use GPU or CPU.")
 add_arg('num_classes',          int,    19,     "Number of classes.")
 add_arg('use_py_reader',        bool,   True,   "Use py_reader.")

fluid/PaddleCV/human_pose_estimation/README.md

@@ -31,7 +31,7 @@ The code is developed and tested under 4 Tesla K40/P40 GPUS cards on CentOS with
 | 256x192\_pose\_resnet\_50 in PyTorch | 0.704 | 0.886 | 0.783 | 0.671 | 0.772 | 0.763 | 0.929 | 0.834 | 0.721 | 0.824 | - |
 | 256x192\_pose\_resnet\_50 in Fluid   | 0.712 | 0.897 | 0.786 | 0.683 | 0.756 | 0.741 | 0.906 | 0.806 | 0.709 | 0.790 | [`link`](https://paddlemodels.bj.bcebos.com/pose/pose-resnet50-coco-256x192.tar.gz) |
 | 384x288\_pose\_resnet\_50 in PyTorch | 0.722 | 0.893 | 0.789 | 0.681 | 0.797 | 0.776 | 0.932 | 0.838 | 0.728 | 0.846 | - |
-| 384x288\_pose\_resnet\_50 in Fluid   | 0.727 | 0.897 | 0.796 | 0.690 | 0.783 | 0.754 | 0.907 | 0.813 | 0.714 | 0.814 | [`link`](https://paddlemodels.bj.bcebos.com/pose/pose-resnet50-coco-384x288.tar.gz |
+| 384x288\_pose\_resnet\_50 in Fluid   | 0.727 | 0.897 | 0.796 | 0.690 | 0.783 | 0.754 | 0.907 | 0.813 | 0.714 | 0.814 | [`link`](https://paddlemodels.bj.bcebos.com/pose/pose-resnet50-coco-384x288.tar.gz) |
 
 ### Notes:
 

fluid/PaddleCV/human_pose_estimation/README_cn.md

@@ -29,7 +29,7 @@
 | 256x192\_pose\_resnet\_50 in PyTorch | 0.704 | 0.886 | 0.783 | 0.671 | 0.772 | 0.763 | 0.929 | 0.834 | 0.721 | 0.824 | - |
 | 256x192\_pose\_resnet\_50 in Fluid   | 0.712 | 0.897 | 0.786 | 0.683 | 0.756 | 0.741 | 0.906 | 0.806 | 0.709 | 0.790 | [`link`](https://paddlemodels.bj.bcebos.com/pose/pose-resnet50-coco-256x192.tar.gz) |
 | 384x288\_pose\_resnet\_50 in PyTorch | 0.722 | 0.893 | 0.789 | 0.681 | 0.797 | 0.776 | 0.932 | 0.838 | 0.728 | 0.846 | - |
-| 384x288\_pose\_resnet\_50 in Fluid   | 0.727 | 0.897 | 0.796 | 0.690 | 0.783 | 0.754 | 0.907 | 0.813 | 0.714 | 0.814 | [`link`](https://paddlemodels.bj.bcebos.com/pose/pose-resnet50-coco-384x288.tar.gz |
+| 384x288\_pose\_resnet\_50 in Fluid   | 0.727 | 0.897 | 0.796 | 0.690 | 0.783 | 0.754 | 0.907 | 0.813 | 0.714 | 0.814 | [`link`](https://paddlemodels.bj.bcebos.com/pose/pose-resnet50-coco-384x288.tar.gz) |
 
 ### 说明