[Deploy] update deploy serving (#2203)

* fix crash bug when using cpp infer, test=dygraph

* update PaddleServing deploy, test=dygraph

* Update export_model.py

* Update export_model.py
Co-authored-by: Nqingqing01 <dangqingqing@baidu.com>

dygraph/deploy/README.md

@@ -16,10 +16,15 @@
 使用`tools/export_model.py`脚本导出模型已经部署时使用的配置文件，配置文件名字为`infer_cfg.yml`。模型导出脚本如下：
 ```bash
 # 导出YOLOv3模型
-python tools/export_model.py -c configs/yolov3/yolov3_darknet53_270e_coco.yml --output_dir=./inference_model \
- -o weights=weights/yolov3_darknet53_270e_coco.pdparams
+python tools/export_model.py -c configs/yolov3/yolov3_darknet53_270e_coco.yml -o weights=weights/yolov3_darknet53_270e_coco.pdparams
 ```
-预测模型会导出到`inference_model/yolov3_darknet53_270e_coco`目录下，分别为`infer_cfg.yml`, `model.pdiparams`,  `model.pdiparams.info`, `model.pdmodel`。
+预测模型会导出到`output_inference/yolov3_darknet53_270e_coco`目录下，分别为`infer_cfg.yml`, `model.pdiparams`,  `model.pdiparams.info`, `model.pdmodel`。
+
+如果需要导出`PaddleServing`格式的模型，需要设置`export_serving_model=True`:
+```buildoutcfg
+python tools/export_model.py -c configs/yolov3/yolov3_darknet53_270e_coco.yml -o weights=weights/yolov3_darknet53_270e_coco.pdparams --export_serving_model=True
+```
+预测模型会导出到`output_inference/yolov3_darknet53_270e_coco`目录下，分别为`infer_cfg.yml`, `model.pdiparams`,  `model.pdiparams.info`, `model.pdmodel`, `serving_client/`文件夹, `serving_server/`文件夹。
 
 模型导出具体请参考文档[PaddleDetection模型导出教程](EXPORT_MODEL.md)。
 
@@ -34,10 +39,12 @@ python tools/export_model.py -c configs/yolov3/yolov3_darknet53_270e_coco.yml --
   （1）Paddle官网提供在不同平台、不同环境下编译好的预测库，您可以直接使用，请在这里[Paddle预测库](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/05_inference_deployment/inference/build_and_install_lib_cn.html) 选择。
   （2）如果您将要部署的平台环境，Paddle官网上没有提供已编译好的预测库，您可以自行编译，编译过程请参考[Paddle源码编译](https://www.paddlepaddle.org.cn/documentation/docs/zh/install/compile/linux-compile.html)。
 
+**注意:**  Paddle预测库版本需要>=2.0
+
 - Python语言部署，需要在对应平台上安装Paddle Python包。如果Paddle官网上没有提供该平台下的Paddle Python包，您可以自行编译，编译过程请参考[Paddle源码编译](https://www.paddlepaddle.org.cn/documentation/docs/zh/install/compile/linux-compile.html)。
 
 - PaddleServing部署
-  PaddleServing 0.4.0是基于Paddle 1.8.4开发，PaddleServing 0.4.1是基于Paddle2.0开发。
+  PaddleServing 0.4.0是基于Paddle 1.8.4开发，PaddleServing 0.5.0是基于Paddle2.0开发。
 
 - Paddle-Lite部署
   Paddle-Lite支持OP列表请参考：[Paddle-Lite支持的OP列表](https://paddle-lite.readthedocs.io/zh/latest/source_compile/library.html) ，请跟进所部署模型中使用到的op选择Paddle-Lite版本。

dygraph/deploy/cpp/include/preprocess_op.h

@@ -22,6 +22,7 @@
 #include <utility>
 #include <memory>
 #include <unordered_map>
+#include <iostream>
 
 #include <opencv2/core/core.hpp>
 #include <opencv2/imgproc/imgproc.hpp>
@@ -138,9 +139,10 @@ class Preprocessor {
       return std::make_shared<Permute>();
     } else if (name == "NormalizeImageOp") {
       return std::make_shared<Normalize>();
-    } else if (name == "PadBatchOp") {
+    } else if (name == "PadBatchOp" || name == "PadStride") {
       return std::make_shared<PadStride>();
     }
+    std::cerr << "can not find function of OP: " << name << " and return: nullptr" << std::endl;
     return nullptr;
   }
 

dygraph/deploy/serving/README.md

 # 服务端预测部署
 
 `PaddleDetection`训练出来的模型可以使用[Serving](https://github.com/PaddlePaddle/Serving) 部署在服务端。  
-本教程以在路标数据集[roadsign_voc](https://paddlemodels.bj.bcebos.com/object_detection/roadsign_voc.tar) 使用`configs/yolov3_mobilenet_v1_roadsign.yml`算法训练的模型进行部署。  
-预训练模型权重文件为[yolov3_mobilenet_v1_roadsign.pdparams](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v1_roadsign.pdparams) 。
+本教程以在COCO数据集上用`configs/yolov3/yolov3_darknet53_270e_coco.yml`算法训练的模型进行部署。  
+预训练模型权重文件为[yolov3_darknet53_270e_coco.pdparams](https://paddlemodels.bj.bcebos.com/object_detection/dygraph/yolov3_darknet53_270e_coco.pdparams) 。
 
 ## 1. 首先验证模型
 ```
-python tools/infer.py -c configs/yolov3_mobilenet_v1_roadsign.yml -o use_gpu=true weights=https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v1_roadsign.pdparams --infer_img=demo/road554.png
+python tools/infer.py -c  --infer_img=demo/000000014439.jpg -o use_gpu=True weights=https://paddlemodels.bj.bcebos.com/object_detection/dygraph/yolov3_darknet53_270e_coco.pdparams --infer_img=demo/000000014439.jpg
 ```
 
 ## 2. 安装 paddle serving
-```
-# 安装 paddle-serving-client
-pip install paddle-serving-client -i https://mirror.baidu.com/pypi/simple
-
-# 安装 paddle-serving-server
-pip install paddle-serving-server -i https://mirror.baidu.com/pypi/simple
-
-# 安装 paddle-serving-server-gpu
-pip install paddle-serving-server-gpu -i https://mirror.baidu.com/pypi/simple
-```
+请参考[PaddleServing](https://github.com/PaddlePaddle/Serving/tree/v0.5.0) 中安装教程安装
 
 ## 3. 导出模型
 PaddleDetection在训练过程包括网络的前向和优化器相关参数，而在部署过程中，我们只需要前向参数，具体参考:[导出模型](https://github.com/PaddlePaddle/PaddleDetection/blob/master/docs/advanced_tutorials/deploy/EXPORT_MODEL.md)
 
 ```
-python tools/export_serving_model.py -c configs/yolov3_mobilenet_v1_roadsign.yml -o use_gpu=true weights=https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v1_roadsign.pdparams --output_dir=./inference_model
+python tools/export_model.py -c configs/yolov3/yolov3_darknet53_270e_coco.yml -o weights=weights/yolov3_darknet53_270e_coco.pdparams --export_serving_model=True
 ```
 
-以上命令会在./inference_model文件夹下生成一个`yolov3_mobilenet_v1_roadsign`文件夹：
+以上命令会在`output_inference/`文件夹下生成一个`yolov3_darknet53_270e_coco`文件夹：
 ```
-inference_model
-│   ├── yolov3_mobilenet_v1_roadsign
+output_inference
+│   ├── yolov3_darknet53_270e_coco
 │   │   ├── infer_cfg.yml
+│   │   ├── model.pdiparams
+│   │   ├── model.pdiparams.info
+│   │   ├── model.pdmodel
 │   │   ├── serving_client
 │   │   │   ├── serving_client_conf.prototxt
 │   │   │   ├── serving_client_conf.stream.prototxt
 │   │   ├── serving_server
-│   │   │   ├── conv1_bn_mean
-│   │   │   ├── conv1_bn_offset
-│   │   │   ├── conv1_bn_scale
+│   │   │   ├── __model__
+│   │   │   ├── __params__
+│   │   │   ├── serving_server_conf.prototxt
+│   │   │   ├── serving_server_conf.stream.prototxt
 │   │   │   ├── ...
 ```
 
 `serving_client`文件夹下`serving_client_conf.prototxt`详细说明了模型输入输出信息
 `serving_client_conf.prototxt`文件内容为：
 ```
+lient_conf.prototxt
+feed_var {
+  name: "im_shape"
+  alias_name: "im_shape"
+  is_lod_tensor: false
+  feed_type: 1
+  shape: 2
+}
 feed_var {
   name: "image"
   alias_name: "image"
@@ -56,25 +59,32 @@ feed_var {
   shape: 608
 }
 feed_var {
-  name: "im_size"
-  alias_name: "im_size"
+  name: "scale_factor"
+  alias_name: "scale_factor"
   is_lod_tensor: false
-  feed_type: 2
+  feed_type: 1
   shape: 2
 }
 fetch_var {
-  name: "multiclass_nms_0.tmp_0"
-  alias_name: "multiclass_nms_0.tmp_0"
+  name: "save_infer_model/scale_0.tmp_1"
+  alias_name: "save_infer_model/scale_0.tmp_1"
   is_lod_tensor: true
   fetch_type: 1
   shape: -1
 }
+fetch_var {
+  name: "save_infer_model/scale_1.tmp_1"
+  alias_name: "save_infer_model/scale_1.tmp_1"
+  is_lod_tensor: true
+  fetch_type: 2
+  shape: -1
+}
 ```
 
 ## 4. 启动PaddleServing服务
 
 ```
-cd inference_model/yolov3_mobilenet_v1_roadsign/
+cd output_inference/yolov3_darknet53_270e_coco/
 
 # GPU
 python -m paddle_serving_server_gpu.serve --model serving_server --port 9393 --gpu_ids 0
@@ -87,20 +97,19 @@ python -m paddle_serving_server.serve --model serving_server --port 9393
 准备`label_list.txt`文件
 ```
 # 进入到导出模型文件夹
-cd inference_model/yolov3_mobilenet_v1_roadsign/
+cd output_inference/yolov3_darknet53_270e_coco/
 
-# 将数据集对应的label_list.txt文件拷贝到当前文件夹下
-cp ../../dataset/roadsign_voc/label_list.txt .
+# 将数据集对应的label_list.txt文件放到当前文件夹下
 ```
 
 设置`prototxt`文件路径为`serving_client/serving_client_conf.prototxt` 。  
-设置`fetch`为`fetch=["multiclass_nms_0.tmp_0"])`
+设置`fetch`为`fetch=["save_infer_model/scale_0.tmp_1"])`
 
 测试
 ```
 # 进入目录
-cd inference_model/yolov3_mobilenet_v1_roadsign/
+cd output_inference/yolov3_darknet53_270e_coco/
 
-# 测试代码 test_client.py 会自动创建output文件夹，并在output下生成`bbox.json`和`road554.png`两个文件
-python ../../deploy/serving/test_client.py ../../demo/road554.png
+# 测试代码 test_client.py 会自动创建output文件夹，并在output下生成`bbox.json`和`000000014439.jpg`两个文件
+python ../../deploy/serving/test_client.py ../../demo/000000014439.jpg
 ```

dygraph/tools/export_model.py

@@ -43,6 +43,11 @@ def parse_args():
         type=str,
         default="output_inference",
         help="Directory for storing the output model files.")
+    parser.add_argument(
+        "--export_serving_model",
+        type=bool,
+        default=False,
+        help="Whether to export serving model or not.")
     parser.add_argument(
         "--slim_config",
         default=None,
@@ -62,6 +67,19 @@ def run(FLAGS, cfg):
     # export model
     trainer.export(FLAGS.output_dir)
 
+    if FLAGS.export_serving_model:
+        from paddle_serving_client.io import inference_model_to_serving
+        model_name = os.path.splitext(os.path.split(cfg.filename)[-1])[0]
+
+        inference_model_to_serving(
+            dirname="{}/{}".format(FLAGS.output_dir, model_name),
+            serving_server="{}/{}/serving_server".format(FLAGS.output_dir,
+                                                         model_name),
+            serving_client="{}/{}/serving_client".format(FLAGS.output_dir,
+                                                         model_name),
+            model_filename="model.pdmodel",
+            params_filename="model.pdiparams")
+
 
 def main():
     paddle.set_device("cpu")