Merge pull request #1030 from wangjiawei04/pddet2

add paddle detection and fix dependency

core/general-client/CMakeLists.txt

 if(CLIENT)
 add_subdirectory(pybind11)
 pybind11_add_module(serving_client src/general_model.cpp src/pybind_general_model.cpp)
+add_dependencies(serving_client sdk_cpp)
 target_link_libraries(serving_client PRIVATE -Wl,--whole-archive utils sdk-cpp pybind python -Wl,--no-whole-archive -lpthread -lcrypto -lm -lrt -lssl -ldl -lz -Wl,-rpath,'$ORIGIN'/lib)
 endif()

core/pdcodegen/CMakeLists.txt

@@ -7,7 +7,7 @@ PROTOBUF_GENERATE_CPP(pdcodegen_proto_srcs pdcodegen_proto_hdrs
 LIST(APPEND pdcodegen_srcs ${pdcodegen_proto_srcs})
 
 add_executable(pdcodegen ${pdcodegen_srcs})
-add_dependencies(pdcodegen boost)
+add_dependencies(pdcodegen boost protobuf)
 target_link_libraries(pdcodegen protobuf ${PROTOBUF_PROTOC_LIBRARY})
 
 # install

core/sdk-cpp/CMakeLists.txt

@@ -2,7 +2,7 @@
 include(src/CMakeLists.txt)
 include(proto/CMakeLists.txt)
 add_library(sdk-cpp ${sdk_cpp_srcs})
-add_dependencies(sdk-cpp pdcodegen configure)
+add_dependencies(sdk-cpp pdcodegen configure protobuf brpc leveldb)
 target_link_libraries(sdk-cpp brpc configure protobuf leveldb)
 
 # install

python/CMakeLists.txt

@@ -86,7 +86,7 @@ if (SERVER)
         elseif(CUDA_VERSION EQUAL 10.2)
             set(SUFFIX 102)
         elseif(CUDA_VERSION EQUAL 11.0)
-            set(SUFFIX 110)
+            set(SUFFIX 11)
 
         endif()
         add_custom_command(

python/examples/detection/README.md

+# Serve models from Paddle Detection
+
+(English|[简体中文](./README_CN.md))
+
+### Introduction
+
+PaddleDetection flying paddle target detection development kit is designed to help developers complete the whole development process of detection model formation, training, optimization and deployment faster and better. For details, see [Github](https://github.com/PaddlePaddle/PaddleDetection/tree/master/dygraph)
+
+This article mainly introduces the deployment of Paddle Detection's dynamic graph model on Serving.
+
+Paddle Detection provides a large number of [Model Zoo](https://github.com/PaddlePaddle/PaddleDetection/blob/master/dygraph/docs/MODEL_ZOO_cn.md), these model libraries can be used in Paddle Serving with export tools Model. For the export tutorial, please refer to [Paddle Detection Export Model Tutorial (Simplified Chinese)](https://github.com/PaddlePaddle/PaddleDetection/blob/master/dygraph/deploy/EXPORT_MODEL.md).
+
+### Serving example
+Several examples of PaddleDetection models used in Serving are given in this folder
+
+-[Faster RCNN](./faster_rcnn_r50_fpn_1x_coco)
+-[PPYOLO](./ppyolo_r50vd_dcn_1x_coco)
+-[TTFNet](./ttfnet_darknet53_1x_coco)
+-[YOLOv3](./yolov3_darknet53_270e_coco)

python/examples/detection/README_CN.md

+## 使用Paddle Detection模型
+
+([English](./README.md)|简体中文)
+
+### 简介
+
+PaddleDetection飞桨目标检测开发套件，旨在帮助开发者更快更好地完成检测模型的组建、训练、优化及部署等全开发流程。详情参见[Github](https://github.com/PaddlePaddle/PaddleDetection/tree/master/dygraph)
+
+本文主要是介绍Paddle Detection的动态图模型在Serving上的部署。
+
+### 导出模型
+
+Paddle Detection提供了大量的[模型库](https://github.com/PaddlePaddle/PaddleDetection/blob/master/dygraph/docs/MODEL_ZOO_cn.md), 这些模型库配合导出工具都可以得到可以用于Paddle Serving的模型。导出教程参见[Paddle Detection模型导出教程](https://github.com/PaddlePaddle/PaddleDetection/blob/master/dygraph/deploy/EXPORT_MODEL.md)。
+
+### Serving示例
+本文件夹下给出了多个PaddleDetection模型用于Serving的范例
+
+- [Faster RCNN](./faster_rcnn_r50_fpn_1x_coco)
+- [PPYOLO](./ppyolo_r50vd_dcn_1x_coco)
+- [TTFNet](./ttfnet_darknet53_1x_coco)
+- [YOLOv3](./yolov3_darknet53_270e_coco)
+

python/examples/faster_rcnn/README.md → python/examples/detection/faster_rcnn_r50_fpn_1x_coco/README.md

@@ -4,21 +4,18 @@
 
 ### Get The Faster RCNN Model
 ```
-wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/faster_rcnn_model.tar.gz
-wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/infer_cfg.yml
+wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/2.0/faster_rcnn_r50_fpn_1x_coco.tar
 ```
-If you want to have more detection models, please refer to [Paddle Detection Model Zoo](https://github.com/PaddlePaddle/PaddleDetection/blob/release/0.2/docs/MODEL_ZOO_cn.md)
 
 ### Start the service
 ```
-tar xf faster_rcnn_model.tar.gz
-mv faster_rcnn_model/pddet* .
-GLOG_v=2 python -m paddle_serving_server_gpu.serve --model pddet_serving_model --port 9494 --gpu_ids 0
+tar xf faster_rcnn_r50_fpn_1x_coco.tar
+python -m paddle_serving_server_gpu.serve --model serving_server --port 9494 --gpu_ids 0
 ```
 
 ### Perform prediction
 ```
-python test_client.py pddet_client_conf/serving_client_conf.prototxt infer_cfg.yml 000000570688.jpg
+python test_client.py 000000570688.jpg
 ```
 
 ## 3. Result analysis

python/examples/faster_rcnn/README_CN.md → python/examples/detection/faster_rcnn_r50_fpn_1x_coco/README_CN.md

@@ -4,21 +4,19 @@
 
 ## 获得Faster RCNN模型
 ```
-wget https://paddle-serving.bj.bcebos.com/pddet_demo/faster_rcnn_model.tar.gz
-wget https://paddle-serving.bj.bcebos.com/pddet_demo/infer_cfg.yml
+wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/2.0/faster_rcnn_r50_fpn_1x_coco.tar
 ```
-如果你想要更多的检测模型，请参考[Paddle检测模型库](https://github.com/PaddlePaddle/PaddleDetection/blob/release/0.2/docs/MODEL_ZOO_cn.md)
+
 
 ### 启动服务
 ```
-tar xf faster_rcnn_model.tar.gz
-mv faster_rcnn_model/pddet* ./
-GLOG_v=2 python -m paddle_serving_server_gpu.serve --model pddet_serving_model --port 9494 --gpu_ids 0
+tar xf faster_rcnn_r50_fpn_1x_coco.tar
+python -m paddle_serving_server_gpu.serve --model pddet_serving_model --port 9494 --gpu_ids 0
 ```
 
 ### 执行预测
 ```
-python test_client.py pddet_client_conf/serving_client_conf.prototxt infer_cfg.yml 000000570688.jpg
+python test_client.py 000000570688.jpg
 ```
 
 ## 3. 结果分析

python/examples/faster_rcnn/label_list.txt → python/examples/detection/faster_rcnn_r50_fpn_1x_coco/label_list.txt

-background
 person
 bicycle
 car

python/examples/faster_rcnn/test_client.py → python/examples/detection/faster_rcnn_r50_fpn_1x_coco/test_client.py

@@ -26,17 +26,18 @@ preprocess = Sequential([
 postprocess = RCNNPostprocess("label_list.txt", "output")
 client = Client()
 
-client.load_client_config(sys.argv[1])
+client.load_client_config("serving_client/serving_client_conf.prototxt")
 client.connect(['127.0.0.1:9494'])
 
-im = preprocess(sys.argv[3])
+im = preprocess(sys.argv[1])
 fetch_map = client.predict(
     feed={
         "image": im,
-        "im_info": np.array(list(im.shape[1:]) + [1.0]),
-        "im_shape": np.array(list(im.shape[1:]) + [1.0])
+        "im_shape": np.array(list(im.shape[1:])).reshape(-1),
+        "scale_factor": np.array([1.0, 1.0]).reshape(-1),
     },
-    fetch=["multiclass_nms"],
+    fetch=["save_infer_model/scale_0.tmp_1"],
     batch=False)
-fetch_map["image"] = sys.argv[3]
+print(fetch_map)
+fetch_map["image"] = sys.argv[1]
 postprocess(fetch_map)

python/examples/detection/ppyolo_r50vd_dcn_1x_coco/README.md

+# PP-YOLO model on Paddle Serving
+
+([简体中文](./README_CN.md)|English)
+
+### Get The Model
+```
+wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/2.0/ppyolo_r50vd_dcn_1x_coco.tar
+```
+
+### Start the service
+```
+tar xf ppyolo_r50vd_dcn_1x_coco.tar
+python -m paddle_serving_server_gpu.serve --model serving_server --port 9494 --gpu_ids 0
+```
+
+### Perform prediction
+```
+python test_client.py 000000570688.jpg
+```
+

python/examples/detection/ppyolo_r50vd_dcn_1x_coco/README_CN.md

+# 使用Paddle Serving部署PP-YOLO模型
+
+(简体中文|[English](./README.md))
+
+## 获得模型
+```
+wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/2.0/ppyolo_r50vd_dcn_1x_coco.tar
+```
+
+
+### 启动服务
+```
+tar xf ppyolo_r50vd_dcn_1x_coco.tar
+python -m paddle_serving_server_gpu.serve --model pddet_serving_model --port 9494 --gpu_ids 0
+```
+
+### 执行预测
+```
+python test_client.py 000000570688.jpg
+```
+

python/examples/detection/ppyolo_r50vd_dcn_1x_coco/label_list.txt

+person
+bicycle
+car
+motorcycle
+airplane
+bus
+train
+truck
+boat
+traffic light
+fire hydrant
+stop sign
+parking meter
+bench
+bird
+cat
+dog
+horse
+sheep
+cow
+elephant
+bear
+zebra
+giraffe
+backpack
+umbrella
+handbag
+tie
+suitcase
+frisbee
+skis
+snowboard
+sports ball
+kite
+baseball bat
+baseball glove
+skateboard
+surfboard
+tennis racket
+bottle
+wine glass
+cup
+fork
+knife
+spoon
+bowl
+banana
+apple
+sandwich
+orange
+broccoli
+carrot
+hot dog
+pizza
+donut
+cake
+chair
+couch
+potted plant
+bed
+dining table
+toilet
+tv
+laptop
+mouse
+remote
+keyboard
+cell phone
+microwave
+oven
+toaster
+sink
+refrigerator
+book
+clock
+vase
+scissors
+teddy bear
+hair drier
+toothbrush

python/examples/detection/ppyolo_r50vd_dcn_1x_coco/test_client.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from paddle_serving_client import Client
+from paddle_serving_app.reader import *
+import sys
+import numpy as np
+
+preprocess = Sequential([
+    File2Image(), BGR2RGB(), Div(255.0),
+    Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225], False),
+    Resize((608, 608)), Transpose((2, 0, 1))
+])
+
+postprocess = RCNNPostprocess("label_list.txt", "output")
+client = Client()
+
+client.load_client_config("serving_client/serving_client_conf.prototxt")
+client.connect(['127.0.0.1:9494'])
+
+im = preprocess(sys.argv[1])
+fetch_map = client.predict(
+    feed={
+        "image": im,
+        "im_shape": np.array(list(im.shape[1:])).reshape(-1),
+        "scale_factor": np.array([1.0, 1.0]).reshape(-1),
+    },
+    fetch=["save_infer_model/scale_0.tmp_1"],
+    batch=False)
+print(fetch_map)
+fetch_map["image"] = sys.argv[1]
+postprocess(fetch_map)

python/examples/detection/ttfnet_darknet53_1x_coco/README.md

+# TTF-Net model on Paddle Serving
+
+([简体中文](./README_CN.md)|English)
+
+### Get Model
+```
+wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/2.0/ttfnet_darknet53_1x_coco.tar
+```
+
+### Start the service
+```
+tar xf ttfnet_darknet53_1x_coco.tar
+python -m paddle_serving_server_gpu.serve --model serving_server --port 9494 --gpu_ids 0
+```
+
+### Perform prediction
+```
+python test_client.py 000000570688.jpg
+```
+

python/examples/detection/ttfnet_darknet53_1x_coco/README_CN.md

+# 使用Paddle Serving部署TTF-Net模型
+
+(简体中文|[English](./README.md))
+
+## 获得模型
+```
+wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/2.0/ttfnet_darknet53_1x_coco.tar
+```
+
+
+### 启动服务
+```
+tar xf ttfnet_darknet53_1x_coco.tar
+python -m paddle_serving_server_gpu.serve --model pddet_serving_model --port 9494 --gpu_ids 0
+```
+
+### 执行预测
+```
+python test_client.py 000000570688.jpg
+```
+

python/examples/detection/ttfnet_darknet53_1x_coco/label_list.txt

+person
+bicycle
+car
+motorcycle
+airplane
+bus
+train
+truck
+boat
+traffic light
+fire hydrant
+stop sign
+parking meter
+bench
+bird
+cat
+dog
+horse
+sheep
+cow
+elephant
+bear
+zebra
+giraffe
+backpack
+umbrella
+handbag
+tie
+suitcase
+frisbee
+skis
+snowboard
+sports ball
+kite
+baseball bat
+baseball glove
+skateboard
+surfboard
+tennis racket
+bottle
+wine glass
+cup
+fork
+knife
+spoon
+bowl
+banana
+apple
+sandwich
+orange
+broccoli
+carrot
+hot dog
+pizza
+donut
+cake
+chair
+couch
+potted plant
+bed
+dining table
+toilet
+tv
+laptop
+mouse
+remote
+keyboard
+cell phone
+microwave
+oven
+toaster
+sink
+refrigerator
+book
+clock
+vase
+scissors
+teddy bear
+hair drier
+toothbrush

python/examples/detection/ttfnet_darknet53_1x_coco/test_client.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from paddle_serving_client import Client
+from paddle_serving_app.reader import *
+import sys
+import numpy as np
+
+preprocess = Sequential([
+    File2Image(), BGR2RGB(), 
+    Normalize([123.675, 116.28, 103.53], [58.395, 57.12, 57.375], False),
+    Resize((512, 512)), Transpose((2, 0, 1))
+])
+
+postprocess = RCNNPostprocess("label_list.txt", "output")
+client = Client()
+
+client.load_client_config("serving_client/serving_client_conf.prototxt")
+client.connect(['127.0.0.1:9494'])
+
+im = preprocess(sys.argv[1])
+fetch_map = client.predict(
+    feed={
+        "image": im,
+        "scale_factor": np.array([1.0, 1.0]).reshape(-1),
+    },
+    fetch=["save_infer_model/scale_0.tmp_1"],
+    batch=False)
+print(fetch_map)

python/examples/detection/yolov3_darknet53_270e_coco/README.md

+# YOLOv3 model on Paddle Serving
+
+([简体中文](./README_CN.md)|English)
+
+### Get Model
+```
+wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/2.0/yolov3_darknet53_270e_coco.tar
+```
+
+### Start the service
+```
+tar xf yolov3_darknet53_270e_coco.tar
+python -m paddle_serving_server_gpu.serve --model serving_server --port 9494 --gpu_ids 0
+```
+
+### Perform prediction
+```
+python test_client.py 000000570688.jpg
+```
+

python/examples/detection/yolov3_darknet53_270e_coco/README_CN.md

+# 使用Paddle Serving部署YOLOv3模型
+
+(简体中文|[English](./README.md))
+
+## 获得模型
+```
+wget --no-check-certificate https://paddle-serving.bj.bcebos.com/pddet_demo/2.0/yolov3_darknet53_270e_coco.tar
+```
+
+
+### 启动服务
+```
+tar xf yolov3_darknet53_270e_coco.tar
+python -m paddle_serving_server_gpu.serve --model pddet_serving_model --port 9494 --gpu_ids 0
+```
+
+### 执行预测
+```
+python test_client.py 000000570688.jpg
+```
+

python/examples/detection/yolov3_darknet53_270e_coco/label_list.txt

+person
+bicycle
+car
+motorcycle
+airplane
+bus
+train
+truck
+boat
+traffic light
+fire hydrant
+stop sign
+parking meter
+bench
+bird
+cat
+dog
+horse
+sheep
+cow
+elephant
+bear
+zebra
+giraffe
+backpack
+umbrella
+handbag
+tie
+suitcase
+frisbee
+skis
+snowboard
+sports ball
+kite
+baseball bat
+baseball glove
+skateboard
+surfboard
+tennis racket
+bottle
+wine glass
+cup
+fork
+knife
+spoon
+bowl
+banana
+apple
+sandwich
+orange
+broccoli
+carrot
+hot dog
+pizza
+donut
+cake
+chair
+couch
+potted plant
+bed
+dining table
+toilet
+tv
+laptop
+mouse
+remote
+keyboard
+cell phone
+microwave
+oven
+toaster
+sink
+refrigerator
+book
+clock
+vase
+scissors
+teddy bear
+hair drier
+toothbrush

python/examples/detection/yolov3_darknet53_270e_coco/test_client.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from paddle_serving_client import Client
+from paddle_serving_app.reader import *
+import sys
+import numpy as np
+
+preprocess = Sequential([
+    File2Image(), BGR2RGB(), Div(255.0),
+    Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225], False),
+    Resize((608, 608)), Transpose((2, 0, 1))
+])
+
+postprocess = RCNNPostprocess("label_list.txt", "output")
+client = Client()
+
+client.load_client_config("serving_client/serving_client_conf.prototxt")
+client.connect(['127.0.0.1:9494'])
+
+im = preprocess(sys.argv[1])
+fetch_map = client.predict(
+    feed={
+        "image": im,
+        "im_shape": np.array(list(im.shape[1:])).reshape(-1),
+        "scale_factor": np.array([1.0, 1.0]).reshape(-1),
+    },
+    fetch=["save_infer_model/scale_0.tmp_1"],
+    batch=False)
+print(fetch_map)
+fetch_map["image"] = sys.argv[1]
+postprocess(fetch_map)

python/examples/faster_rcnn/benchmark.py

-# -*- coding: utf-8 -*-
-#
-# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-# pylint: disable=doc-string-missing
-
-from __future__ import unicode_literals, absolute_import
-import os
-import sys
-import time
-import json
-import requests
-from paddle_serving_client import Client
-from paddle_serving_client.utils import MultiThreadRunner
-from paddle_serving_client.utils import benchmark_args, show_latency
-from paddle_serving_app.reader import ChineseBertReader
-
-from paddle_serving_app.reader import *
-import numpy as np
-
-args = benchmark_args()
-
-
-def single_func(idx, resource):
-    img = "./000000570688.jpg"
-    profile_flags = False
-    latency_flags = False
-    if os.getenv("FLAGS_profile_client"):
-        profile_flags = True
-    if os.getenv("FLAGS_serving_latency"):
-        latency_flags = True
-        latency_list = []
-
-    if args.request == "rpc":
-        preprocess = Sequential([
-            File2Image(), BGR2RGB(), Div(255.0),
-            Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225], False),
-            Resize(640, 640), Transpose((2, 0, 1))
-        ])
-
-        postprocess = RCNNPostprocess("label_list.txt", "output")
-        client = Client()
-
-        client.load_client_config(args.model)
-        client.connect([resource["endpoint"][idx % len(resource["endpoint"])]])
-
-        start = time.time()
-        for i in range(turns):
-            if args.batch_size >= 1:
-                l_start = time.time()
-                feed_batch = []
-                b_start = time.time()
-                im = preprocess(img)
-                for bi in range(args.batch_size):
-                    print("1111batch")
-                    print(bi)
-                    feed_batch.append({
-                        "image": im,
-                        "im_info": np.array(list(im.shape[1:]) + [1.0]),
-                        "im_shape": np.array(list(im.shape[1:]) + [1.0])
-                    })
-            # im = preprocess(img)
-                b_end = time.time()
-
-                if profile_flags:
-                    sys.stderr.write(
-                        "PROFILE\tpid:{}\tbert_pre_0:{} bert_pre_1:{}\n".format(
-                            os.getpid(),
-                            int(round(b_start * 1000000)),
-                            int(round(b_end * 1000000))))
-                #result = client.predict(feed=feed_batch, fetch=fetch)
-                fetch_map = client.predict(
-                    feed=feed_batch, fetch=["multiclass_nms"])
-                fetch_map["image"] = img
-                postprocess(fetch_map)
-
-                l_end = time.time()
-                if latency_flags:
-                    latency_list.append(l_end * 1000 - l_start * 1000)
-            else:
-                print("unsupport batch size {}".format(args.batch_size))
-    else:
-        raise ValueError("not implemented {} request".format(args.request))
-    end = time.time()
-    if latency_flags:
-        return [[end - start], latency_list]
-    else:
-        return [[end - start]]
-
-
-if __name__ == '__main__':
-    multi_thread_runner = MultiThreadRunner()
-    endpoint_list = ["127.0.0.1:7777"]
-    turns = 10
-    start = time.time()
-    result = multi_thread_runner.run(
-        single_func, args.thread, {"endpoint": endpoint_list,
-                                   "turns": turns})
-    end = time.time()
-    total_cost = end - start
-
-    avg_cost = 0
-    for i in range(args.thread):
-        avg_cost += result[0][i]
-    avg_cost = avg_cost / args.thread
-
-    print("total cost: {}s".format(total_cost))
-    print("each thread cost: {}s. ".format(avg_cost))
-    print("qps: {}samples/s".format(args.batch_size * args.thread * turns /
-                                    total_cost))
-    if os.getenv("FLAGS_serving_latency"):
-        show_latency(result[1])

python/examples/faster_rcnn/benchmark.sh

-rm profile_log*
-export CUDA_VISIBLE_DEVICES=0
-export FLAGS_profile_server=1
-export FLAGS_profile_client=1
-export FLAGS_serving_latency=1
-
-gpu_id=0
-#save cpu and gpu utilization log
-if [ -d utilization ];then
-    rm -rf utilization
-else
-    mkdir utilization
-fi
-#start server
-$PYTHONROOT/bin/python3 -m paddle_serving_server_gpu.serve --model $1 --port 7777 --thread 4 --gpu_ids 0  --ir_optim >  elog  2>&1 &
-sleep 5
-
-#warm up
-$PYTHONROOT/bin/python3 benchmark.py --thread 4 --batch_size 1 --model $2/serving_client_conf.prototxt --request rpc > profile 2>&1
-echo -e "import psutil\ncpu_utilization=psutil.cpu_percent(1,False)\nprint('CPU_UTILIZATION:', cpu_utilization)\n" > cpu_utilization.py
-for thread_num in 1 4 8 16
-do
-for batch_size in 1
-do
-    job_bt=`date '+%Y%m%d%H%M%S'`
-    nvidia-smi --id=0 --query-compute-apps=used_memory --format=csv -lms 100 > gpu_use.log 2>&1 &
-    nvidia-smi --id=0 --query-gpu=utilization.gpu --format=csv -lms 100 > gpu_utilization.log 2>&1 &
-    gpu_memory_pid=$!
-    $PYTHONROOT/bin/python3 benchmark.py --thread $thread_num --batch_size $batch_size --model $2/serving_client_conf.prototxt --request rpc > profile 2>&1
-    kill ${gpu_memory_pid}
-    kill `ps -ef|grep used_memory|awk '{print $2}'`
-    echo "model_name:" $1
-    echo "thread_num:" $thread_num
-    echo "batch_size:" $batch_size
-    echo "=================Done===================="
-    echo "model_name:$1" >> profile_log_$1
-    echo "batch_size:$batch_size" >> profile_log_$1
-    $PYTHONROOT/bin/python3 cpu_utilization.py >> profile_log_$1
-    job_et=`date '+%Y%m%d%H%M%S'`
-    awk 'BEGIN {max = 0} {if(NR>1){if ($1 > max) max=$1}} END {print "MAX_GPU_MEMORY:", max}' gpu_use.log >> profile_log_$1
-    awk 'BEGIN {max = 0} {if(NR>1){if ($1 > max) max=$1}} END {print "GPU_UTILIZATION:", max}' gpu_utilization.log >> profile_log_$1
-    rm -rf gpu_use.log gpu_utilization.log
-    $PYTHONROOT/bin/python3 ../util/show_profile.py profile $thread_num >> profile_log_$1
-    tail -n 8 profile >> profile_log_$1
-    echo "" >> profile_log_$1
-done
-done
-
-#Divided log
-awk 'BEGIN{RS="\n\n"}{i++}{print > "bert_log_"i}' profile_log_$1
-mkdir bert_log && mv bert_log_* bert_log
-ps -ef|grep 'serving'|grep -v grep|cut -c 9-15 | xargs kill -9

python/paddle_serving_server_gpu/__init__.py

@@ -442,7 +442,7 @@ class Server(object):
         for line in version_file.readlines():
             if re.match("cuda_version", line):
                 cuda_version = line.split("\"")[1]
-                if cuda_version == "101" or cuda_version == "102" or cuda_version == "110":
+                if cuda_version == "101" or cuda_version == "102":
                     device_version = "serving-gpu-" + cuda_version + "-"
                 elif cuda_version == "arm" or cuda_version == "arm-xpu":
                     device_version = "serving-" + cuda_version + "-"

python/setup.py.app.in

@@ -32,7 +32,7 @@ if '${PACK}' == 'ON':
 
 
 REQUIRED_PACKAGES = [
-    'six >= 1.10.0', 'sentencepiece', 'opencv-python', 'pillow',
+    'six >= 1.10.0', 'sentencepiece<=0.1.83', 'opencv-python<=4.2.0.32', 'pillow',
     'pyclipper'
 ]
 

python/setup.py.client.in

@@ -44,7 +44,7 @@ if '${PACK}' == 'ON':
 
 REQUIRED_PACKAGES = [
     'six >= 1.10.0', 'protobuf >= 3.11.0', 'numpy >= 1.12', 'grpcio <= 1.33.2',
-    'grpcio-tools <= 1.33.2'
+    'grpcio-tools <= 1.33.2', 'requests'
 ]