Merge branch 'develop' into xpu-example

CMakeLists.txt

@@ -60,6 +60,13 @@ option(WITH_TRT             "Compile Paddle Serving with TRT"
 option(PADDLE_ON_INFERENCE  "Compile for encryption"                             ON)
 option(WITH_OPENCV	    "Compile Paddle Serving with OPENCV"                    OFF)
 
+if(NOT DEFINED VERSION_TAG)
+  set(VERSION_TAG "0.0.0")
+endif()
+if (WITH_PYTHON)
+  message(STATUS "Compile Version Tag for wheel: ${VERSION_TAG}")
+endif()
+
 if (WITH_OPENCV)
     SET(OPENCV_DIR "" CACHE PATH "Location of libraries")
     if(NOT DEFINED OPENCV_DIR)

doc/BAIDU_KUNLUN_XPU_SERVING.md

@@ -59,11 +59,11 @@ For example, after the Server Compiation step，the whl package will be produced
 # Request parameters description
 In order to deploy serving
  service on the arm server with Baidu Kunlun xpu chips and use the acceleration capability of Paddle-Lite，please specify the following parameters during deployment。
-|param|param description|about|
-|:--|:--|:--|
-|use_lite|using Paddle-Lite Engine|use the inference capability of Paddle-Lite|
-|use_xpu|using Baidu Kunlun for inference|need to be used with the use_lite option|
-|ir_optim|open the graph optimization|refer to[Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite)|
+| param    | param description                | about                                                              |
+| :------- | :------------------------------- | :----------------------------------------------------------------- |
+| use_lite | using Paddle-Lite Engine         | use the inference capability of Paddle-Lite                        |
+| use_xpu  | using Baidu Kunlun for inference | need to be used with the use_lite option                           |
+| ir_optim | open the graph optimization      | refer to[Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) |
 # Deplyment examples
 ## Download the model
 ```
@@ -78,15 +78,15 @@ There are mainly three deployment methods：
     
 The first two deployment methods are recommended。
 
-Start the rpc service, deploying on ARM server with Baidu Kunlun chips，and accelerate with Paddle-Lite and Baidu Kunlun xpu.
+Start the rpc service, deploying on ARM server with Baidu Kunlun chips，and accelerate with Paddle-Lite and Baidu Kunlun xpu.
 ```
 python3 -m paddle_serving_server_gpu.serve --model uci_housing_model --thread 6 --port 9292 --use_lite --use_xpu --ir_optim
 ```
-Start the rpc service, deploying on ARM server，and accelerate with Paddle-Lite.
+Start the rpc service, deploying on ARM server，and accelerate with Paddle-Lite.
 ```
 python3 -m paddle_serving_server_gpu.serve --model uci_housing_model --thread 6 --port 9292 --use_lite --ir_optim
 ```
-Start the rpc service, deploying on ARM server.
+Start the rpc service, deploying on ARM server.
 ```
 python3 -m paddle_serving_server_gpu.serve --model uci_housing_model --thread 6 --port 9292
 ```
@@ -103,7 +103,7 @@ fetch_map = client.predict(feed={"x": np.array(data).reshape(1,13,1)}, fetch=["p
 print(fetch_map)
 ```
 Some examples are provided below, and other models can be modifed with reference to these examples。
-|sample name|sample links|
-|:-----|:--|
-|fit_a_line|[fit_a_line_xpu](../python/examples/xpu/fit_a_line_xpu)|
-|resnet|[resnet_v2_50_xpu](../python/examples/xpu/resnet_v2_50_xpu)|
+| sample name | sample links                                                |
+| :---------- | :---------------------------------------------------------- |
+| fit_a_line  | [fit_a_line_xpu](../python/examples/xpu/fit_a_line_xpu)     |
+| resnet      | [resnet_v2_50_xpu](../python/examples/xpu/resnet_v2_50_xpu) |

doc/LOW_PRECISION_DEPLOYMENT.md

+# Low-Precision Deployment for Paddle Serving
+(English|[简体中文](./LOW_PRECISION_DEPLOYMENT_CN.md))
+
+Intel CPU supports int8 and bfloat16 models, NVIDIA TensorRT supports int8 and float16 models.
+
+## Obtain the quantized model through PaddleSlim tool
+Train the low-precision models please refer to [PaddleSlim](https://paddleslim.readthedocs.io/zh_CN/latest/tutorials/quant/overview.html).
+
+## Deploy the quantized model from PaddleSlim using Paddle Serving with Nvidia TensorRT int8 mode
+
+Firstly, download the [Resnet50 int8 model](https://paddle-inference-dist.bj.bcebos.com/inference_demo/python/resnet50/ResNet50_quant.tar.gz) and convert to Paddle Serving's saved model。
+```
+wget https://paddle-inference-dist.bj.bcebos.com/inference_demo/python/resnet50/ResNet50_quant.tar.gz
+tar zxvf ResNet50_quant.tar.gz
+
+python -m paddle_serving_client.convert --dirname ResNet50_quant
+```
+Start RPC service, specify the GPU id and precision mode
+```
+python -m paddle_serving_server.serve --model serving_server --port 9393 --gpu_ids 0 --use_gpu --use_trt --precision int8 
+```
+Request the serving service with Client
+```
+from paddle_serving_client import Client
+from paddle_serving_app.reader import Sequential, File2Image, Resize, CenterCrop
+from paddle_serving_app.reader import RGB2BGR, Transpose, Div, Normalize
+
+client = Client()
+client.load_client_config(
+    "resnet_v2_50_imagenet_client/serving_client_conf.prototxt")
+client.connect(["127.0.0.1:9393"])
+
+seq = Sequential([
+    File2Image(), Resize(256), CenterCrop(224), RGB2BGR(), Transpose((2, 0, 1)),
+    Div(255), Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225], True)
+])
+
+image_file = "daisy.jpg"
+img = seq(image_file)
+fetch_map = client.predict(feed={"image": img}, fetch=["score"])
+print(fetch_map["score"].reshape(-1))
+```
+
+## Reference
+* [PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim)
+* [Deploy the quantized model Using Paddle Inference on Intel CPU](https://paddle-inference.readthedocs.io/en/latest/optimize/paddle_x86_cpu_int8.html)
+* [Deploy the quantized model Using Paddle Inference on Nvidia GPU](https://paddle-inference.readthedocs.io/en/latest/optimize/paddle_trt.html)
\ No newline at end of file

doc/LOW_PRECISION_DEPLOYMENT_CN.md

+# Paddle Serving低精度部署
+(简体中文|[English](./LOW_PRECISION_DEPLOYMENT.md))
+
+低精度部署, 在Intel CPU上支持int8、bfloat16模型，Nvidia TensorRT支持int8、float16模型。
+
+## 通过PaddleSlim量化生成低精度模型
+详细见[PaddleSlim量化](https://paddleslim.readthedocs.io/zh_CN/latest/tutorials/quant/overview.html)
+
+## 使用TensorRT int8加载PaddleSlim Int8量化模型进行部署
+首先下载Resnet50 [PaddleSlim量化模型](https://paddle-inference-dist.bj.bcebos.com/inference_demo/python/resnet50/ResNet50_quant.tar.gz),并转换为Paddle Serving支持的部署模型格式。
+```
+wget https://paddle-inference-dist.bj.bcebos.com/inference_demo/python/resnet50/ResNet50_quant.tar.gz
+tar zxvf ResNet50_quant.tar.gz
+
+python -m paddle_serving_client.convert --dirname ResNet50_quant
+```
+启动rpc服务, 设定所选GPU id、部署模型精度
+```
+python -m paddle_serving_server.serve --model serving_server --port 9393 --gpu_ids 0 --use_gpu --use_trt --precision int8 
+```
+使用client进行请求
+```
+from paddle_serving_client import Client
+from paddle_serving_app.reader import Sequential, File2Image, Resize, CenterCrop
+from paddle_serving_app.reader import RGB2BGR, Transpose, Div, Normalize
+
+client = Client()
+client.load_client_config(
+    "resnet_v2_50_imagenet_client/serving_client_conf.prototxt")
+client.connect(["127.0.0.1:9393"])
+
+seq = Sequential([
+    File2Image(), Resize(256), CenterCrop(224), RGB2BGR(), Transpose((2, 0, 1)),
+    Div(255), Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225], True)
+])
+
+image_file = "daisy.jpg"
+img = seq(image_file)
+fetch_map = client.predict(feed={"image": img}, fetch=["score"])
+print(fetch_map["score"].reshape(-1))
+```
+
+## 参考文档
+* [PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim)
+* PaddleInference Intel CPU部署量化模型[文档](https://paddle-inference.readthedocs.io/en/latest/optimize/paddle_x86_cpu_int8.html)
+* PaddleInference NV GPU部署量化模型[文档](https://paddle-inference.readthedocs.io/en/latest/optimize/paddle_trt.html)
\ No newline at end of file

python/gen_version.py

@@ -47,3 +47,5 @@ elif package_name.endswith('xpu'):
 path = "paddle_serving_" + sys.argv[1]
 commit_id = subprocess.check_output(['git', 'rev-parse', 'HEAD'])
 update_info(path + "/version.py", "commit_id", commit_id)
+
+update_info(path + "/version.py", "version_tag", "${VERSION_TAG}")

python/paddle_serving_app/__init__.py

@@ -12,3 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 from .models import ServingModels
+from . import version
+
+__version__ = version.version_tag

python/paddle_serving_app/version.py

@@ -12,5 +12,5 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """ Paddle Serving App version string """
-serving_app_version = "0.0.0"
+version_tag = "0.0.0"
 commit_id = ""

python/paddle_serving_client/__init__.py

@@ -17,4 +17,4 @@ from . import version
 from . import client
 from .client import *
 
-__version__ = version.serving_client_version
+__version__ = version.version_tag

python/paddle_serving_client/version.py

@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """ Paddle Serving Client version string """
-serving_client_version = "0.0.0"
-serving_server_version = "0.0.0"
+version_tag = "0.0.0"
 module_proto_version = "0.0.0"
 commit_id = ""

python/paddle_serving_server/__init__.py

@@ -31,4 +31,4 @@ from paddle_serving_server import (
 from .dag import *
 from .server import *
 
-__version__ = version.serving_server_version
+__version__ = version.version_tag

python/paddle_serving_server/server.py

@@ -22,7 +22,7 @@ from .proto import general_model_config_pb2 as m_config
 from .proto import multi_lang_general_model_service_pb2_grpc
 import google.protobuf.text_format
 import time
-from .version import serving_server_version, version_suffix, device_type
+from .version import version_tag, version_suffix, device_type
 from contextlib import closing
 import argparse
 
@@ -369,7 +369,7 @@ class Server(object):
         version_file = open("{}/version.py".format(self.module_path), "r")
 
         folder_name = "serving-%s-%s" % (self.get_serving_bin_name(),
-                                         serving_server_version)
+                                         version_tag)
         tar_name = "%s.tar.gz" % folder_name
         bin_url = "https://paddle-serving.bj.bcebos.com/bin/%s" % tar_name
 

python/paddle_serving_server/version.py

@@ -12,10 +12,9 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """ Paddle Serving Server version string """
-serving_client_version = "0.0.0"
-serving_server_version = "0.0.0"
-module_proto_version = "0.0.0"
+version_tag = "0.0.0"
 version_suffix = ""
+module_proto_version = "0.0.0"
 device_type = "0"
 cuda_version = "9"
 commit_id = ""

python/setup.py.app.in

@@ -22,7 +22,7 @@ import os
 from setuptools import setup, Distribution, Extension
 from setuptools import find_packages
 from setuptools import setup
-from paddle_serving_app.version import serving_app_version
+from paddle_serving_app.version import version_tag
 from pkg_resources import DistributionNotFound, get_distribution
         
 def python_version():
@@ -78,7 +78,7 @@ package_dir={'paddle_serving_app':
 
 setup(
     name='paddle-serving-app',
-    version=serving_app_version.replace('-', ''),
+    version=version_tag.replace('-', ''),
     description=
     ('Paddle Serving Package for saved model with PaddlePaddle'),
     url='https://github.com/PaddlePaddle/Serving',

python/setup.py.client.in

@@ -22,7 +22,7 @@ import sys
 from setuptools import setup, Distribution, Extension
 from setuptools import find_packages
 from setuptools import setup
-from paddle_serving_client.version import serving_client_version
+from paddle_serving_client.version import version_tag
 import util
 
 py_version = sys.version_info
@@ -79,7 +79,7 @@ package_dir={'paddle_serving_client':
 
 setup(
     name='paddle-serving-client',
-    version=serving_client_version.replace('-', ''),
+    version=version_tag.replace('-', ''),
     description=
     ('Paddle Serving Package for saved model with PaddlePaddle'),
     url='https://github.com/PaddlePaddle/Serving',

python/setup.py.in

@@ -19,7 +19,7 @@ from __future__ import print_function
 from setuptools import setup, Distribution, Extension
 from setuptools import find_packages
 from setuptools import setup
-from paddle_serving.version import serving_client_version
+from paddle_serving.version import version_tag
 from grpc_tools import protoc
 import util
 
@@ -43,7 +43,7 @@ package_dir={'paddle_serving.serving_client':
 
 setup(
     name='paddle-serving-client',
-    version=serving_client_version.replace('-', ''),
+    version=version_tag.replace('-', ''),
     description=
     ('Paddle Serving Package for saved model with PaddlePaddle'),
     url='https://github.com/PaddlePaddle/Serving',

python/setup.py.server.in

@@ -19,10 +19,10 @@ from __future__ import print_function
 from setuptools import setup, Distribution, Extension
 from setuptools import find_packages
 from setuptools import setup
-from paddle_serving_server.version import serving_server_version, version_suffix
+from paddle_serving_server.version import version_tag, version_suffix
 import util
 
-package_version = serving_server_version.replace('-', '')
+package_version = version_tag.replace('-', '')
 if version_suffix != "":
     version_suffix = "post" + version_suffix
     package_version = package_version + "." + version_suffix