未验证 提交 f385cb02 编写于 作者: T TeslaZhao 提交者: GitHub

Merge pull request #1166 from zhangjun/download-fix

fix serving_bin download, add low-precision example
......@@ -17,7 +17,7 @@ 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
python -m paddle_serving_server.serve --model serving_server --port 9393 --gpu_ids 0 --use_trt --precision int8
```
Request the serving service with Client
```
......@@ -27,7 +27,7 @@ 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")
"serving_client/serving_client_conf.prototxt")
client.connect(["127.0.0.1:9393"])
seq = Sequential([
......@@ -37,11 +37,11 @@ seq = Sequential([
image_file = "daisy.jpg"
img = seq(image_file)
fetch_map = client.predict(feed={"image": img}, fetch=["score"])
print(fetch_map["score"].reshape(-1))
fetch_map = client.predict(feed={"image": img}, fetch=["save_infer_model/scale_0.tmp_0"])
print(fetch_map["save_infer_model/scale_0.tmp_0"].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
* [Deploy the quantized model Using Paddle Inference on Nvidia GPU](https://paddle-inference.readthedocs.io/en/latest/optimize/paddle_trt.html)
......@@ -16,7 +16,7 @@ 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
python -m paddle_serving_server.serve --model serving_server --port 9393 --gpu_ids 0 --use_trt --precision int8
```
使用client进行请求
```
......@@ -43,4 +43,4 @@ 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
* PaddleInference NV GPU部署量化模型[文档](https://paddle-inference.readthedocs.io/en/latest/optimize/paddle_trt.html)
# resnet50 int8 example
(English|[简体中文](./README_CN.md))
## 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_trt --precision int8
```
Request the serving service with Client
```
python resnet50_client.py
```
## 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)
# resnet50 int8示例
(简体中文|[English](./README.md))
## 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_trt --precision int8
```
使用client进行请求
```
python resnet50_client.py
```
## 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)
# 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 Sequential, File2Image, Resize, CenterCrop
from paddle_serving_app.reader import RGB2BGR, Transpose, Div, Normalize
client = Client()
client.load_client_config(
"serving_client/serving_client_conf.prototxt")
client.connect(["127.0.0.1:9303"])
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=["save_infer_model/scale_0.tmp_0"])
print(fetch_map["save_infer_model/scale_0.tmp_0"].reshape(-1))
......@@ -386,8 +386,6 @@ class Server(object):
return
if not os.path.exists(self.server_path):
os.system("touch {}/{}.is_download".format(self.module_path,
folder_name))
print('Frist time run, downloading PaddleServing components ...')
r = os.system('wget ' + bin_url + ' --no-check-certificate')
......@@ -403,9 +401,10 @@ class Server(object):
tar = tarfile.open(tar_name)
tar.extractall()
tar.close()
open(download_flag, "a").close()
except:
if os.path.exists(exe_path):
os.remove(exe_path)
if os.path.exists(self.server_path):
os.remove(self.server_path)
raise SystemExit(
'Decompressing failed, please check your permission of {} or disk space left.'
.format(self.module_path))
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册