From efa634ebf8da128e35aa6d3041c69acb42c7af99 Mon Sep 17 00:00:00 2001 From: yunyaoXYY <109218879+yunyaoXYY@users.noreply.github.com> Date: Mon, 14 Nov 2022 15:16:14 +0800 Subject: [PATCH] [Quantization] Change the usage of FastDeploy auto compression tool. (#576) * Add PaddleOCR Support * Add PaddleOCR Support * Add PaddleOCRv3 Support * Add PaddleOCRv3 Support * Update README.md * Update README.md * Update README.md * Update README.md * Add PaddleOCRv3 Support * Add PaddleOCRv3 Supports * Add PaddleOCRv3 Suport * Fix Rec diff * Remove useless functions * Remove useless comments * Add PaddleOCRv2 Support * Add PaddleOCRv3 & PaddleOCRv2 Support * remove useless parameters * Add utils of sorting det boxes * Fix code naming convention * Fix code naming convention * Fix code naming convention * Fix bug in the Classify process * Imporve OCR Readme * Fix diff in Cls model * Update Model Download Link in Readme * Fix diff in PPOCRv2 * Improve OCR readme * Imporve OCR readme * Improve OCR readme * Improve OCR readme * Imporve OCR readme * Improve OCR readme * Fix conflict * Add readme for OCRResult * Improve OCR readme * Add OCRResult readme * Improve OCR readme * Improve OCR readme * Add Model Quantization Demo * Fix Model Quantization Readme * Fix Model Quantization Readme * Add the function to do PTQ quantization * Improve quant tools readme * Improve quant tool readme * Improve quant tool readme * Add PaddleInference-GPU for OCR Rec model * Add QAT method to fastdeploy-quantization tool * Remove examples/slim for now * Move configs folder * Add Quantization Support for Classification Model * Imporve ways of importing preprocess * Upload YOLO Benchmark on readme * Upload YOLO Benchmark on readme * Upload YOLO Benchmark on readme * Improve Quantization configs and readme * Add support for multi-inputs model * Add backends and params file for YOLOv7 * Add quantized model deployment support for YOLO series * Fix YOLOv5 quantize readme * Fix YOLO quantize readme * Fix YOLO quantize readme * Improve quantize YOLO readme * Improve quantize YOLO readme * Improve quantize YOLO readme * Improve quantize YOLO readme * Improve quantize YOLO readme * Fix bug, change Fronted to ModelFormat * Change Fronted to ModelFormat * Add examples to deploy quantized paddleclas models * Fix readme * Add quantize Readme * Add quantize Readme * Add quantize Readme * Modify readme of quantization tools * Modify readme of quantization tools * Improve quantization tools readme * Improve quantization readme * Improve PaddleClas quantized model deployment readme * Add PPYOLOE-l quantized deployment examples * Improve quantization tools readme * Improve Quantize Readme * Fix conflicts * Fix conflicts * improve readme * Improve quantization tools and readme * Improve quantization tools and readme * Add quantized deployment examples for PaddleSeg model * Fix cpp readme * Fix memory leak of reader_wrapper function * Fix model file name in PaddleClas quantization examples * Update Runtime and E2E benchmark * Update Runtime and E2E benchmark * Rename quantization tools to auto compression tools * Remove PPYOLOE data when deployed on MKLDNN * Fix readme * Support PPYOLOE with OR without NMS and update readme * Update Readme * Update configs and readme * Update configs and readme * Add Paddle-TensorRT backend in quantized model deploy examples * Support PPYOLOE+ series * Add reused_input_tensors for PPYOLOE * Improve fastdeploy tools usage * improve fastdeploy tool * Improve fastdeploy auto compression tool * Improve fastdeploy auto compression tool * Improve fastdeploy auto compression tool * Improve fastdeploy auto compression tool * Improve fastdeploy auto compression tool * remove modify * Improve fastdeploy auto compression tool * Improve fastdeploy auto compression tool * Improve fastdeploy auto compression tool * Improve fastdeploy auto compression tool * Improve fastdeploy auto compression tool --- python/requirements.txt | 1 + tools/README.md | 35 +++++++++++++ tools/README_EN.md | 35 +++++++++++++ tools/auto_compression/README.md | 14 ++--- tools/auto_compression/README_EN.md | 16 +++--- .../fd_auto_compress/__init__.py | 1 + .../fd_auto_compress/fd_auto_compress.py | 10 +--- tools/auto_compression/setup.py | 15 +++--- tools/common_tools/__init__.py | 0 tools/common_tools/common_tools.py | 51 +++++++++++++++++++ tools/setup.py | 20 ++++++++ 11 files changed, 169 insertions(+), 29 deletions(-) create mode 100644 tools/README.md create mode 100644 tools/README_EN.md create mode 100644 tools/common_tools/__init__.py create mode 100644 tools/common_tools/common_tools.py create mode 100644 tools/setup.py diff --git a/python/requirements.txt b/python/requirements.txt index b4fc12689..03755825d 100644 --- a/python/requirements.txt +++ b/python/requirements.txt @@ -3,3 +3,4 @@ requests tqdm numpy opencv-python +fd-auto-compress>=0.0.0 diff --git a/tools/README.md b/tools/README.md new file mode 100644 index 000000000..ccc6cca69 --- /dev/null +++ b/tools/README.md @@ -0,0 +1,35 @@ +# FastDeploy 工具包 +FastDeploy提供了一系列高效易用的工具优化部署体验, 提升推理性能. +例如, FastDeploy基于PaddleSlim的Auto Compression Toolkit(ACT), 给用户提供了一键模型自动化压缩的工具, 用户可以轻松地通过一行命令对模型进行自动化压缩, 并在FastDeploy上部署压缩后的模型, 提升推理速度. 本文档将以FastDeploy一键模型自动化压缩工具为例, 介绍如何安装此工具, 并提供相应的使用文档. + +## FastDeploy一键模型自动化压缩工具 + +### 环境准备 +1.用户参考PaddlePaddle官网, 安装develop版本 +``` +https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/develop/install/pip/linux-pip.html +``` + +2.安装PaddleSlim develop版本 +```bash +git clone https://github.com/PaddlePaddle/PaddleSlim.git & cd PaddleSlim +python setup.py install +``` + +3.安装fd-auto-compress一键模型自动化压缩工具 +```bash +# 通过pip安装fd-auto-compress. +# FastDeploy的python包已包含此工具, 不需重复安装. +pip install fd-auto-compress + +# 在当前目录执行以下命令 +python setup.py install +``` + +### 一键模型自动化压缩工具的使用 +按照以上步骤成功安装后,即可使用FastDeploy一键模型自动化压缩工具, 示例如下. + +```bash +fastdeploy --auto_compress --config_path=./configs/detection/yolov5s_quant.yaml --method='PTQ' --save_dir='./yolov5s_ptq_model/' +``` +详细使用文档请参考[FastDeploy一键模型自动化压缩工具](./auto_compression/README.md) diff --git a/tools/README_EN.md b/tools/README_EN.md new file mode 100644 index 000000000..4e30972b1 --- /dev/null +++ b/tools/README_EN.md @@ -0,0 +1,35 @@ +# FastDeploy Toolkit +FastDeploy provides a series of efficient and easy-to-use tools to optimize the deployment experience and improve inference performance. +For example, based on PaddleSlim's Auto Compression Toolkit (ACT), FastDeploy provides users with a one-click model automation compression tool that allows users to easily compress the model with a single command. This document will take FastDeploy's one-click model automation compression tool as an example, introduce how to install the tool, and provide the corresponding documentation for usage. + + +## FastDeploy One-Click Model Auto Compression Tool + +### Environmental Preparation +1.Install PaddlePaddle develop version +``` +https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/develop/install/pip/linux-pip.html +``` + +2.Install PaddleSlim dev version +```bash +git clone https://github.com/PaddlePaddle/PaddleSlim.git & cd PaddleSlim +python setup.py install +``` + +3.Install fd-auto-compress package +```bash +# Installing fd-auto-compress via pip +# This tool is included in the python installer of FastDeploy, so you don't need to install it again. +pip install fd-auto-compress + +# Execute in the current directory +python setup.py install +``` + +### The Usage of One-Click Model Auto Compression Tool +After the above steps are successfully installed, you can use FastDeploy one-click model automation compression tool, as shown in the following example. +```bash +fastdeploy --auto_compress --config_path=./configs/detection/yolov5s_quant.yaml --method='PTQ' --save_dir='./yolov5s_ptq_model/' +``` +For detailed documentation, please refer to [FastDeploy One-Click Model Auto Compression Tool](./auto_compression/README.md) diff --git a/tools/auto_compression/README.md b/tools/auto_compression/README.md index 4fbc8dcf0..8c1439997 100644 --- a/tools/auto_compression/README.md +++ b/tools/auto_compression/README.md @@ -18,8 +18,11 @@ python setup.py install ``` ### fastdeploy-auto-compression 一键模型自动化压缩工具安装方式 -用户在当前目录下，运行如下命令: ``` +# 通过pip安装fd-auto-compress包 +pip install fd-auto-compress + +# 并在上一层目录(非本级目录)执行如下命令 python setup.py install ``` @@ -43,12 +46,11 @@ wget https://bj.bcebos.com/paddlehub/fastdeploy/COCO_val_320.tar.gz tar -xvf COCO_val_320.tar.gz ``` -##### 2.使用fastdeploy_auto_compress命令，执行一键模型自动化压缩: +##### 2.使用fastdeploy --auto_compress命令，执行一键模型自动化压缩: 以下命令是对yolov5s模型进行量化, 用户若想量化其他模型, 替换config_path为configs文件夹下的其他模型配置文件即可. ```shell -fastdeploy_auto_compress --config_path=./configs/detection/yolov5s_quant.yaml --method='PTQ' --save_dir='./yolov5s_ptq_model/' +fastdeploy --auto_compress --config_path=./configs/detection/yolov5s_quant.yaml --method='PTQ' --save_dir='./yolov5s_ptq_model/' ``` -【说明】离线量化（训练后量化）：post-training quantization，缩写是PTQ ##### 3.参数说明 @@ -78,12 +80,12 @@ wget https://bj.bcebos.com/paddlehub/fastdeploy/COCO_train_320.tar tar -xvf COCO_train_320.tar ``` -##### 2.使用fastdeploy_auto_compress命令，执行一键模型自动化压缩: +##### 2.使用fastdeploy --auto_compress命令，执行一键模型自动化压缩: 以下命令是对yolov5s模型进行量化, 用户若想量化其他模型, 替换config_path为configs文件夹下的其他模型配置文件即可. ```shell # 执行命令默认为单卡训练，训练前请指定单卡GPU, 否则在训练过程中可能会卡住. export CUDA_VISIBLE_DEVICES=0 -fastdeploy_auto_compress --config_path=./configs/detection/yolov5s_quant.yaml --method='QAT' --save_dir='./yolov5s_qat_model/' +fastdeploy --auto_compress --config_path=./configs/detection/yolov5s_quant.yaml --method='QAT' --save_dir='./yolov5s_qat_model/' ``` ##### 3.参数说明 diff --git a/tools/auto_compression/README_EN.md b/tools/auto_compression/README_EN.md index e53a1a37a..6c591897a 100644 --- a/tools/auto_compression/README_EN.md +++ b/tools/auto_compression/README_EN.md @@ -1,7 +1,5 @@ # FastDeploy One-Click Model Auto Compression - - FastDeploy, based on PaddleSlim's Auto Compression Toolkit(ACT), provides developers with a one-click model auto compression tool that supports post-training quantization and knowledge distillation training. We take the Yolov5 series as an example to demonstrate how to install and execute FastDeploy's one-click model auto compression. @@ -24,9 +22,13 @@ python setup.py install ### Install Fastdeploy Auto Compression Toolkit -Run the following command in the current directory +Run the following command to install ``` +# Install fd-auto-compress package using pip +pip install fd-auto-compress + +# Execute the following command in the previous directory (not in the current directory) python setup.py install ``` @@ -52,12 +54,12 @@ wget https://bj.bcebos.com/paddlehub/fastdeploy/COCO_val_320.tar.gz tar -xvf COCO_val_320.tar.gz ``` -##### 2. Run fastdeploy_auto_compress command to compress the model +##### 2. Run fastdeploy --auto_compress command to compress the model The following command is to quantize the yolov5s model, if developers want to quantize other models, replace the config_path with other model configuration files in the configs folder. ```shell -fastdeploy_quant --config_path=./configs/detection/yolov5s_quant.yaml --method='PTQ' --save_dir='./yolov5s_ptq_model/' +fastdeploy --auto_compress --config_path=./configs/detection/yolov5s_quant.yaml --method='PTQ' --save_dir='./yolov5s_ptq_model/' ``` [notice] PTQ is short for post-training quantization @@ -89,14 +91,14 @@ wget https://bj.bcebos.com/paddlehub/fastdeploy/COCO_val_320.tar.gz tar -xvf COCO_val_320.tar.gz ``` -##### 2.Use fastdeploy_auto_compress command to compress models +##### 2.Use fastdeploy --auto_compress command to compress models The following command is to quantize the yolov5s model, if developers want to quantize other models, replace the config_path with other model configuration files in the configs folder. ```shell # Please specify the single card GPU before training, otherwise it may get stuck during the training process. export CUDA_VISIBLE_DEVICES=0 -fastdeploy_quant --config_path=./configs/detection/yolov5s_quant.yaml --method='QAT' --save_dir='./yolov5s_qat_model/' +fastdeploy --auto_compress --config_path=./configs/detection/yolov5s_quant.yaml --method='QAT' --save_dir='./yolov5s_qat_model/' ``` ##### 3.Parameters diff --git a/tools/auto_compression/fd_auto_compress/__init__.py b/tools/auto_compression/fd_auto_compress/__init__.py index e69de29bb..ea473ddf1 100644 --- a/tools/auto_compression/fd_auto_compress/__init__.py +++ b/tools/auto_compression/fd_auto_compress/__init__.py @@ -0,0 +1 @@ +import fd_auto_compress diff --git a/tools/auto_compression/fd_auto_compress/fd_auto_compress.py b/tools/auto_compression/fd_auto_compress/fd_auto_compress.py index 145f4a468..5f5b62798 100644 --- a/tools/auto_compression/fd_auto_compress/fd_auto_compress.py +++ b/tools/auto_compression/fd_auto_compress/fd_auto_compress.py @@ -82,13 +82,11 @@ def reader_wrapper(reader, input_list): return gen -def main(): +def auto_compress(FLAGS): + #FLAGS needs parse time_s = time.time() - paddle.enable_static() - parser = argsparser() - FLAGS = parser.parse_args() assert FLAGS.devices in ['cpu', 'gpu', 'xpu', 'npu'] paddle.set_device(FLAGS.devices) @@ -189,7 +187,3 @@ def main(): time_total = time.time() - time_s print("Finish Compression, total time used is : ", time_total, "seconds.") - - -if __name__ == '__main__': - main() diff --git a/tools/auto_compression/setup.py b/tools/auto_compression/setup.py index 9ee3ce28f..5938874bc 100644 --- a/tools/auto_compression/setup.py +++ b/tools/auto_compression/setup.py @@ -1,26 +1,25 @@ import setuptools import fd_auto_compress -long_description = "fastdeploy-auto-compression is a toolkit for model auto compression of FastDeploy.\n\n" -long_description += "Usage: fastdeploy_auto_compress --config_path=./yolov7_tiny_qat_dis.yaml --method='QAT' --save_dir='../v7_qat_outmodel/' \n" +long_description = "fd_auto_compress is a toolkit for model auto compression of FastDeploy.\n\n" +long_description += "Usage: fastdeploy --auto_compress --config_path=./yolov7_tiny_qat_dis.yaml --method='QAT' --save_dir='../v7_qat_outmodel/' \n" with open("requirements.txt") as fin: REQUIRED_PACKAGES = fin.read() setuptools.setup( - name="fastdeploy-auto-compression", # name of package + name="fd_auto_compress", # name of package description="A toolkit for model auto compression of FastDeploy.", long_description=long_description, long_description_content_type="text/plain", packages=setuptools.find_packages(), + author='fastdeploy', + author_email='fastdeploy@baidu.com', + url='https://github.com/PaddlePaddle/FastDeploy.git', install_requires=REQUIRED_PACKAGES, classifiers=[ "Programming Language :: Python :: 3", "License :: OSI Approved :: Apache Software License", "Operating System :: OS Independent", ], - license='Apache 2.0', - entry_points={ - 'console_scripts': - ['fastdeploy_auto_compress=fd_auto_compress.fd_auto_compress:main', ] - }) + license='Apache 2.0', ) diff --git a/tools/common_tools/__init__.py b/tools/common_tools/__init__.py new file mode 100644 index 000000000..e69de29bb diff --git a/tools/common_tools/common_tools.py b/tools/common_tools/common_tools.py new file mode 100644 index 000000000..04fa215f7 --- /dev/null +++ b/tools/common_tools/common_tools.py @@ -0,0 +1,51 @@ +import argparse + + +def argsparser(): + + parser = argparse.ArgumentParser(description=__doc__) + ## argumentments for auto compression + parser.add_argument('--auto_compress', default=False, action='store_true') + parser.add_argument( + '--config_path', + type=str, + default=None, + help="path of compression strategy config.", + required=True) + parser.add_argument( + '--method', + type=str, + default=None, + help="choose PTQ or QAT as quantization method", + required=True) + parser.add_argument( + '--save_dir', + type=str, + default='./output', + help="directory to save compressed model.") + parser.add_argument( + '--devices', + type=str, + default='gpu', + help="which device used to compress.") + + ## arguments for other tools + return parser + + +def main(): + + args = argsparser().parse_args() + if args.auto_compress == True: + try: + from fd_auto_compress.fd_auto_compress import auto_compress + print("Welcome to use FastDeploy Auto Compression Toolkit!") + auto_compress(args) + except ImportError: + print( + "Can not start auto compresssion successfully! Please check if you have installed it!" + ) + + +if __name__ == '__main__': + main() diff --git a/tools/setup.py b/tools/setup.py new file mode 100644 index 000000000..e0430f2df --- /dev/null +++ b/tools/setup.py @@ -0,0 +1,20 @@ +import setuptools + +long_description = "fastdeploy-tools is a toolkit for FastDeploy, including auto compression .etc.\n\n" +long_description += "Usage of auto compression: fastdeploy --auto_compress --config_path=./yolov7_tiny_qat_dis.yaml --method='QAT' --save_dir='./v7_qat_outmodel/' \n" + +setuptools.setup( + name="fastdeploy-tools", # name of package + description="A toolkit for FastDeploy.", + long_description=long_description, + long_description_content_type="text/plain", + packages=setuptools.find_packages(), + classifiers=[ + "Programming Language :: Python :: 3", + "License :: OSI Approved :: Apache Software License", + "Operating System :: OS Independent", + ], + license='Apache 2.0', + entry_points={ + 'console_scripts': ['fastdeploy = common_tools.common_tools:main', ] + })