PaddleOCR推理的pytorch实现和模型转换

Overview

PaddleOCR2Pytorch

简介

”真·白嫖“PaddleOCR

注意

PytorchOCRPaddleOCR-2.0rc1+动态图版本移植。

特性

高质量推理模型,准确的识别效果

  • 超轻量ptocr_mobile移动端系列
  • 通用ptocr_server系列
  • 支持中英文数字组合识别、竖排文本识别、长文本识别

模型列表

.pth模型下载链接:https://pan.baidu.com/s/1r1DELT8BlgxeOP2RqREJEg 提取码:6clx

检测模型

模型名称 PaddleOCR对应模型
ch_ptocr_mobile_v2.0_det_infer.pth ch_ppocr_mobile_v2.0_det: 推理模型 / 训练模型
ch_ptocr_server_v2.0_det_infer.pth ch_ppocr_server_v2.0_det: 推理模型 / 训练模型

识别模型

模型名称 PaddleOCR对应模型
ch_ptocr_mobile_v2.0_rec_infer.pth ch_ppocr_mobile_v2.0_rec: 推理模型 / 训练模型
ch_ptocr_server_v2.0_rec_infer.pth ch_ppocr_server_v2.0_rec: 推理模型 / 训练模型

分类模型

模型名称 PaddleOCR对应模型
ch_ptocr_mobile_v2.0_cls_infer.pth 推理模型 / 训练模型

效果展示

  • 中文模型

Pipeline

PP-OCR是一个实用的超轻量OCR系统。主要由DB文本检测、检测框矫正和CRNN文本识别三部分组成。该系统从骨干网络选择和调整、预测头部的设计、数据增强、学习率变换策略、正则化参数选择、预训练模型使用以及模型自动裁剪量化8个方面,采用19个有效策略,对各个模块的模型进行效果调优和瘦身,最终得到整体大小为3.5M的超轻量中英文OCR和2.8M的英文数字OCR。更多细节请参考PP-OCR技术方案 https://arxiv.org/abs/2009.09941 。其中FPGM裁剪器和PACT量化的实现可以参考PaddleSlim

推理

环境

shapely
numpy
pillow
pyclipper
opencv-python <= 4.2.0.32
pytorch

中文检测模型推理

python3 ./tools/infer/predict_det.py --image_dir ./doc/imgs --model_path your_det_pth_path.pth

中文识别模型推理

python3 ./tools/infer/predict_rec.py --image_dir ./doc/imgs_words --model_path your_rec_pth_path.pth

中文方向分类模型推理

python3 ./tools/infer/predict_cls.py --image_dir ./doc/imgs_words --model_path your_cls_pth_path.pth

文本检测、方向分类和文字识别串联推理

使用方向分类器

python3 ./tools/infer/predict_system.py --image_dir ./doc/imgs --det_model_path your_det_pth_path.pth --rec_model_path your_rec_pth_path.pth --use_angle_cls --cls_model_path your_cls_pth_path.pth

不适用方向分类器

python3 ./tools/infer/predict_system.py --image_dir ./doc/imgs --det_model_path your_det_pth_path.pth --rec_model_path your_rec_pth_path.pth

PaddleOCR2Pytorch

环境

shapely
numpy
pillow
pyclipper
opencv-python <= 4.2.0.32
pytorch
paddlepaddle==2.0rc1

模型转换

转换模型使用PaddleOCR的训练模型

模型路径详见PaddleOCR对应模型或者百度网盘链接https://pan.baidu.com/s/1getAprT2l_JqwhjwML0g9g 提取码:lmv7

python3 ./converter/ch_ppocr_mobile_v2.0_det_converter.py --src_model_path paddle_ch_ppocr_mobile_v2.0_det_train_dir

python3 ./converter/ch_ppocr_server_v2.0_det_converter.py --src_model_path paddle_ch_ppocr_server_v2.0_det_train_dir

python3 ./converter/ch_ppocr_mobile_v2.0_rec_converter.py --src_model_path paddle_ch_ppocr_mobile_v2.0_rec_train_dir

python3 ./converter/ch_ppocr_server_v2.0_rec_converter.py --src_model_path paddle_ch_ppocr_server_v2.0_rec_train_dir

python3 ./converter/ch_ppocr_mobile_v2.0_cls_converter.py --src_model_path paddle_ch_ppocr_mobile_v2.0_cls_train_dir

FAQ

参考

domhttpx is a google search engine dorker with HTTP toolkit built with python, can make it easier for you to find many URLs/IPs at once with fast time.

domhttpx is a google search engine dorker with HTTP toolkit built with python, can make it easier for you to find many URLs/IPs at once with fast time

Naufal Ardhani 59 Dec 04, 2022
Python client for Invidious' JSON API

Python project template A template for new Python projects. Features Automatically builds PDoc documentation & uploads package to PyPI on new GitHub r

Kevo 2 Jun 05, 2022
NekoRobot-2 - Neko is An Anime themed advance Telegram group management bot.

NekoRobot A modular telegram Python bot running on python3 with an sqlalchemy, mongodb database. ╒═══「 Status 」 Maintained Support Group Included Free

Lovely Boy 19 Nov 12, 2022
A badge generator service to count visitors of your markdown file.

Github Visitors Badge A badge generator service to count visitors of your markdown file. Hello every one! In this post, I will tell you the story of m

Kɪꜱᴀʀᴀ Pᴇꜱᴀɴᴊɪᴛʜ Pᴇʀᴇʀᴀ 〄 1 Feb 06, 2022
A python Discord wrapper made in well, python.

discord.why A python Discord wrapper made in well, python. Made to be used by devs who want something a bit more, general. Basic Examples Sending a me

HellSec 6 Mar 26, 2022
Telephus is a connection pooled, low-level client API for Cassandra in Twisted python.

Telephus Son of Heracles who loved Cassandra. He went a little crazy, at one point. One might almost say he was twisted. Description Telephus is a con

Brandon Williams 93 Apr 29, 2021
Python API for British Geological Survey magnetic field calculator

Magnetic field calculator Python API for British Geological Survey magnetic field calculator. Description This project magnetic field calculator. It u

Filip Š 3 Mar 11, 2022
A Discord Token Spammer, multi webhooks compatibility, made in python +3.7. By Ezermoz

DiscordWebhookSpammer A Discord Token Spammer, multi webhooks compatibility, made in python +3.7. By Ezermoz Put you webhook in webhooks.txt if you wa

3 Nov 24, 2021
This project is based on discord.py and is meant to be a 'Quick Start Bot' to cut down on the time it takes to write complex discord bots.

This project is based on discord.py and is meant to be a 'Quick Start Bot' to cut down on the time it takes to write complex discord bots.

Alec Ibarra 1 Mar 03, 2022
The gPodder podcast client.

___ _ _ ____ __ _| _ \___ __| |__| |___ _ _ |__ / / _` | _/ _ \/ _` / _` / -_) '_| |_ \ \__, |_| \___/\__,_\__,_\___|_| |_

gPodder and related projects 1.1k Jan 04, 2023
This is an Advanced Calculator maybe with Discord Buttons in python.

Welcome! This is an Advanced Calculator maybe with Discord Buttons in python. This was the first version of the calculator, made for my discord bot, P

Polsulpicien 18 Dec 24, 2022
Notion4ever - Python tool for export all your content of Notion page using official Notion API

NOTION4EVER Notion4ever is a small python tool that allows you to free your cont

50 Dec 30, 2022
You can share your Chegg account for answers using this bot with your friends without getting your account blocked/flagged

Chegg-Answer-Bot You can share your Chegg account for answers using this bot with your friends without getting your account blocked/flagged Reuirement

Ammey Saini 27 Dec 24, 2022
Automatically mass follows tons of NameMC profiles.

Automatically mass follows tons of NameMC profiles. (Creates REAL traffic to your profile)

Jam 3 Jun 29, 2022
Wordnik Python public library

Python 2.7 client for Wordnik.com API Overview This is a Python 2.7 client for the Wordnik.com v4 API. For more information, see http://developer.word

Wordnik 224 Dec 29, 2022
A Python app which retrieves the rank and players' equipped skins during a match

VALORANT rank yoinker About The Project Usage Contributing Contact Acknowledgements Disclaimer About The Project Their Queue Current Skin Current Rank

Isaac Kenyon 270 Jan 04, 2023
This tool is created by Shahzain and is one of the best self bots out there!

Shahzain SelfBot This tool is created by Shahzain and is one of the best self bots out there! Features Token Destroyer! Server Nuker(50-100 Bans Per S

Shahzain 6 Apr 02, 2022
AWS EC2 S3 Automated With python

AWS_EC2_S3_Automated Description This programme is a Python3 script that utilizes Boto3 to automate the process of creating an AWS EC2 instance with a

niall_crowe 2 Nov 16, 2021
Python based Algo trading bot for Nifty / Banknifty futures and options

Fully automated Alice Blue Algo Trading with Python on NSE and MCX for Nifty / Crude / Banknifty futures and options , absolutely FREE ! This algo tra

Rajesh Sivadasan 49 Dec 31, 2022
Deleting someone else's Instagram account, repeat until the target account is blocked.

Program Features 📌 Instagram report V4. 📌 Coded with the latest version of Python. 📌 Has automatic scheduling. 📌 Full account report. 📌 Report a

hack4lx 16 Oct 25, 2022