China
lvjianjin/crnn
一个基于TensorFlow2的CRNN项目
English | 简体中文
Introduction
A crnn project based on Tensorflow2.
Demo
digit
chinese
Features
- Refactoring.
Requirements
- Python 3.6+
- Tensorflow 2.2.0+
Recent updates
2020-10-13: The first version is online.
2021-3-6: Open source data construction tools 。
Contents
Installation
python
Install the required Python packages,
pip install -r requirements.txt
docker
Usage
train
- Pull code
git clone https://github.com/lvjianjin/crnn.git
- Prepare the data set
digit dataset
Generate simulation data through this tool, and modify the local dataset path to dataset_path in ./configs/config.py.
chinese dataset
Generate simulation data through this tool or download the following Chinese data set to the local, and modify the local data set path to dataset_path in ./configs/config.py.
链接：https://pan.baidu.com/s/1FgdITVrM_HsyNh7QSpePjw
提取码：iakr
解压密码:chineseocr
- training
执行训练代码,
python main.py -m train
test
Execute training code,
python main.py -m test
Pre-trained model
The pre-trained model will be released soon, so stay tuned.
Contact
- e-mail：jianjinlv@163.com
- QQ Group：1081332609
License
The codes are released under the MIT License.
Similar Projects
Discover other similar projects and AI models.
Handwriting OCR for Vietnamese Address using state-of-the-art CRNN model implemented with Tensorflow. This was a challenge proposed by the Cinnamon AI Marathon.
Ready-to-use OCR with 80+ supported languages and all popular writing scripts including Latin, Chinese, Arabic, Devanagari, Cyrillic and etc.
📄 Awesome OCR multiple programing languages toolkits based on ONNX Runtime, OpenVINO, MNN, PaddlePaddle, TensorRT and PyTorch.
PaddleOCR inference in PyTorch. Converted from [PaddleOCR](https://github.com/PaddlePaddle/PaddleOCR)
Lightweight CRNN for OCR (including handwritten text) with depthwise separable convolutions and spatial transformer module [keras+tf]
chineseocr lite onnx，超轻量级中文ocr demo，支持onnx推理 ( dbnet+crnn+anglenet)
Convolutional Recurrent Neural Network(CRNN) for End-to-End Text Recognition - TensorFlow 2