PaddleOCR
PaddleOCR copied to clipboard
Awesome multilingual OCR toolkits based on PaddlePaddle (practical ultra lightweight OCR system, support 80+ languages recognition, provide data annotation and synthesis tools, support training and de...
运行PaddleOCR C++代码时,将gpu_mem参数设置成=500 1.使用GPU进行运行推理,加载图片时候,显存先占大约1.3G,然后涨到1.9个G (无论gpu_mem设置多少,都是稳定显存占用1.9G) 2.然后使用Tensorrt运行推理,显存占用大约3.6G。 请问:为什么使用Tensorrt后,显存会占用这么高? 有什么方法能降低显存占用么?
请问文档中提供的配置都是基于8卡训练的吗,ch_PP-OCRv3/ch_PP-OCRv3_rec.yml中默认的epoch_num=500,在论文中提到识别任务共用到了18.5M训练图像,请问在8卡的情况下训练耗时大约是多少天呢
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem - 系统环境/System Environment:Ubuntu 20 - 版本号/Version:Paddle:2.3 PaddleOCR:2.6 问题相关组件/Related components:python ./tools/infer_kie_token_ser_re.py - 运行指令/Command Code:python ./tools/infer_kie_token_ser_re.py -c configs/kie/vi_layoutxlm/re_vi_layoutxlm_xfund_zh.yml -o Architecture.Backbone.checkpoints=./pretrained_model/re_vi_layoutxlm_xfund_pretrained/best_accuracy Global.infer_img=./ppstructure/docs/kie/input/zh_val_42.jpg -c_ser...
I trained a recognition model in Persian. The issue is when the model predicts the output the text is reversed for some reason. Any idea how can I resolve that?...
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem - 系统环境/System Environment: **Linux** - 版本号/Version:Paddle: paddlepaddle-gpu 2.3.2 - PaddleOCR: paddleocr 2.6.0.1 - 问题相关组件/Related components: Python 3.8.13 - 运行指令/Command Code:...
1、修复bug:删除1.docx保存 2、2种类型docx保存,修改不同的命名 3、删除recovery requirement.txt的PyMupdf 4、更新requirement.txt安装版本
Hi Team, I am new to Machine learning. Request your help with how to make paddleOCR detect special characters like ←→↑↓/~. apart from English and numeric characters. Kindly help with...
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem - 系统环境/System Environment: - 版本号/Version:Paddle: PaddleOCR: 问题相关组件/Related components: - 运行指令/Command Code: - 完整报错/Complete Error Message:
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem - 系统环境/System Environment: - 版本号/Version:Paddle: PaddleOCR: 问题相关组件/Related components: - 运行指令/Command Code: - 完整报错/Complete Error Message:  ##### 如上图,【货物或应税劳务、服务名称】这一列,下面的【小米红米3时尚金色】和【移动联通电信4G手机双卡双待】这两个是分开的两行, 1. 如果是当做key-value抽取任务(SER+RE)的话,该如何处理?一个key...
- 系统环境/System Environment:Docker [2.1.3-gpu-cuda11.2-cudnn8](https://hub.docker.com/layers/paddlepaddle/paddle/2.1.3-gpu-cuda11.2-cudnn8/images/sha256-ad3b494a60a52390199902e284c1d87e40defe390ca702619bde51f5aa904364?context=explore) - 版本号/Version:Paddle:2.1.3 PaddleOCR 2.6: 问题相关组件/Related components:SVTR 直接调用[readme](https://github.com/PaddlePaddle/PaddleOCR/blob/release%2F2.6/doc/doc_ch/algorithm_rec_svtr.md#svtr%E7%AE%97%E6%B3%95%E7%AE%80%E4%BB%8B)中的提供的配置和权重对svtr_base模型进行评测时,会报找不到SVTRLabelDecode的错误。我留意到配置文件里提到这个设定只在评测时用到,请问这一部分的模块是不是还没开源?实现起来跟CTC有什么区别呢?