PaddleOCR icon indicating copy to clipboard operation
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...

Results 1088 PaddleOCR issues
Sort by recently updated
recently updated
newest added

- 系统环境/System Environment:Linux - 版本号/Version:Paddle-2.7 - 运行指令/Command Code: 可正常执行: paddleocr --image_dir ./imgs/11.jpg --use_angle_cls true --use_gpu false 改成了: paddleocr --image_dir ./imgs/11.jpg --use_angle_cls true 或者 paddleocr --image_dir ./imgs/11.jpg --use_angle_cls true --use_gpu true...

documentation

请提供下述完整信息以便快速定位问题: 使用官方例子,CPU模式能正常识别出文字,使用gpu模式没有报错,但文字识别率为零,无法识别。 - 系统环境/win10,cuda11.7,cuddn8.9.4:显卡:960m - 版本号/Version:Paddle:gpu版本2.5 PaddleOCR: 2.7,系统自动使用pp—ocrv4模型 - 运行指令/Command Code: - img="./data/11.jpg" - res=OCR.ocr(img,cls=trur) - print(res) - 完整报错/Complete Error Message: - ppocr debug:rec_res num:0,elapsed:0.0 我们提供了AceIssueSolver来帮助你解答问题,你是否想要它来解答:yes 请尽量不要包含图片在问题中/Please try to not...

bug

Your library requires super-old libssl1_1 dependency. Please, get rid of it, the lib is already at 3.x versions!

bug

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:centos7.6 - 版本号/Version:Paddle:2.5.1.post112 PaddleOCR:2.7.0.3 问题相关组件/Related components:ppocrV4_server - 运行指令/Command Code: - paddle_ocr = PaddleOCR(use_angle_cls=True, use_gpu=True, use_space_char=True, lang="ch", det_model_dir='paddleocr_model/ch_PP-OCRv4_det_server_infer/', rec_model_dir='paddleocr_model/ch_PP-OCRv4_rec_server_infer/',...

bug

首先paddle大大们带来惊人提升的ppocrv4系列,以及开源了大模型! 但是在Model convert时出现一些问题: ppocrv4 检测模型转成tensorrt,FP32 结果正常,FP16结果不对 - System Environment:Ubuntu 20.04.3 LTS (Focal Fossa)" - Version: - Paddle:2.4.2 - PaddleOCR:2.6.1.3 - 问题相关组件/Related components: - paddle2onnx==1.0.9 - trtexec:8.0.3-1+cuda11.3 - Command Code:onnx2tensorrt We...

TensorRT

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 - 版本号/Version:Paddle: PaddleOCR:2.7.0.3,paddlepaddle-gpu:2.5.1, CUDA:11.2.142,问题相关组件/Related components:PaddleOCR - 运行指令/Command Code:paddleocr --image_dir=2.jpg --type=structure use_gpu=True - 完整报错/Complete Error Message:无报错,但是识别结果为空;我是用use_gpu=False时,有识别结果,但是use_gpu=True时,结果为空

bug

这是一个有关PaddleOCR.page_num的问题,通过阅读源码,我的理解是:这个page_num是初始化PaddleOCR时指定的页码选项,比如说page_num设置为2,就只会识别前两幅图片的内容。 我观察了下源码里page_num没有重新归零的设定,这样的话,就会导致我在复用PaddleOCR对象时出现问题,比如说我第一次.ocr()时传进来一个2页的PDF,或者2张图片,此时page_num会被赋值为2,那么如果第二次传进来一个3页的PDF,或者3张图片的话,这个page_num就会生效,ocr()函数只会识别前2张图片,请问我的理解对吗? 相关代码: https://github.com/PaddlePaddle/PaddleOCR/blob/5ce67ec2e7cc5436be8cfb7b983cf88e9521644c/paddleocr.py#L650 ``` python if isinstance(img, list): if self.page_num > len(img) or self.page_num == 0: self.page_num = len(img) imgs = img[:self.page_num] else: imgs = [img] ```

bug

![image](https://github.com/PaddlePaddle/PaddleOCR/assets/31125043/4104c02a-b6b2-4aa3-9178-3bdbdb10270c) 如图我paddleocr 2.7 paddlepaddle 2.4.2 cuda 11.6官网推荐的一套版本环境 推理的时候初始化ocr有这个报错,初始化用了178秒,谁知道怎么解决嘛?

TensorRT

System Environment:Windows 11 Versions: Paddle:paddlepaddle-gpu 2.5.0 PaddleOCR: 2.6 tensorrt:8.6.1 i am runing tools.infer.predict_system as module, one time with the flag "--use_tensorrt" as false and one time with it as true....

TensorRT

To avoid re-downloading these files each time a new container is created, is it possible to pre-download the files and store them in a local directory? This way, when a...

feature request