PaddleOCR icon indicating copy to clipboard operation
PaddleOCR copied to clipboard

为什么我用自己的数据集训练了PP-OCRv4去识别却不返回结果?

Open youde2000 opened this issue 1 year ago • 0 comments

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:Macos M1
  • 版本号/Version:Paddle: PaddleOCR:2.7 问题相关组件/Related components:
  • 运行指令/Command Code: /Users/2mythvector/miniconda3/envs/ppocr_rosetta/bin/python
    /Users/2mythvector/Downloads/PaddleOCR-release-2.6/tools/infer/predict_rec.py
    --image_dir="/Users/2mythvector/opt/oh_text_renderer/ALL/A/A_image1_7.png"
    --rec_char_dict_path="/Users/2mythvector/Downloads/PaddleOCR-release-2.6/ppocr/utils/ppocr_keys_v1.txt"
    --rec_model_dir="/Users/2mythvector/Downloads/infer"
  • 完整报错/Complete Error Message:

我们提供了AceIssueSolver来帮助你解答问题,你是否想要它来解答(请填写yes/no)?/We provide AceIssueSolver to solve issues, do you want it? (Please write yes/no):

请尽量不要包含图片在问题中/Please try to not include the image in the issue.

这是我用自己的数据集训练了PP-OCRv4的结果 image 但是我用官方提供的模型却可以,条件都一样,就是模型不一样 image

这是我的训练时的超参数,用的是官方提供的训练模型 image

youde2000 avatar Nov 16 '23 00:11 youde2000