<img width="1565" height="852" alt="Image" src="https://github.com/user-attachments/assets/3634d735-bf9a-4e6e-9d43-d23ba77db017" /> 我的目的:需要使用dots.ocr的表格识别能力从图片中提取表格 问题:prompt设置为dict_promptmode_to_prompt,会有很多表格被检测为image或text,无法作为表格被还原;prompt设置prompt_table_html,表格识别结果都不够准确/不够稳定。