如何更换其他检测和识别模型？ - RapidOCR 文档

[giscus[bot]]

如何更换其他检测和识别模型？ - RapidOCR 文档

Awesome OCR multiple programing languages toolkits based on ONNXRuntime, OpenVION and PaddlePaddle.

https://rapidai.github.io/RapidOCRDocs/blog/2023/10/17/%E5%A6%82%E4%BD%95%E6%9B%B4%E6%8D%A2%E5%85%B6%E4%BB%96%E6%A3%80%E6%B5%8B%E5%92%8C%E8%AF%86%E5%88%AB%E6%A8%A1%E5%9E%8B/

[xuanyucode]

您好，cnocr的2.3基础识别模型能否集成进来，ppv4的准确率还是偏低，https://github.com/breezedeus/cnocr?tab=readme-ov-file

[SWHL]

那好奇为啥不直接用CnOCR呢？我看CnOCR项目也挺好的呀

[xuanyucode]

因为您集成的表格检测很准确，就是ocr识别比较小的字会有些错误，所以要是能适配更准确的文字识别就完美了

[SWHL]

懂了，我建了一个issue来跟踪这件事，后续看看能否整合进来。#275

[pashangshangpo]

使用 PP-OCRv4_server_rec_doc_infer.onnx 后,CUP识别要几分钟,太慢了

[ybyghuber]

用自己训练的模型以及指定对应的字典文件后，报错如下
Traceback (most recent call last):
File "/workspace/RapidTable/demo.py", line 24, in
table_engine = RapidTable(input_args)
File "/workspace/RapidTable/rapid_table/main.py", line 44, in init
self.ocr_engine = self._init_ocr_engine(self.cfg.ocr_params)
File "/workspace/RapidTable/rapid_table/main.py", line 57, in init_ocr_engine
return rapidocr.RapidOCR(params=params)
File "/root/miniconda3/envs/tableocr/lib/python3.10/site-packages/rapidocr/main.py", line 39, in init
cfg = self._load_config(config_path, params)
File "/root/miniconda3/envs/tableocr/lib/python3.10/site-packages/rapidocr/main.py", line 54, in _load_config
cfg = ParseParams.update_batch(cfg, params)
File "/root/miniconda3/envs/tableocr/lib/python3.10/site-packages/rapidocr/utils/parse_parameters.py", line 48, in update_batch
if k.split(".")[1] in enum_params and not isinstance(v, Enum):
IndexError: list index out of range

[SWHL]

贴一些完整代码