请问如果想设置成高并发应该怎么做

[sunzx8]

想让它能做并发度为20的识别，但是单纯启动好像也很慢，请问应该怎么做

[SWHL]

目前只能自己手动开启多进程实现。还要看资源。如果一个进程，CPU都满了，那就别高并发了

[sunzx8]

请问如果想在python中切换中英文模型识别需要设置什么参数 获取 Outlook for iOS<https://aka.ms/o0ukef>

…

________________________________ 发件人: SWHL ***@***.***> 发送时间: Thursday, July 11, 2024 6:09:39 PM 收件人: RapidAI/RapidOCR ***@***.***> 抄送: ***@***.*** ***@***.***>; Author ***@***.***> 主题: Re: [RapidAI/RapidOCR] 请问如果想设置成高并发应该怎么做 (Discussion #201) This is an external email. 目前只能自己手动开启多进程实现。还要看资源。如果一个进程，CPU都满了，那就别高并发了 ― Reply to this email directly, view it on GitHub<#201 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BAS3H6BJUKOSYIB54VUJF2LZLZKWHAVCNFSM6AAAAABKUVBRKOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTAMBRHE2DQOI>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[SWHL]

去查一下官方API文档

…

---- 回复的原邮件 ---- | 发件人 | ***@***.***> | | 发送日期 | 2024年07月16日 11:56 | | 收件人 | RapidAI/RapidOCR ***@***.***> | | 抄送人 | SWHL ***@***.***>, Comment ***@***.***> | | 主题 | Re: [RapidAI/RapidOCR] 请问如果想设置成高并发应该怎么做 (Discussion #201) | 请问如果想在python中切换中英文模型识别需要设置什么参数 获取 Outlook for iOS<https://aka.ms/o0ukef>

________________________________ 发件人: SWHL ***@***.***> 发送时间: Thursday, July 11, 2024 6:09:39 PM 收件人: RapidAI/RapidOCR ***@***.***> 抄送: ***@***.*** ***@***.***>; Author ***@***.***> 主题: Re: [RapidAI/RapidOCR] 请问如果想设置成高并发应该怎么做 (Discussion #201) This is an external email. 目前只能自己手动开启多进程实现。还要看资源。如果一个进程，CPU都满了，那就别高并发了 ― Reply to this email directly, view it on GitHub<#201 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BAS3H6BJUKOSYIB54VUJF2LZLZKWHAVCNFSM6AAAAABKUVBRKOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTAMBRHE2DQOI>. You are receiving this because you authored the thread.Message ID: ***@***.***> — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>

[SWHL]

从你贴的看不出来啥。这个模型我有转好的：https://huggingface.co/SWHL/RapidOCR/tree/main/PP-OCRv4

[sunzx8]

从你贴的看不出来啥。这个模型我有转好的：https://huggingface.co/SWHL/RapidOCR/tree/main/PP-OCRv4

您好，我用了您这个模型还是报错
报错代码：
ocr_engine = RapidOCR(rec_model_path='/home/ubuntu/shawn/table_recognition/YOLOv10-Document-Layout-Analysis/ch_PP-OCRv4_det_server_infer.onnx')
报错内容：
Traceback (most recent call last):
File "/home/ubuntu/shawn/table_recognition/YOLOv10-Document-Layout-Analysis/demo.py", line 19, in
ocr_engine = RapidOCR(rec_model_path='/home/ubuntu/shawn/table_recognition/YOLOv10-Document-Layout-Analysis/ch_PP-OCRv4_det_server_infer.onnx')
File "/home/ubuntu/miniconda3/envs/internvl/lib/python3.9/site-packages/rapidocr_onnxruntime/main.py", line 54, in init
self.text_rec = TextRecognizer(config["Rec"])
File "/home/ubuntu/miniconda3/envs/internvl/lib/python3.9/site-packages/rapidocr_onnxruntime/ch_ppocr_v3_rec/text_recognize.py", line 35, in init
self.postprocess_op = CTCLabelDecode(character_dict_path)
File "/home/ubuntu/miniconda3/envs/internvl/lib/python3.9/site-packages/rapidocr_onnxruntime/ch_ppocr_v3_rec/utils.py", line 12, in init
assert character_dict_path is not None, "character_dict_path should not be None"
AssertionError: character_dict_path should not be None
版本：
rapidocr-onnxruntime 1.3.22

[SWHL]

等我晚上回去看看吧

[SWHL]

首先，我直接下载我给你的链接，rec模型是没有问题的。从报错来看，你下载错了模型，应该是ch_PP-OCRv4_rec_server_infer.onnx，你下载的是检测的模型ch_PP-OCRv4_det_server_infer.onnx。改正一下，应该就没问题了。

其次，我在线转换的模型也没有问题的。转换网址：link

问题如果解决了，请顺手关闭一下Discussions哈

[sunzx8]

谢谢您

[sunzx8]

谢谢

…

________________________________ 发件人: SWHL ***@***.***> 发送时间: 2024年7月16日 21:24 收件人: RapidAI/RapidOCR ***@***.***> 抄送: ***@***.*** ***@***.***>; Author ***@***.***> 主题: Re: [RapidAI/RapidOCR] 请问如果想设置成高并发应该怎么做 (Discussion #201) This is an external email. 从你贴的看不出来啥。这个模型我有转好的：https://huggingface.co/SWHL/RapidOCR/tree/main/PP-OCRv4 ― Reply to this email directly, view it on GitHub<#201 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BAS3H6B2IZMVIK3L3DW2WSDZMUNKJAVCNFSM6AAAAABKUVBRKOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTAMBWGI3TKMQ>. You are receiving this because you authored the thread.Message ID: ***@***.***>