中文文字检测及识别（ORC）

最新推荐文章于 2024-05-12 21:26:51 发布

hailongzhang26

最新推荐文章于 2024-05-12 21:26:51 发布

阅读量3.7k

点赞数 2

分类专栏：预审文件的识别文章标签： ocr 边缘检测图像识别人工智能

版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。

本文链接：https://blog.csdn.net/hailongzhang26/article/details/104638435

版权

预审文件的识别专栏收录该内容

1 篇文章 0 订阅

订阅专栏

中文文字检测及识别（ORC）

https://github.com/471417367/chinese_ocr_api

首先基于CTPN检测到文字（可以是中英文以及数字），然后基于RCNN进行文字识别。

关注

2
点赞
踩
11

收藏

觉得还不错? 一键收藏
0
评论
中文文字检测及识别（ORC）

中文文字检测及识别（ORC）https://github.com/471417367/chinese_ocr_api首先基于CTPN检测到文字（可以是中英文以及数字），然后基于RCNN进行文字识别。
复制链接

扫一扫

专栏目录

hailongzhang26 CSDN认证博客专家 CSDN认证企业博客

码龄5年

6: 原创

112万+: 周排名

171万+: 总排名

8089: 访问

: 等级

153: 积分

7: 粉丝

12: 获赞

16: 评论

27: 收藏

私信

关注

热门文章

分类专栏

最新评论

bert4keras模型转onnx
源代码杀手: 记录一下： python -m tf2onnx.convert --saved-model .\model_path\generate_model_tf --output generate_simbert.onnx --opset 13 2022-12-31 14:28:58.820069: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library cudart64_110.dll D:\Anaconda3\envs\py36\lib\runpy.py:125: RuntimeWarning: 'tf2onnx.convert' found in sys.modules after import of package 'tf2onnx', but prior to execution of 'tf2onnx.convert'; this may result in unpredictable behaviour warn(RuntimeWarning(msg)) 2022-12-31 14:29:00.986324: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library nvcuda.dll 2022-12-31 14:29:01.016570: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1733] Found device 0 with properties: pciBusID: 0000:01:00.0 name: NVIDIA GeForce GTX 1650 Ti computeCapability: 7.5 coreClock: 1.485GHz coreCount: 16 deviceMemorySize: 4.00GiB deviceMemoryBandwidth: 178.84GiB/s 2022-12-31 14:29:01.016850: I tensorflow/stream_executor
bert4keras模型转onnx
源代码杀手: 2022-12-31 14:29:36.267868: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:1144] Optimization results for grappler item: graph_to_optimize constant_folding: Graph size after: 5109 nodes (-691), 7878 edges (-766), time = 9720.63477ms. function_optimizer: function_optimizer did nothing. time = 52.447ms. constant_folding: Graph size after: 5109 nodes (0), 7878 edges (0), time = 1114.88696ms. function_optimizer: function_optimizer did nothing. time = 67.862ms. 2022-12-31 14:29:41,076 - INFO - Using tensorflow=2.5.0, onnx=1.9.0, tf2onnx=1.9.1/8e8c23 2022-12-31 14:29:41,076 - INFO - Using opset <onnx, 13> 2022-12-31 14:29:55,495 - INFO - Computed 0 values for constant folding 2022-12-31 14:30:44,805 - INFO - Optimizing ONNX model 2022-12-31 14:31:23,527 - INFO - After optimization: Cast -256 (577->321), Concat -168 (340->172), Const -3968 (4224->256), Gather -49 (148->99), GlobalAveragePool +52 (0->52), GlobalMaxPool +1 (0->1), Identity -169 (169->0), Reduce Max -1 (1->0),
bert4keras模型转onnx
源代码杀手: [code=objc] python -m tf2onnx.convert --saved-model .\model_path\generate_model_tf --output generate_simbert.onnx --opset 13 2022-12-31 14:28:58.820069: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library cudart64_110.dll D:\Anaconda3\envs\py36\lib\runpy.py:125: RuntimeWarning: 'tf2onnx.convert' found in sys.modules after import of package 'tf2onnx', but prior to execution of 'tf2onnx.convert'; this may result in unpredictable behaviour warn(RuntimeWarning(msg)) 2022-12-31 14:29:00.986324: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library nvcuda.dll 2022-12-31 14:29:01.016570: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1733] Found device 0 with properties: pciBusID: 0000:01:00.0 name: NVIDIA GeForce GTX 1650 Ti computeCapability: 7.5 coreClock: 1.485GHz coreCount: 16 deviceMemorySize: 4.00GiB deviceMemoryBandwidth: 178.84GiB/s 2022-12-31 14:29:01.016850: I tensorflow/st [/code]
bert4keras模型转onnx
源代码杀手: 转完后模型反而变大，不转也罢
bert4keras模型转onnx
hailongzhang26: bert_input_process是bert源码里面整理出来的，tokenization源码里面有这个文件，bert4keras里面的tokenization作者有优化，可以看看他的源码。都不难。

最新文章

目录

评论

被折叠的条评论为什么被折叠?

到【灌水乐园】发言

查看更多评论

添加红包

成就一亿技术人!

hope_wisdom

发出的红包

实付元

使用余额支付

点击重新获取

扫码支付

钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。