nickliqian/cnn_captcha
use cnn recognize captcha by tensorflow. 本项目针对字符型图片验证码,使用tensorflow实现卷积神经网络,进行验证码识别。
PythonApache-2.0
Issues
- 1
- 2
Python用哪个版本啊,安装依赖这步就跪了
#216 opened by ishenmao - 6
搞GPU环境走了两小时的坑,给后面的朋友一点建议。
#166 opened by JoJoJoinme - 0
- 0
验证码长度不是固定的,可能是5,6,7个字符项目无法兼容啊
#214 opened by z-one - 0
for model
#212 opened by eddiechina - 0
run
#213 opened by SHAOTAO1998 - 3
10万的样本跑了4万次了还是40%求问负样本具体应该怎么弄
#176 opened by Lction - 0
- 1
- 0
dockfile无法构建
#199 opened by lqszsd - 1
- 0
模型中训练和测试的时候网络的参数是不一致
#195 opened by Mailche - 5
大家一般怎么做数据标注的,如果我的训练样本有5000个图片,怎么快速标注?
#163 opened by 150109514 - 0
- 1
Performance issue in the program
#191 opened by DLPerf - 0
验证码尺寸
#190 opened by JasonDing0401 - 0
- 0
Captcha库生成的验证码最小宽度问题
#188 opened by Ambert1108 - 0
请问运行train_model.py出现以下报错原因是什么
#181 opened by Irises1 - 0
请问为什么测试的时候准确率有90%但是在recognize_local.py上识别出来的都是错的
#184 opened by guoyu1998 - 0
Assign requires shapes of both tensors to match. lhs shape= [248] rhs shape= [144]
#185 opened by mgaic - 1
websever_recognize_api文件运行后,访问url显示链接暂时无法访问
#172 opened by TT159 - 7
大佬,有tensorflow2.0的版本吗?
#160 opened by YouSmart2016 - 0
小写字母正常,大写字母错误,改了工作区间没用
#177 opened by regancz - 0
运行train_model没有输出结果
#175 opened by lushi123-art - 2
如何把训练出来的model转化为通用pb文件供其他语言使用呢,求指点
#159 opened by mooooooonkey - 5
请问大神我把图片大小改成100*200,代码怎么修改,还有增加大写字母,代码怎么修改
#150 opened by blankandwhite - 1
生成验证码训练集问题
#174 opened by bilibilipala - 1
验证码中的字符间距问题
#173 opened by ThomasSun001 - 1
配置中使用GPU训练,但是在任务管理器下GPU没有反应
#170 opened by doubleee - 1
大写字母后报错,图片格式仍为100*60,新模型训练
#161 opened by reporter-law - 1
FileNotFoundError
#169 opened by nckf - 0
- 0
Input to reshape is a tensor with 10 values, but the requested shape requires a multiple of 40
#165 opened by 85757 - 2
- 2
自己训练需要尺寸是70*30 的验证码。请告知一下怎么修改
#164 opened by xiaoliangli1128 - 1
- 1
- 1
修改两个配置文件中的图片的宽高 跑tensorflow会报错 宽高只能固定给100x60吗
#153 opened by mgaic - 0
生成图片宽高给87x26就报错 100x60和800x250就可以跑 环境ubantu16.0
#157 opened by mgaic - 0
生成图片宽高给87x26就报错 100x26和800x250就可以跑 环境ubantu16.0
#156 opened by mgaic - 0
- 0
图片宽高给87x26就报错 100x26和800x250就可以跑 环境ubantu16.04
#154 opened by mgaic - 1
更换数据集后报错:Assign requires shapes of both tensors to match. lhs shape= [4608,1024] rhs
#152 opened by JoeMY96 - 0
recognize_time_test.py运行报错
#151 opened by lonleylkk - 0
- 0
recognize_api.py 运行报错
#147 opened by 1604051738 - 0
想请问大佬,最后训练数据统计的折线图怎么生成的
#146 opened by jkhohai - 0