Issues
- 0
bn层 lr_mult:0
#145 opened by yu1206 - 2
求更新网盘链接
#144 opened by chungexintudi - 3
训练数据集图片数量比train.txt里面的标签数量多了1个。
#136 opened by jjz-learning - 0
百度链接没一个能用的
#143 opened by vector2myth - 0
中文数据集链接失效,能否更新下?
#142 opened by ViviKing414 - 0
SynthText Data
#141 opened by laofeiwei - 2
从词典随机生成数据训练不收敛
#117 opened by wenlive - 2
合成的中文数据百度盘已的失效
#125 opened by waji3000 - 0
编译报错
#140 opened by ctgushiwei - 0
合作交流
#139 opened by mymutouren - 0
- 17
这么多bug的项目竟然还发在这里到处吹
#109 opened by mengpengfei - 0
c++调用caffemodel问题
#137 opened by xiaotaia - 4
loss 一直在50,降不下去了,是数据少吗,有30000条数据。
#102 opened by shaoshuaiche - 5
- 0
opencv加载caffe
#135 opened by amandazw - 0
数据问题一直不行,求解答
#134 opened by xiaotaia - 0
ctcloss = 0 (* 1 = 0 loss)
#133 opened by Ryansanity - 5
字符长度变化较大 ,在10-50长度,这个数据怎么训练
#107 opened by shaoshuaiche - 5
大佬好!关于 train.prototxt 的 label
#131 opened by EdwardVincentMa - 3
- 0
如何添加未定义的新层?
#132 opened by EdwardVincentMa - 0
- 0
丢字符的问题如何修改网络?
#129 opened by ManiroXi - 0
请问编译需要使用哪个版本的protobuf?感谢回答
#128 opened by vincezengqiang - 2
请问用vs2015打开工程加载caffe失败,是哪里出错了吗
#115 opened by duducode - 2
densenet-no-blstm-vertical-feature
#111 opened by stephonjou - 0
请问有人试过用efficientnet网络来做文字识别的么?
#127 opened by pingzi5233 - 0
含densenet的模型可以转换成tensorflow版吗?
#126 opened by doublegssc - 2
ocr_test 选择CPU_ONLY时,找不到libClassificationCPU-MKL.lib
#112 opened by rpf1019 - 0
准确率是单行全对的准确率?
#124 opened by lightning0016 - 0
为何要更换为junhyukoh实现的lstm版本
#123 opened by zhangsong8 - 0
为什么测试结果表明resnet18比resnet50效果好
#122 opened by mengpengfei - 2
为什么用双层blstm,单层blstm不会收敛?
#121 opened by mengpengfei - 0
- 1
您好,我配置标签时这样的
#118 opened by AnMoran - 5
我创建了普通的densenet 网络,怎么训练效果并不好,你帮我看看,大仙
#114 opened by sunjunlishi - 0
data_layer.cpp has no member named ‘label_size’
#116 opened by zzcqinag - 15
DenseBlockLayer的 openblas实现
#110 opened by sunjunlishi - 0
使用cudnn10.1版本出现问题
#113 opened by zhengqicl - 1
你好作者,请问均值你是怎么算出来的,
#106 opened by yangninghua - 2
- 5
请教作者,关于opencv调用densenet模型
#108 opened by teresasun521 - 0
能否支持多卡训练?
#103 opened by Cherry2410 - 0
作者究竟怎么测的,我测试的准确度和速度都比作者表现好得多
#101 opened by mengpengfei - 0
这个怎么那么多bug啊
#105 opened by mengpengfei - 0
选择哪块显卡运行ocr_test
#104 opened by HouBiaoLiu - 0
支不支持cudnn7,cudnn5比7慢了将近两倍
#100 opened by mengpengfei - 0
英文数据集训练之后能识别验证码吗?
#97 opened by mengpengfei - 0
无法屏蔽glog初始化输出的问题
#99 opened by ZipengFeng