Alpha

Results 8 comments of Alpha

How did you succeed later? I also tried using gpu to speed up training,I also failed!! ![wrong](https://user-images.githubusercontent.com/25379601/34646376-040998ae-f3a1-11e7-9c63-60af43f5adb0.png) my gpu driver ![driver](https://user-images.githubusercontent.com/25379601/34646379-147ca17c-f3a1-11e7-880b-cdbd2cd2b057.png) and my cuda nvcc: NVIDIA (R) Cuda compiler driver...

@Simon-Duan 你是怎样训练以及测试的?我很迷,希望能获得你的帮助

你们加载模型来测试图片是用的哪个脚本啊?caffe test 还是./ocr_test??识别图片文字呢cpp_classfication这个脚本吗?

按照warp-ctc github上的教程安装后,去你编译好的warp-ctc项目下的build文件下找到libwarpctc.so,复制到/usr/lib下即可,然后重新编译此项目即可。祝你好运!

作者的目的只是为了复现cnn+blstm+ctc算法,实验相关网络结构,参数调优及训练集生成,并不是为了做图片识别,所以对识别图像要求比较高,只能单行文字。

you need install OpenBlas , I did it successfully by this way,hope it helps https://www.cnblogs.com/llxrl/p/5292119.html

我当时在docker-nvidia镜像里跑通了这个项目,没有出现上述错误,建议你也这么做。我后来用tensorflow自己动手实现了cnn+lstm+ctc来实现ocr,用的这个项目的训练集训练,感觉也还凑合。

I did something like this on my win10 and succeed,i guess it may be effective on linux, good luck! vim /etc/profile export TESSDATA_PREFIX="/usr/local/Cellar/tesseract/3.05.01/share/" //Add at the end of etc/profile source...