Open xiongwenliu opened 4 years ago
@Shreeshrii Hello, is there any good solution
我遇到了同样的问题 请问你解决了吗
我遇到了同样的问题 请问你解决了吗
Translation: "I encountered the same problem. Did you solve it?"
@xiongwenliu Sorry, without seeing the files you are using, it is difficult to guess what's the problem.
How many images are you training with? How many lines of data?
I encountered the same problem. Did you solve it?
operating system:windows10/x64 Testeract version:v5.0.0.20190623 Training font:Trained models with support for legacy and LSTM OCR engine My steps:1、Using jtessboxeditor to generate TIF 2、tesseract a2.test.exp0.tif a2.test.exp0 -l chi_sim --psm 6 lstmbox 3、Using jtessboxeditor to correct characters 4、tesseract a2.test.exp0.tif a2.test.exp0 -l chi_sim --psm 6 lstm.train 5、generate a2.training_files.txt 6、combine_tessdata -e chi_sim.traineddata chi_sim.lstm 7、lstmtraining --model_output="D:\Users\Welcome\Desktop\ocr\xl" --continue_from="D:\Users\Welcome\Desktop\ocr\xl\chi_sim.lstm" --train_listfile="D:\Users\Welcome\Desktop\ocr\xl\a2.training_files.txt" --traineddata="D:\Users\Welcome\Desktop\ocr\xl\chi_sim.traineddata" --debug_interval -1 --target_error_rate 5 8、lstmtraining --stop_training --continue_from="D:\Users\Welcome\Desktop\ocr\xl\output_checkpoint" --traineddata="D:\Users\Welcome\Desktop\ocr\xl\chi_sim.traineddata" --model_output="D:\Users\Welcome\Desktop\ocr\xl\my_chi_sim.traineddata"
results:
And when I use the best traineddata,No problem,but It does not support tes4j-3.4.8 in Java