I have posted issues in tesseract also you can look here.
Environment
- Tesseract Version: tesseract 5.0.0-alpha-20210401-94-ga968
- Platform: ubuntu 18.04 colab
I have to reduce my trained model time as per the default tesseract model.
I have one more doubt.
The following image described the speed when we use the same image with the default model it takes 3.59 sec, with the trained model 4.82 sec, and the default + trained model 4.79 sec.
we can see default + trained take less time than the only trained model is used.
Image resolution is here 1700*2200 .
The following image described the speed when we use the same image with the default model it takes 1.88 sec, with the trained model 2.57 sec, and the default + trained model 3.88 sec.
we can see default + trained take more time here than the only trained model is used.
Image resolution is here 770*1024.
Can you please give any suggestion how can i decrease the time? I am already using int converted model.