How to implement pytesseract code with opencl to make it run on GPU?

Question

I ran OCR on a document using pytesseract . Results were very good but it took a very long time.

Now I would like to run that code on GPU. The official tesseract documentation says that we can run code on GPU using opencl but I haven't found any implementation till now. Can anyone please explain how should I proceed?

Answer 1

Where the official tesseract documentation says that we can run code on GPU using opencl ? There is experimental and unsupported implementation of opencl for small part of code, but it does not do any significant improvement.

If you are interested in speed improvement, that have a look at possibilities to run tesseract faster + use tesseract wrapper (or write your own - see SimpleTesseractPythonWrapper ) that use its API instead of wrapping tesseract executable (eg pytesseract).