Re:How we made our optical character recognition (OCR) code ...
by booder1 on Thursday May 22, 2025 @06:59PM
(#44068075)
Surya weights for the models are licensed cc-by-nc-sa-4.0 so not free for commercial usage. Also, as far as I know, the training data is 100% unavailable. Given they use well trained, but standard models, it isn't really open source and barely, maybe, open weight. I kinda hate how their repo says gpl cause that is only true for the inference code. The training code is closed source.
Using Surya gets you significantly better results and makes almost all the work detailed in the article largely unnecessary.