Hacker News new | ask | show | jobs
Show HN: CPU-only fast OCR for screenshots, images, PDFs, webpages (github.com)
9 points by mrkn1 20 days ago
4 comments
quanglee 18 days ago
Do you have some benchmark on how fast it is on cpu? I checked the repo and the huggingface page but only found gpu-based benchmarks
link
mrkn1 18 days ago
I have not. My experience has been a few seconds for an 1024x1024 with medium density of text, FWIW. Feel free to try it on a few test images, model is pretty small and fast, but yeah no formal evals on CPU.
link
explosion-s 20 days ago
Why is no GPU a pull here? Is there usually an activation overhead GPU request or something?
link
mrkn1 20 days ago
No, simply, my laptop only has a CPU.
link
arnabdey0503 20 days ago
Have you done benchmark againest other methods?
link
mrkn1 20 days ago
I haven't. But the evals of the underlying model are published here, including on Omnibench. https://huggingface.co/PaddlePaddle/PaddleOCR-VL-1.5
link
ajithhyd 19 days ago
Can it support other languages?
link
mrkn1 19 days ago
It should support 109 languages. More info here: https://huggingface.co/PaddlePaddle/PaddleOCR-VL
link