Skip to content
Better HN
Top
Best
Ask
Show
New
Jobs
Search
⌘K
Show HN: CPU-only fast OCR for screenshots, images, PDFs, webpages
(opens in new tab)
(github.com)
9 points
mrkn1
26d ago
8 comments
Save
Share
8 comments
8 comments · 4 top-level
top
newest
oldest
quanglee
24d ago
· 1 in thread
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
mrkn1
OP
23d 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.
explosion-s
26d ago
· 1 in thread
Why is no GPU a pull here? Is there usually an activation overhead GPU request or something?
mrkn1
OP
26d ago
No, simply, my laptop only has a CPU.
arnabdey0503
26d ago
· 1 in thread
Have you done benchmark againest other methods?
mrkn1
OP
26d 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
ajithhyd
25d ago
· 1 in thread
Can it support other languages?
mrkn1
OP
25d ago
It should support 109 languages. More info here:
https://huggingface.co/PaddlePaddle/PaddleOCR-VL
j
/
k
navigate · click thread line to collapse
