News for hackers. Stuff that matters.
Re:DeepSeek OCR by jakewins on Monday October 20, 2025 @05:05AM (#45641608)
But this was very much the case with existing OCR software as well? I guess the LLMs will end up making up plausible looking text instead of text riddled with errors, which makes it much harder to catch the mistakes, in fairness
Re:DeepSeek OCR by rkagerer on Monday October 20, 2025 @06:50AM (#45642440)
Good libraries gave results with embedded confidence levels for each unit recognized.
Re:DeepSeek OCR by wahnfrieden on Monday October 20, 2025 @09:41AM (#45643820)
Existing ocr doesn’t skip over entire (legible) paragraphs or hallucinate entire sentences
Re:DeepSeek OCR by KoolKat23 on Monday October 20, 2025 @11:58AM (#45645395)
This must be some older/smaller model.
Re:DeepSeek OCR by criddell on Monday October 20, 2025 @10:29AM (#45644305)
I usually run the image(s) through more than one converter then compare the results. They all have problems, but the parts they agree on are usually correct.
Re:DeepSeek OCR by Davidzheng on Monday October 20, 2025 @09:50AM (#45643920)
rarely happens to me using LLMs to transcribe pdfs