HackerTrans
TopNewTrendsCommentsPastAskShowJobs

goldenjm

11 karmajoined 16 lat temu

comments

goldenjm
·4 dni temu·discuss
This model looks very interesting. How does it compare to heuristic-based approaches like Defuddle and Readability?
goldenjm
·4 miesiące temu·discuss
You're welcome!
goldenjm
·4 miesiące temu·discuss
Hey- I just saw your comment here 61 days ago (https://news.ycombinator.com/item?id=46560110) and can't reply there, so I'm replying here since this is your latest post.

You requested that we add inverting colors to PDF mode on Paper2Audio. I have good news: we added it (other users requested it too). It is in the main menu when you have our PDF view open. Please let me know anytime if you have any questions or feedback.

- Joe, Paper2Audio Founder
goldenjm
·6 miesięcy temu·discuss
I'm the Paper2Audio founder and I'm thrilled to see you recommending us here. Paper2Audio specializes in narrating complex documents like research papers to you. It is free for personal use.

This PDF exceeds our page limit, so you would have to split it up. We're working on increasing our page limits.
goldenjm
·8 miesięcy temu·discuss
This would be more helpful if it included DeepSeek-OCR, PaddleOCR-VL and MinerU 2.5. In general, I've found that OmniDocBench is a reliable benchmark, perhaps surprisingly because it is made by the same team as MinerU. They updated their benchmark table recently: https://github.com/opendatalab/OmniDocBench#end-to-end-evalu.... There are some other models that score above DeepSeek-OCR as well that I'm not as familiar with.
goldenjm
·8 miesięcy temu·discuss
Their argument in favor of this change seems extremely reasonable and well-explained.