HackerTrans
TopNewTrendsCommentsPastAskShowJobs

LoMoGan

no profile record

Submissions

[untitled]

1 points·by LoMoGan·14 dagen geleden·0 comments

[untitled]

1 points·by LoMoGan·3 maanden geleden·0 comments

[untitled]

1 points·by LoMoGan·3 maanden geleden·0 comments

DiffRatio: A SOTA one-step Diffusion model with 50% less GPU memory

arxiv.org
1 points·by LoMoGan·6 maanden geleden·0 comments

DiffRatio – A One-Step Diffusion Model with SOTA quality and 50% less memory

arxiv.org
4 points·by LoMoGan·6 maanden geleden·1 comments

Show HN: PageIndex Chat – AI reader for long documents

2 points·by LoMoGan·7 maanden geleden·0 comments

Show HN: ChatIndex – A Lossless Memory System for AI Agents

17 points·by LoMoGan·8 maanden geleden·5 comments

An API for Chating with Nvidia 10-Q Report

github.com
2 points·by LoMoGan·8 maanden geleden·0 comments

Show HN: ChatIndex – an open-sourced long-context managment system

github.com
4 points·by LoMoGan·8 maanden geleden·0 comments

Do we still need OCR when we can build a pure vision-based AI agent

pageindex.ai
4 points·by LoMoGan·9 maanden geleden·1 comments

comments

LoMoGan
·6 maanden geleden·discuss
This paper was posted on arXiv today. It shows surprisingly strong results on ImageNet at 512 resolution (FID 1.41) with one-step generation, while requiring 50% less training-time memory. Do you think this could become the next standard training method for image foundation models? Feel free to leave your comments.
LoMoGan
·8 maanden geleden·discuss
Yeah good point!
LoMoGan
·8 maanden geleden·discuss
Yeah lolll, this is the memory you need
LoMoGan
·9 maanden geleden·discuss
With the rise of vision-language models (VLMs) (such as Qwen-VL and GPT-4.1), new end-to-end OCR models like DeepSeek-OCR have emerged. These models jointly understand visual and textual information, enabling direct interpretation of PDFs without an explicit layout detection step.

However, this paradigm shift raises an important question:

If a VLM can already process both the document images and the query to produce an answer directly, do we still need the intermediate OCR step?
LoMoGan
·9 maanden geleden·discuss
Sounds interesting, will try it out.
LoMoGan
·9 maanden geleden·discuss
Interesting, is this based on an external Vector DB to store and process the PDF?