mariarmestre

6 ms·

1.
RAG Fails on PDF Tables(mrm1001.github.io)
4 points by mariarmestre 1y ago|0 comments
2.
OCR and RAG for Tables(mrm1001.github.io)
2 points by mariarmestre 1y ago|0 comments
3.
by mariarmestre 2y ago
Hmmm... this is not eligible for their zero data retention policy anymore. Not sure how this will go down.
4.
Agents aren't all you need(parcha.com)
2 points by mariarmestre 2y ago|0 comments
5.
Autoquizzer for Trivia Night(theregister.com)
2 points by mariarmestre 2y ago|0 comments
6.
by mariarmestre 2y ago
Are you going to keep maintaining the package?
7.
Comcast Uses LLMs to Assist Agents in Real Time(arxiv.org)
1 points by mariarmestre 2y ago|0 comments
8.
Show HN: Generate a Quiz from Any Url(huggingface.co)
2 points by mariarmestre 2y ago|0 comments
9.
by mariarmestre 2y ago
I just tried it on a AI newsletter: https://a.tldrnewsletter.com/web-version?ep=1&lc=f08d3180-da... and it turns out I'm not keeping up with AI news as well as I thought.
10.
Generate a Quiz from Any Url(github.com)
2 points by mariarmestre 2y ago|1 comments
11.
Improving retrieval with automated metadata extraction(haystack.deepset.ai)
2 points by mariarmestre 2y ago|0 comments
12.
by mariarmestre 2y ago
Sorry, this is the link: https://community.intel.com/t5/Blogs/Tech-Innovation/Cloud/T...
13.
Intel reduces latencies of chat LLM app using quantisation(community.intel.com)
5 points by mariarmestre 2y ago|5 comments
14.
by mariarmestre 3y ago
Thank you! Our goal is to make production-ready code, so we believe that good documentation and stability of the code are paramount. I'll pass the feedback along the rest of the team :)
15.
by mariarmestre 3y ago
This represents a major rewrite of the package with more powerful features than ever. We have introduced the new concept of components. Components are composable and customisable and can be connected into pipelines. Pipelines are dynamic ex
16.
Haystack 2.0 – compound LLM pipelines are here(haystack.deepset.ai)
11 points by mariarmestre 3y ago|4 comments
17.
by mariarmestre 3y ago
Thank you for this! Congrats for the beta release. I guess this is not totally battle-tested yet then?
18.
Ask HN: Has anybody gone beyond prototype stage with LLMs?
3 points by mariarmestre 3y ago|3 comments
19.
by mariarmestre 3y ago
I think a fundamental issue with search, and the reason why many companies do not invest in tuning a good search experience, is that the main metric usually is to minimise embarrassing/irrelevant results, rather than get the best possi
20.
Show HN: App to extract structured fields from text into a spreadsheet(github.com)
1 points by mariarmestre 3y ago|0 comments
21.
Show HN: Build a knowledge base from your Slack content(dataqa.ai)
3 points by mariarmestre 4y ago|2 comments
22.
Show HN: Build a knowledge base from your Slack content(dataqa.ai)
4 points by mariarmestre 4y ago|3 comments
23.
The Race to Break ChatGPT(dataqa.ai)
1 points by mariarmestre 4y ago|0 comments
24.
by mariarmestre 5y ago
The only issue here is what I mentioned in the comment above: how easy is it to read and parse the content of said website and is it legal to read the content programmatically? Do you have any website(s) in mind?
25.
by mariarmestre 5y ago
Thanks so much for your comment! You're right that this annotation tool can be used on any form of free-form documents found online. I tackled Wikipedia first because it was an obvious first choice and they have an API to read the html
26.
by mariarmestre 5y ago
Wikidata is a very complete knowledge base, but I think there is still room for a tool like this to be used on Wikipedia data. There might be missing information on Wikidata which is still found on Wikipedia (e.g. list of Bytedance products
27.
by mariarmestre 5y ago
This is to build your own knowledge base. In many cases, Wikidata might not have the data you're looking for. For example, in the tutorial I have linked, the task is to come up with all the products released by a list of companies. Tou
28.
by mariarmestre 5y ago
Hi! Thanks for your feedback. I had not come across this, but it looks quite similar :) (at least conceptually). I don't have a video, but there is a short gif in the repository. I am planning to make a video at some point though!
29.
Show HN: A labelling tool to easily extract and label Wikipedia data
116 points by mariarmestre 5y ago|13 comments