HN user
You can do hybrid search in Postgres.
Shameless plug: https://github.com/jankovicsandras/plpgsql_bm25 BM25 search implemented in PL/pgSQL ( Unlicense / Public domain )
The repo includes also plpgsql_bm25rrf.sql : PL/pgSQL function for hybrid search ( plpgsql_bm25 + pgvector ) with Reciprocal Rank Fusion; and Jupyter notebook examples.
You can do hybrid search in Postgres.
Shameless plug: https://github.com/jankovicsandras/plpgsql_bm25 BM25 search implemented in PL/pgSQL ( Unlicense / Public domain )
The repo includes plpgsql_bm25rrf.sql : PL/pgSQL function for hybrid search ( plpgsql_bm25 + pgvector ) with Reciprocal Rank Fusion; and Jupyter notebook examples.
Looks cool!
Here's some feedback:
- the diagram is not centered and zoomed by default, this is easy to fix
- it would be great to have better animation controls, like progressbar, play half speed, etc.
- it would be great to be able to export animation to video, animgif, etc. e. g. with ffmpeg.js
Shameless plug: https://github.com/jankovicsandras/plpgsql_bm25 BM25 search implemented in PL/pgSQL ( Unlicense / Public domain )
The repo includes plpgsql_bm25rrf.sql : PL/pgSQL function for Hybrid search ( plpgsql_bm25 + pgvector ) with Reciprocal Rank Fusion; and Jupyter notebook examples.
"It's also possible to do it on top of other DBs like Postgres, but takes more effort."
Shameless plug: plpgsql_bm25: BM25 search implemented in PL/pgSQL (The Unlicense / PUBLIC DOMAIN)
https://github.com/jankovicsandras/plpgsql_bm25
There's an example Postgres_hybrid_search_RRF.ipynb in the repo which shows hybrid search with Reciprocal Rank Fusion ( plpgsql_bm25 + pgvector ).
Very cool robot dog and interesting video! Can the dog climb stairs? Isn't capstan drive temperature sensitive, e.g. the ropes will be shorter in cold and longer in warm wheather?
Can this solve vector tracing? https://github.com/jankovicsandras/imagetracerjs (Public domain) Discaimer: I made this.
Postgres is a good idea.
Shameless plug: https://github.com/jankovicsandras/plpgsql_bm25
BM25 search implemented in PL/pgSQL, there's also an example of Hybrid (BM25+pgvector) search in the repo.
This looks cool!
Shameless plug:
https://github.com/jankovicsandras/plpgsql_bm25
BM25 search implemented in PL/pgSQL, might be useful if one can't use Rust extensions with Postgres, e. g. hosted Postgres without admin rights.
I think the following works in Firefox. WARNING: this is a security risk obviously.
about:config -> security.fileuri.strict_origin_policy -> false
source: https://stackoverflow.com/questions/58067499/runing-javascri...
This is a good intro to text search. Shameless plug: If you throw in a bit more, ca. 250 SLOC, you can have BM25 search: https://github.com/jankovicsandras/bm25opt
Can you please add this shameless plug?
https://github.com/jankovicsandras/plpgsql_bm25
Opensource BM25 search in PL/pgSQL (for example where you can't use Rust extensions), and hybrid search with pgvector and Reciprocal Rank Fusion.
Shameless plug:
This is a good point and was a difficult design decision. The reasons for changing the API are:
- easier to use with untokenized corpus and questions
- to fix issues with the tokenizing ( e.g. https://github.com/dorianbrown/rank_bm25/issues/38 ); also rank_bm25 provides no default tokenizer, a naive split-on-whitespace is a wrong choice
- considerably simplify the code (way less SLOC)
- point out the similarities of the algorithms for educational purpuses / further development
In practice, the differences are minimal ( see Example 3: comparison with rank_bm25 ).
Or you can use Postgres to store the original text, the metadata and the embedding in a single table and then do hybrid queries against them (with pre- or post-filtering, optimized automatically).
Shameless plug:
BM25 search implemented in PL/pgSQL: https://github.com/jankovicsandras/plpgsql_bm25
faster BM25 search algorithms in Python: https://github.com/jankovicsandras/bm25opt
Hi, I don't know what's SOTA, but I got good results with these (open source, on-device) :
https://github.com/SYSTRAN/faster-whisper (speech-to-text) https://github.com/rhasspy/piper (text-to-speech)
Isn't parts of this quoting or paraphrasing from the book "The Tibetan Book of Living and Dying" ? https://en.m.wikipedia.org/wiki/The_Tibetan_Book_of_Living_a...
Free and open source Okapi BM25 search for Postgresql implemented in PL/pgSQL, so it works without extensions.
https://github.com/jankovicsandras/plpgsql_bm25
Might be useful in hosted/cloud Postgres, where you can't use Rust extensions.
A hybrid search approach might help, like combining vector similarity scores with e.g. BM25 scores.
Shameless plug (FOSS): https://github.com/jankovicsandras/plpgsql_bm25 Okapi BM25 search implemented in PL/pgSQL for Postgres.
"Controller and processor Data controllers must clearly disclose any data collection, declare the lawful basis and purpose for data processing, and state how long data is being retained and if it is being shared with any third parties or outside of the EEA."
I don't see the word "personal" in this sentence, only "any data collection". It's clearly not in the users' interest/benefit to activate this data collection, and not required for the normal functioning of the browser/websites. So activating this silently is minimum _very unethical_ and probably illegal, but I'm not a lawyer.
Is this illegal in GDPR countries (EU + Norway + ...) or am I reading GDPR wrong?
"If informed consent is used as the lawful basis for processing, consent must have been explicit for data collected and each purpose data is used for"
Your point 3. is true generally, but - as I understand - parent wanted to operate a customer support channel and dismissed IRC. In this case: just don't lock your channel for Guests. Timezones affect other platforms, but an IRC bot might help customers when admins sleep. But I'm not sure IRC is best for this use case (e.g. forums, wiki are probably better), but "high barrier of entry" is arguable.
"huuuuge barrier to entry for new users"
1. go to https://web.libera.chat/
2. click Start (as a Guest account)
Here's a completely made up point of view. (I haven't read the articles.)
GPT-4 is not an LLM, but a complex software system, which has LLM(s) at its core, but also other components like RAG, toxicity filter, apologizing mechanism, expert systems, etc. "GPT-4" is product name / marketing name. For OpenAI, this would be logical for performance and business reasons. This explains also how they can tune it, the apparent secrecy about the architecture, etc.
It's also logical to make small, incremental changes to this system instead of building whatever GPT-5 would mean from ground up. So I expect "GPT-5" is also just a marketing name for a slightly better black-box (for us) system and product line.
This looks cool!
There's a small ironically funny typo in the first line: knolwedge
One
Another point of view:
https://medium.com/insurge-intelligence/how-the-cia-made-goo...
https://medium.com/insurge-intelligence/why-google-made-the-...
(I don't know how truthful these articles are, but found them interesting to read.)
This. Git is not an Edit history. Another surprising (at least to me) thing is that you can't add empty directories (without placeholder files) : https://stackoverflow.com/questions/115983/how-do-i-add-an-e...
Interesting that Llama.cpp, Ollama or Whisper is not even mentioned.