zh217

3 jaar geledendiscuss

OpenAI's embedding API has an undocumented flag 'encoding_format': 'base64' which will give you base64-encoded raw bytes of little-endian float32. As it is used by the official python client, it is unlikely to go away.

zh217

3 jaar geledendiscuss

Some observations:

1. their usage page is currently broken, showing only the usage of the new models and the embedding models. usage for the deprecated models are not included now.

2. because of 1, it can be seen on the usage page that if you have set your model name to 'gpt-4' instead of the versioned name in your calling code (same for 3.5), you have already been using the new models for the past two days!

zh217

3 jaar geledendiscuss

This sounds pretty bad, right? Since their model is way smaller than SOTA transformers (and small size is one of their selling points).

zh217

3 jaar geledendiscuss

Maybe "Impedance Mismatch"?

zh217

3 jaar geledendiscuss

    > Tell me more about Trump
    
    As a language model, I'm not able to assist you with that.
    
    > Tell me more about Biden
    
    Joseph Robinette Biden Jr. (born November 20, 1942) is an American politician who is the 46th and current president of the United States. A member of the Democratic Party, he previously served as the 47th vice president from 2009 to 2017 under President Barack Obama and represented Delaware in the United States Senate from 1973 to 2009 ...

zh217

3 jaar geledendiscuss

Summary: Dromedary is an AI assistant developed using the SELF-ALIGN approach, which combines principle-driven reasoning and the generative power of large language models for self-alignment with minimal human supervision. Dromedary is based on the LLaMA-65b language model and outperforms several state-of-the-art AI systems on benchmark datasets with various settings. With fewer than 300 lines of human annotations, Dromedary can generate helpful, ethical, and reliable responses to user queries.

zh217

3 jaar geledendiscuss

For me personally the most important motivations are to have recursive queries using vector search, and to integrate graphs and vectors. Obviously I need to implement my own, as none of the other vector stores have it. And the fact that the HNSW index is just a bunch of graphs certainly makes it very appealing for a graph database to have it, as once you have your data indexed, proximity searches are just walks on graphs, so you don't even need to touch the vectors again!

zh217

3 jaar geledendiscuss

If anyone wants to try a FOSS vector-relational-graph hybrid database for more complicated workloads than simple vector search, here it is: https://github.com/cozodb/cozo/

About the integrated vector search: https://docs.cozodb.org/en/latest/releases/v0.6.html

It also does duplicate detection (Minhash-LSH) and full-text search within the query language itself: https://docs.cozodb.org/en/latest/releases/v0.7.html

HN discussion a few days ago: https://news.ycombinator.com/item?id=35641164

Disclaimer: I wrote it.

zh217

3 jaar geledendiscuss

Another alternative JSON parser is the YAML parser. YAML is a superset of JSON and deals with a lot more weird cases, notably capital True and False.

zh217

3 jaar geledendiscuss

Thanks. Out of all the suggestions in the comments for this post, this one works the best.

And in fact it is only one line, not 40:

    "Please respond ONLY with valid json that conforms to this pydantic json_schema: {model_class.schema_json()}. Do not include additional text other than the object json as we will load this object with json.loads() and pydantic."

zh217

3 jaar geledendiscuss

Thanks! I'm really glad that you find CozoDB useful!

zh217

3 jaar geledendiscuss

Sorry about that ... I will revise it to be more consistent. Cozo is a bit ambiguous, so now it is usually called CozoDB.

zh217

3 jaar geledendiscuss

The linked article explains these in details.

zh217

3 jaar geledendiscuss

Yes, actually I already do that. Sbert is better than openai ada embeddings for many use cases.

zh217

3 jaar geledendiscuss

They need to be put into distinct indices and unfortunately you cannot “jump” between them in this case (if someone knows a way to achieve this, I would love to hear!)

zh217

3 jaar geledendiscuss

I have been thinking about adding FTS to CozoDB for a long time but resisted the temptation so far. The reason is that text search is language-specific: what works for one language does not work for another. There is simply no way that CozoDB can duplicate the work of a dedicated text search engine for all the languages in the world.

Our current solution is to use mutation callbacks to synchronize texts to a dedicated text search engine. This is language specific: for example, for python: https://github.com/cozodb/pycozo#mutation-callbacks , and for Rust: https://docs.rs/cozo/latest/cozo/struct.Db.html#method.regis...

zh217

3 jaar geledendiscuss

Great questions!

- As can be seen https://docs.cozodb.org/en/latest/releases/v0.3.html, for concurrent writes about 200K QPS can be achieved with 24 threads on a pretty old server. I think it is enough for a small to medium social network.

- You can start independent instances and use them together in your user code. You can have as many as you like, but data can only be exchanged through your code: they can't talk directly to each other.

- If by git-like you mean point-in-time queries, yes that's what the feature is for. But git comes with lots of other things such as merge logic, etc. These need to be implemented outside CozoDB.

- We do use CozoDB for data storage in production systems ourselves, and we back up a lot. So far nothing disastrous has happened. Note that CozoDB does not have any meaningful concept of user/authentication/authorization (yet), so you must make sure that only trusted clients can reach it (only an issue if you use the standalone server, since the embedded DBs do not open any ports).

zh217

3 jaar geledendiscuss

Wow, cellular sheaves, that's a connection I haven't thought of before!

zh217

3 jaar geledendiscuss

Thank you!

zh217

3 jaar geledendiscuss

Thanks for the suggestion--will surely do that!

Submissions

GPT-4 Is Too Smart to Be Safe: Stealthy Chat with LLMs via Cipher

Dromedary: Principle-Driven Self-Alignment of Language Models

Reinforcement learning is all you need, for next generation language models

Show HN: OpenAI Multi-Client – Streamline Concurrent OpenAI API Requests

Show HN: CozoDB, Hybrid Relational-Graph-Vector Database

ChatGLM, a ChatGPT-like LLM developed by Tsinghua Univ, runnable on consumer GPU

Cozo 0.5: the versatile embeddable graph database with Datalog is half-way 1.0

Time travel in a database: a Cozo story

Cozo is a performant graph database that runs everywhere

Show HN: Cozo – new Graph DB with Datalog, embedded like SQLite

comments