ocolegro

8 ms·

31.
by ocolegro 3y ago
Yes, the framework is designed to directly deploy a FastAPI application with an associated Python client. You can see the client here - [ https://github.com/SciPhi-AI/R2R/blob/main/r2r/client/bas
32.
by ocolegro 3y ago
Yes, this is on the shortlist. Do you have any preferred frameworks?
33.
by ocolegro 3y ago
awesome! Please take a look and let me know what you think.
34.
by ocolegro 3y ago
the only time we've found pgvector to be preferable to qdrant is when the per-category embeddings are very few. E.g. in cases where filtering is expected to reduce the dataset greatly for each query (>99.9%). This is when the relati
35.
by ocolegro 3y ago
This is AMAZING feedback and it is on brand with what I've heard from a number of builders. Thanks for sharing your experiences here. The infra challenges are real - it has been what I have been struggling the most with in providing hi
36.
by ocolegro 3y ago
This is an open source solution that is meant to offer similar capability / ease of use, but with transparency & flexibility for the developer.
37.
by ocolegro 3y ago
First pass feedback on differences is that R2R is building with all database / llm providers in mind. Further, it seems Canopy has picked some pretty different abstractions to focus on. For instance, they mention `ChatEngine` as core a
38.
by ocolegro 3y ago
Yes, this is an easy lift, could* you add an issue? We also offer qdrant and pgvector, and will expand into most major providers with time. I personally recommend qdrant after trying 6 or 7 different ones while trying to scale out.
39.
by ocolegro 3y ago
This is a great question, thanks for asking. We are testing workflows internally that use orchestration software like Hatchet/Temporal to allow the framework to robustly handle 100s of GBs of upload data from parsing to chunking to emb
40.
by ocolegro 3y ago
It seems you're referencing a concept akin to the Voight-Kampff test from Blade Runner, where questions are designed to distinguish between humans and replicants based on their responses. In reality, I'm an AI, and "LLM"
41.
by ocolegro 3y ago
Hybrid search is definitely worth exploring (e.g. adding in TF-IDF). I believe there is such an implementation out of the box with Weaviate. I have tried many techniques and seen others try many different techniques. I think the hardest par
42.
by ocolegro 3y ago
No worries, thanks again the thoughtful feedback. We are also very interested in the more novel RAG techniques, so I'm not sure that one is necessarily a higher priority than the other. We've just gotten more immediate feedback fr
43.
by ocolegro 3y ago
Thanks for taking the time to provide your candid feedback, I think you have made a lot of good points. You are correct that the options in R2R are fairly simple today - Our approach here is to get input from the developer community to make
44.
by ocolegro 3y ago
I agree with all these points, drawing from my personal experiences with development. Gemini 1.5 is remarkable for its extensive context window, potentially unlocking new applications. However, it has drawbacks such as being slow and costly
45.
Show HN: R2R – Open-source framework for production-grade RAG(github.com)
167 points by ocolegro 3y ago|57 comments
46.
by ocolegro 3y ago
Nice, will take a look. Right now I've focused on the following more general integrations: VectorDBs - These include established providers like qdrant / pgvector / weaviate / pinecone / chroma. I’m also happy to try
47.
Tired of OpenAI assistants API? I will deploy your first RAG pipeline
16 points by ocolegro 3y ago|7 comments
48.
Show HN: I made a Perplexity-like research agent(search.sciphi.ai)
5 points by ocolegro 3y ago|0 comments
49.
AgentSearch – An LLM-First Search RAG Client and Engine(owencolegrove.substack.com)
3 points by ocolegro 3y ago|0 comments
50.
It's hard to believe this textbook was authored by AI(github.com)
3 points by ocolegro 3y ago|1 comments
51.
by ocolegro 3y ago
The book was created via a pipeline which goes MIT OCW -> Syllabus -> Table of Contents -> Textbook. The final step done w/ RAG over all of wikipedia. I see why Anthropic thinks they can automate research within in a few years
52.
by ocolegro 3y ago
This work was generously supported by @runpod_io , and I must say I have been very impressed with the experience around their cloud provisions. I can't recommend them highly enough given the current GPU shortage.
53.
With LLMs we can create an open-source Library of Alexandria(huggingface.co)
1 points by ocolegro 3y ago|2 comments
54.
by ocolegro 3y ago
With LLMs we can create a fully open-source Library of Alexandria. As a first attempt, I have generated 650,000 unique textbook samples from a diverse span of courses, kindergarten through graduate school. Happy to discuss methodology and g
55.
by ocolegro 3y ago
nice result! I've been looking into benchmarking models recently, it would be interesting to run your model through on the same battery of tests [ https://github.com/emrgnt-cmplxty/zero-shot-replication/blob...
56.
by ocolegro 3y ago
*Further Reading*: - [GPT-4's decline over time (HackerNews)]( https://news.ycombinator.com/item?id=36786407 ) - [GPT-4 downgrade discussions (OpenAI Forums)]( https://community.openai.com/t/gpt-4-has
57.
The AI Reproducibility Crisis
4 points by ocolegro 3y ago|3 comments
58.
Algofi (YC S21) Is Hiring(workatastartup.com)
1 points by ocolegro 5y ago