Andy Oakley
@andyoakley.com
Search relevance and quality. Machine learning, information retrieval, and data analysis.
How to use LLMs to make knowledge work easier? Sometimes it's helpful to think of LLMs as a team of enthusiastic, capable interns. Magentic-One proposes a way to put that into practice. arxiv.org/html/2411.04...
Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks
arxiv.org
November 16, 2024 at 10:29 PM
How to use LLMs to make knowledge work easier? Sometimes it's helpful to think of LLMs as a team of enthusiastic, capable interns. Magentic-One proposes a way to put that into practice. arxiv.org/html/2411.04...
Sometimes ideas that don’t seem like they should work, actually do, and do so surprisingly well. emschwartz.me/binary-vecto...
Binary vector embeddings are so cool
Vector embeddings by themselves are pretty neat. Binary quantized vector embeddings are extra impressive. In short, they can retain 95+% retrieval accuracy with 32x compression 🤯.
emschwartz.me
November 16, 2024 at 10:21 PM
Sometimes ideas that don’t seem like they should work, actually do, and do so surprisingly well. emschwartz.me/binary-vecto...
Open WebUI makes a strong entry for "most generically named tool", but is among the best for working with locally hosted LLMs github.com/open-webui/o...
GitHub - open-webui/open-webui: User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
User-friendly AI Interface (Supports Ollama, OpenAI API, ...) - open-webui/open-webui
github.com
November 13, 2024 at 5:57 AM
Open WebUI makes a strong entry for "most generically named tool", but is among the best for working with locally hosted LLMs github.com/open-webui/o...
Impressive performance from Qwen2.5-Coder. Even just a few years ago it was hard to imagine we'd be running models with these capabilities at home qwenlm.github.io/blog/qwen2.5...
Qwen2.5-Coder Series: Powerful, Diverse, Practical.
GITHUB HUGGING FACE MODELSCOPE KAGGLE DEMO DISCORD
Introduction Today, we are excited to open source the “Powerful”, “Diverse”, and “Practical” Qwen2.5-Coder series, dedicated to continuously promotin...
qwenlm.github.io
November 13, 2024 at 5:54 AM
Impressive performance from Qwen2.5-Coder. Even just a few years ago it was hard to imagine we'd be running models with these capabilities at home qwenlm.github.io/blog/qwen2.5...