Jesse

jessepisel
ยท

AI & ML interests

computer vision, generative ai, agentic

Recent Activity

Organizations

ThinkOnward's profile picture

jessepisel's activity

upvoted 2 articles 4 days ago
reacted to as-cle-bert's post with ๐Ÿ‘ 8 days ago
view post
Post
2642
I just released a fully automated evaluation framework for your RAG applications!๐Ÿ“ˆ

GitHub ๐Ÿ‘‰ https://github.com/AstraBert/diRAGnosis
PyPi ๐Ÿ‘‰ https://pypi.org/project/diragnosis/

It's called ๐๐ข๐‘๐€๐†๐ง๐จ๐ฌ๐ข๐ฌ and is a lightweight framework that helps you ๐—ฑ๐—ถ๐—ฎ๐—ด๐—ป๐—ผ๐˜€๐—ฒ ๐˜๐—ต๐—ฒ ๐—ฝ๐—ฒ๐—ฟ๐—ณ๐—ผ๐—ฟ๐—บ๐—ฎ๐—ป๐—ฐ๐—ฒ ๐—ผ๐—ณ ๐—Ÿ๐—Ÿ๐— ๐˜€ ๐—ฎ๐—ป๐—ฑ ๐—ฟ๐—ฒ๐˜๐—ฟ๐—ถ๐—ฒ๐˜ƒ๐—ฎ๐—น ๐—บ๐—ผ๐—ฑ๐—ฒ๐—น๐˜€ ๐—ถ๐—ป ๐—ฅ๐—”๐—š ๐—ฎ๐—ฝ๐—ฝ๐—น๐—ถ๐—ฐ๐—ฎ๐˜๐—ถ๐—ผ๐—ป๐˜€.

You can launch it as an application locally (it's Docker-ready!๐Ÿ‹) or, if you want more flexibility, you can integrate it in your code as a python package๐Ÿ“ฆ

The workflow is simple:
๐Ÿง  You choose your favorite LLM provider and model (supported, for now, are Mistral AI, Groq, Anthropic, OpenAI and Cohere)
๐Ÿง  You pick the embedding models provider and the embedding model you prefer (supported, for now, are Mistral AI, Hugging Face, Cohere and OpenAI)
๐Ÿ“„ You prepare and provide your documents
โš™๏ธ Documents are ingested into a Qdrant vector database and transformed into a synthetic question dataset with the help of LlamaIndex
๐Ÿ“Š The LLM is evaluated for the faithfulness and relevancy of its retrieval-augmented answer to the questions
๐Ÿ“Š The embedding model is evaluated for hit rate and mean reciprocal ranking (MRR) of the retrieved documents

And the cool thing is that all of this is ๐—ถ๐—ป๐˜๐˜‚๐—ถ๐˜๐—ถ๐˜ƒ๐—ฒ ๐—ฎ๐—ป๐—ฑ ๐—ฐ๐—ผ๐—บ๐—ฝ๐—น๐—ฒ๐˜๐—ฒ๐—น๐˜† ๐—ฎ๐˜‚๐˜๐—ผ๐—บ๐—ฎ๐˜๐—ฒ๐—ฑ: you plug it in, and it works!๐Ÿ”Œโšก

Even cooler? This is all built on top of LlamaIndex and its integrations: no need for tons of dependencies or fancy workarounds๐Ÿฆ™
And if you're a UI lover, Gradio and FastAPI are there to provide you a seamless backend-to-frontend experience๐Ÿ•ถ๏ธ

So now it's your turn: you can either get diRAGnosis from GitHub ๐Ÿ‘‰ https://github.com/AstraBert/diRAGnosis
or just run a quick and painless:

uv pip install diragnosis


To get the package installed (lightning-fast) in your environment๐Ÿƒโ€โ™€๏ธ

Have fun and feel free to leave feedback and feature/integrations requests on GitHub issuesโœจ
reacted to fdaudens's post with โค๏ธ about 1 month ago
view post
Post
2691
โญ๏ธ The AI Energy Score project just launched - this is a game-changer for making informed decisions about AI deployment.

You can now see exactly how much energy your chosen model will consume, with a simple 5-star rating system. Think appliance energy labels, but for AI.

Looking at transcription models on the leaderboard is fascinating: choosing between whisper-tiny or whisper-large-v3 can make a 7x difference. Real-time data on these tradeoffs changes everything.

166 models already evaluated across 10 different tasks, from text generation to image classification. The whole thing is public and you can submit your own models to test.

Why this matters:
- Teams can pick efficient models that still get the job done
- Developers can optimize for energy use from day one
- Organizations can finally predict their AI environmental impact

If you're building with AI at any scale, definitely worth checking out.

๐Ÿ‘‰ leaderboard: https://lnkd.in/esrSxetj
๐Ÿ‘‰ blog post: https://lnkd.in/eFJvzHi8

Huge work led by @sasha with @bgamazay @yjernite @sarahooker @regisss @meg
  • 1 reply
ยท
upvoted an article about 2 months ago
view article
Article

Welcome to Inference Providers on the Hub ๐Ÿ”ฅ

โ€ข 435
reacted to fdaudens's post with โค๏ธ about 2 months ago
view post
Post
8827
Yes, DeepSeek R1's release is impressive. But the real story is what happened in just 7 days after:

- Original release: 8 models, 540K downloads. Just the beginning...

- The community turned those open-weight models into +550 NEW models on Hugging Face. Total downloads? 2.5Mโ€”nearly 5X the originals.

The reason? DeepSeek models are open-weight, letting anyone build on top of them. Interesting to note that the community focused on quantized versions for better efficiency & accessibility. They want models that use less memory, run faster, and are more energy-efficient.

When you empower builders, innovation explodes. For everyone. ๐Ÿš€

The most popular community model? @bartowski 's DeepSeek-R1-Distill-Qwen-32B-GGUF version โ€” 1M downloads alone.
ยท
upvoted an article about 2 months ago
view article
Article

Open-R1: a fully open reproduction of DeepSeek-R1

โ€ข 810