Hampus Londögård
@hlondogard.bsky.social
14 followers 59 following 9 posts
Posts Media Videos Starter Packs
hlondogard.bsky.social
Thanks, both @marimo.io and @pyodide.org makes an insanely productive and funny platform to work with!
marimo.io
marimo @marimo.io · May 29
🌟 Spotlight on @hlondogard.bsky.social!

Detect objects in-browser — zero servers.

Hampus’s marimo notebook uses @pyodide.org to proxy transformers-js-py and onnxruntime for on-device DETR — upload an image or URL for instant bounding boxes.

Try it live → marimo.io/p/@hlondoga...
hlondogard.bsky.social
(1/2)
Just dropped a new blog post on accelerating on-device inference with Transformers.js.py (wraps @hf.co transformers.js) and Pyodide (Py WASM) in #marimo app! 🚀 Dive into how to run efficient object detection right in your browser.
Link in post below 👇
Reposted by Hampus Londögård
yjernite.bsky.social
Amazing work from HF colleagues that's super relevant to AI policy - open development matches "frontier" performance 🤗

For policymakers, it means we need regulation that works for both open and closed model - not closed-first that disproportionately harms open!

Good news is...
1/2
A graph comparing the performance of different AI systems on a benchmark derived from the Informatics Olympiads, showing an open Hugging Face model coming third ahead of DeepSeek-R1 and Claude-3.7 Sonnet
hlondogard.bsky.social
Smallpond by DeepSeek!

I'm a bit late, but grateful that DeepSeek made their (distributed) data processing framework OSS.

Take-away?
<= 10TB? keep using Polars/DuckDB
> 10TB? think about the complexity added

But! I believe the future holds simpler tooling.
A small blog below!
Reposted by Hampus Londögård
ditzkoff.bsky.social
Guy who doesn’t even respond to the mothers of his children: reply to my email or you’re fired
Reposted by Hampus Londögård
hlondogard.bsky.social
🚀 Just published my latest holiday project: Image to LEGO! 🎄
Turn any photo into a buildable LEGO model with step-by-step instructions. 🧱
Deep Learning, voxelization, and some LEGO magic—read how I built it! 👇
📖 blog.londogard.com/posts/2025-0...

#python #LEGO #AI #DeepLearning
Image to Lego (Xmas Project) – Londogard Blog
A childs dream - turning any image into a LEGO Build. In this blog I walk through my xmas project.
blog.londogard.com
Reposted by Hampus Londögård
merve.bsky.social
supercharge your LLM apps with smolagents 🔥

however cool your LLM is, without being agentic it can only go so far

enter smolagents: a new agent library by @hf.co to make the LLM write code, do analysis and automate boring stuff! huggingface.co/blog/smolage...
thumbnail that says introducing smolagents
Reposted by Hampus Londögård
gradio-hf.bsky.social
Leffa from Meta is a 🆕 unified framework for controllable person image generation that enables precise manipulation of both appearance (i.e., Virtual try-on) and pose (i.e., Pose transfer).

Learn more about 📚 Code,🔥 Demo, and 🤗 Model ⬇️
hlondogard.bsky.social
Object Detection is still only supported using TensorFlow? Was a big let-down for our team with the 3.0 release..

Hoping it'll be truly multi-backend soon and we forget the days of TF RaggedTensor requirement